This action might not be possible to undo. Are you sure you want to continue?
P. Geoffrey Allen*
Department of Resource Economics, University of Massachusetts, Amherst, MA 01003, USA
Forecasts of agricultural production and prices are intended to be useful for farmers, governments, and agribusiness industries. Because of the special position of food production in a nation’s security, governments have become both principal suppliers and main users of agricultural forecasts. They need internal forecasts to execute policies that provide technical and market support for the agricultural sector. Government publications routinely provide private decision makers with commodity price and output forecasts at regional and national levels and at various horizons. Routine forecasts are not found in the agricultural economics journals that are the sources for most of this review. The review emphasizes methodological contributions and changes. Short-term output or ‘outlook’ forecasting uses a unique form of leading indicator. Because the production process has long been well understood, production forecasts are based on the quantifiable features of livestock or a growing crop. Price forecasts are largely made by conventional econometric methods, with time series approaches occupying minor roles. Because of the dominance of agricultural economists, there has been an overemphasis on explanation, and little interest in the predictive power of models. In recent years, some agricultural economists have begun to compare forecasts from different methods. Findings generally conform to widely held beliefs. For short-term forecasting, combining leads to more accurate forecasts, better than those produced by vector autoregression, which surprisingly is the best single method. Also surprising is that econometric models and univariate methods both do badly compared with naive models.
Key words: Agricultural prices; Meta-analysis; Sector modeling Agricultural production: Forecast comparisons: Econometric forecasting; Judgmental forecasting;
Economic forecasting in agriculture has some features in common with business forecasting and with macroeconomic forecasting. But over time, it has developed a focus of its own. Just (1993) the first and Just and Rausser of (1993) characterize econquarter century agricultural
fax: (413) 545-5853
omits research (from about 1925-1950) as prescriptive: recommendations were made to farmers and managers in order to increase profits. During the second quarter century, the profession shifted toward prediction, broadly defined, including use of econometric techniques for estimating elasticities and forecasting prices. The third quarter century, from 1975 onwards, has been characterized by research on policy, trade and the global economy and expansion to environmental and resource problems. Throughout
0169-2070/94/$07.00 0 1994 Elsevier SSDZ: 0169-2070(94)00519-I
B.V. All rights
i International Journal of Forecasting
10 (1994) 81-13.5
the entire period, and more markedly of late, explanation of past behavior has been the dominant focus of agricultural supply modeling, which is the area to which most agricultural forecasting belongs. Because an assured food supply is important to national security, governments have attempted to quantify agricultural production and to exert some control over it. In the beginning, simply collecting and tabulating data on the current agricultural situation was a major challenge, and agricultural statisticians played a major role in the development of statistical methods [USDA (1969)]. Data revision was frequent. Estimates of production, for example, were subject to revision after a new census had been tabulated. The large number of ‘Situation reports’ or similarly titled publications indicates the fascination of agricultural statisticians with estimating the current status of a data series. Most agricultural forecasters were trained as either statisticians or agricultural economists. The two professions have formed what has been, at times, an uneasy alliance. Statisticians have been largely responsible for developing the approach to outlook forecasting that relies on indicator analysis. Agricultural economists have tended to emphasize ever more complicated econometric models. They have worried a great deal about providing convincing explanations of economic phenomena, with the assumption (generally untested) that this would be useful not only for decision making but also for forecasting.
A brief history
The aim of the review is to provide a summary of the main approaches used by agricultural forecasters, with an assessment of the strengths and weaknesses of each approach. The review tries to answer two opposing questions: which results from research into agricultural forecasting can be generalized to all kinds of forecasting? Which conclusions from general forecasting research apply to agriculture? The major sections discuss the methods of forecasting as they appeared chronologically. Correspondingly, the methods become increasingly complex.
Judgmental forecasts appeared first and are still significant components of short-term outlook forecasts. There is a long history of econometric analysis, starting with single equation studies. Greater computing power saw larger multiequation analyses appear. First came studies that performed partial analysis on a single commodity sector. The interrelation among certain sectors, particularly livestock and feed, was recognized early on. Early studies on the agricultural sector in aggregate contained few equations and were of simple form. Later, multiequation, multisectoral econometric studies appeared. Although trend extrapolation methods were widely used in commodity outlook studies, agricultural applications of modern time series methods did not appear until the early 1970s. Gradually, more sophisticated efforts were made by a handful of agricultural economists, their work ranging from various forms of composite forecasts to vector autoregression (VAR) and state space models. Because of the historical interest in decisionmaking by micro-economists, there has been a scattering of articles relating forecasting to the making of choices, including those concerning probabilistic forecasts, value of information and comparison of forecasting methods when used to aid a specific purchase or sales decision. Present work in agricultural forecasting reflects the culmination of two strands of research. From earliest times, statisticians have analyzed agricultural data, in part because it was available, but also because the results were of value to farmers and other business people. Second, predicting the outcomes of different policies is a major activity of many agricultural economists. 1.2. Scope of the review
Articles on the methods or results of forecasting were extracted from an exhaustive search of the main agricultural economics journals. Searches of DIALOG databases from 1969 (Agricola) or 1970 (Journal of Economic Literature) to 1992 and of Government indexes added to the list of studies. Coverage of ‘journals and other sources is shown in the appendix. Expanded
I International Journal of Forecasting 10 (1994) 81-135
tables that list the source studies for each table entry can be obtained from the author. The review covers those agricultural commodities and inputs that are the subject of forecasts regularly made by government departments of agriculture and reported in their publications. These include, at national, regional and the production and value local aggregations, (equivalently area, yield and price) of crops, and livestock numbers, production and value. Also included in the review are industrial products such as vegetable oils and meals, grain by-products, and agricultural inputs (e.g. fertilizer, pesticides, but not general petroleum products). I attempted to include every study that compared forecasts of agricultural commodities or inputs done by different methods, as well as all articles that evaluated the performance of agricultural forecasters and their methods. Studies that focus strictly on market efficiency are excluded, as are studies that use the commodity futures markets as a test of efficiency rather than in a comparison of forecasting ability. Readers interested in such issues should refer to two recent studies on livestock futures price movements around the date of release of USDA inventory reports [Colling and Irwin (1990), Schroeder et al. (1990)]. These studies also review earlier work in that area. Also excluded from this review are commodity forecasts where the focus is the industrial use of food and fibre products. Forestry, fishing and aquaculture sectors are omitted as well. The review relies on published work. It would be a mammoth task to collect and assess a representative sample of published government forecasts and even more difficult to acquire private company forecasts and unpublished government forecasts. [Two studies that have performed this task on USDA short-term outlook forecasts are included in the comparative review: Gunnelson et al. (1972), Surls and Gajewski (1990).] 1.3. A note on evaluation Evaluation of the different forecasting approaches is a key feature of the review. Making
fair comparisons can be difficult for two reasons. First, the only true test of a forecasting model is its forecasting performance in the post-estimation period. Partly, no doubt, because of short data series, testing, if done at all, has often been by within-sample simulation. With post-sample testing, particularly of econometric models, several problems arise. Parameter updating may be done as data become ‘known’ in the post-sample period, or updating may be totally omitted. A form of forecast contamination frequently occurs when actual values of exogenous variables are used, though these would be unknown at the time of the forecast. We do not at present know how much such contamination misleads model selection and model accuracy. In real-world forecasting, unknown exogenous variables would themselves need to be forecast. And forecasts may be modified by the analyst’s judgment before release-a kind of informal combined forecast, even if not acknowledged as such. Second, studies use various criteria for measuring how well a method makes point forecasts and turning point predictions. The criterion used by each study to make comparative rankings is noted in the detailed tables available from the author. Root mean square error (RMSE) is the most widely reported accuracy measure and was used to construct the table entries wherever possible. Aggregating rankings from studies using different criteria is cavalier, to say the least. Even worse, the careful study by Armstrong and Collopy (1992) shows that neither of the commonest criteria (RMSE and mean absolute percentage error, MAPE) is the most reliable for choosing the best method. Similar sensitivity to choice of criterion occurs when attempts are made to rank methods for their ability to forecast turning-points (discussed further in section 7.2).
The nature of agricultural production and the historical relations among the different groups of participants in agriculture make agriculture different from most economic activity. Most prod-
P.G. Allen I International Journal of Forecasting 10 (1994) 81-135
uct is unbranded and sold in markets where individual suppliers have no say in price determination. Both nature and government policy can have a major impact on a farmer’s production and profits. Farmers and others connected with agriculture are used to receiving technical and economic information from publicly supported institutions. 2.1. Characteristics of agricultural production
reasonably well after they have occurred, which is useful in making post-harvest production estimates, but not pre-harvest forecasts. 2.2. Producers of agricultural forecasts
Agricultural production is unusual compared with most business activity in its strong dependence on biological processes. Farmers have minimal ability to alter the rate of development of a crop or animal. Second, for most commodities, the production cycle is measured in months or years. Other features impose dynamic structure, especially on prices: seasonal impacts on production, high cost of adjustment once production is underway and the need to carry inventory. Estimation of leading indicators therefore became a major part of short-term agricultural production forecasting, dominating any work on price forecasting. The estimation of leading indicators was a natural extension of the data gathering activity concerning current production or inventories. For example, estimation of acres planted to spring wheat is a good indication of harvested acreage. In no other sector has leading indicator analysis found such long-term and widespread use. Agricultural production appears to meet the four conditions laid down by Armstrong (1985, p. 196) for good forecasts by econometric methods: there should be strong causal relationships, relations should be capable of being measured accurately, causal variables should change substantially and it should be possible to forecast changes in causal variables. Unfortunately, econometric methods do poorly at forecasting agricultural production and prices. The most likely reason is the great influence on production of random shocks. Relative to most manufacturing activity, agriculture is greatly influenced by unpredictable random events such as droughts, hoods and attacks by pests. The consequence of these shocks on production can be assessed
The predominant forecaster of production, prices and trade of agricultural commodities and inputs in most countries is central government. The Economic Research Service of the United States Department of Agriculture (USDA-ERS) contains the largest agglomeration of agricultural economists and produces the greatest number of agricultural forecasts. Government commodity specialists are the main providers of outlook information in Australia, Canada and the US [Johnson et al. (1982)]. Reports on the situation and outlook for commodity and input markets at local, national and world levels are issued from one to twelve times a year depending on commodity and country. Some agencies issue regular medium-term forecasts (2-5 years ahead). For example, Agriculture Canada has issued medium-term outlook reports twice a year since 1987 [Cluff (1990)]. Long-term projections are generally issued only irregularly, and usually for groups of commodities. Although governments publish many forecasts, often as regular series, they also make many forecasts solely for internal use, for example, the USDA forecasts of the budgetary cost of the farm program. Other public agencies, from the Food and Agricultural Organization of the United Nations to regional or provincial governments, also produce forecasts. University faculty and (in the US) extension economists prepare forecasts for general release as part of short-term outlook programs for local farmers and agribusinesses. They may also present forecasts in scholarly publications; these usually have a methodological focus. Private companies that process or trade commodities or supply inputs produce forecasts for with relatively simple in-house use, typically models combined with judgment. They are probably closest to business forecasters in both approach and objectives. Private consultants also produce forecasts for sale, most frequently as
P.G. Allen I International Journal of Forecasting 10 (1994) 81-135
adjuncts to large-scale macroeconomic models. Farmers practically never produce formal forecasts, though most of them doubtlessly form a judgment about future outcomes of their business choices. 2.3. Users of agricultural forecasts Farmers may rarely make forecasts, but they form the largest group of users. They need to make production and marketing decisions that may have financial repercussions many months in the future. Short-run commodity outlook forecasts, at least in the US, have tended to emphasize production and inventory information. Farmers have more use for price forecasts. Once committed to a product, farmers are price takers. They produce goods that are homogeneous or highly substitutable with the goods of their competitors, who may either be their neighbors or live halfway round the world. They have no concern with problems common in manufacturing, such as the amount of sales of a branded product or what quantity of a specific model to keep in inventory. But farmers, especially those in developed countries, must also be concerned with the ways in which changes in government policy will alter their business conditions. Agricultural journalists represent a second kind of audience for commodity forecasts. They are not users in the sense of being makers of decisions based on forecast information. They provide an indirect way for readers and listeners (mainly farmers) to receive outlook forecasts. Processors of food and fiber, and others in the marketing chain, need forecasts to aid in their purchasing and storing decisions. They too would probably like price forecasts, but would be able to make greater use of production forecasts in their decisions than would farmers. Larger businesses also supplement public forecasts with their own in-house ones. Governments in many countries intervene in agricultural production to protect domestic agriculture and provide food security. For this they need two kinds of information. First, for legislation and, to a much lesser extent, for program implementation, governments need to know the
consequences of different policy choices on different groups in society. Agricultural economists have been especially willing, over the last 30 years, to build ever larger models to provide answers to policy questions. Emphasis has been placed on comparing proposed policies via simulations, which has measurably assisted legislators. Forecasts of output and prices are conditional on the policy actually selected. To date, efforts to forecast which policy will be selected have been minimal. [See Rausser (1982), Chapter 18 for a review of the theory and empirical applications of endogenous government behavior.] Neither have government or academic economists done much to evaluate a model’s ability to forecast the actual consequences of an adopted policy. Second, in monitoring the progress of farm programs designed to control supplies or support prices, governments would like to know about the effectiveness of the program and anticipated budget outlays.
Government agencies have issued short-term forecasts of prices and production for many years. In the early years, the reports contained much about current situation and little about outlook [Hudson and Furniss (1966)]. The development of methods of estimating and forecasting agricultural production in the United States forms the basis for the organization of this section. [For a politically oriented statistical history, see US Department of Agriculture (1969). For detailed technical descriptions of data gathering and analysis, see USDA (1983). For a summary of statistical methods and detailed information on timing and content, e.g. estimates, forecasts and intentions of crop and livestock reports, see USDA (1989).] The first agricultural forecasts were estimates of crop appearance, referred to as ‘condition’. Initially these were purely judgmental assessments made by crop reporters, who compared the crop’s current appearance and vitality with that of a normal year. These assessments were soon used to calibrate formal production fore-
Because some crop reports covered more acreage of a given crop than others. In 1910. then aggregated using estimated acreage in the same area as weights [USDA (1983)]. By taking. for example. when the final yield in that state for that year was 28 bushels. For longer horizons. and inventory estimates for five livestock species. Formal correction for sampling bias and the relation of past intentions to past actual performance followed. and by 1929 had evolved into a standard procedure [Kunze (1990)]. of 19 crops in 21 Northern states and the Nebraska Territory [Newell and Warrington (1962)]. the editor of the American Agriculturalist sought and published monthly crop ‘condition’ summaries from May to September using information submitted by farmer subscribers [Ebling (1939)].G. the crop reporting agency (at that time the USDA Bureau of Statistics) was issuing quantitative estimates of acreage. Allen I International Journal of Forecasting 10 (1994) 81-13s casts. Later. It was run by the appropriate government agency and attended by government. provided the USDA with its main means of crop production forecasting for nearly 100 years. 3. Quantitative analysis of condition and use of par The first USDA forecast. a second type of judgmental forecast resulted from surveys of farmers’ intentions to plant specific crops or to breed specific animals. 3. Between 1912 and 1929. biases from problems in forming judgments or from technology changes. production and value for 13 crops. such as number of flowers or number of ears. By 1920. gave a 100% equivalent yield of 35 bushels (28 + 80 X 100). the first federal and provincial conference was held in Ottawa in February 1934. academic and private agricultural economists. a condition value of 80% for winter wheat on 1 July. In Canada. as assessed by farmer or government crop reporters. Mid-season sampling of observable characteristics enabled forecasters to improve their predictions on ultimate yield. was made in May 1912 for winter wheat and from the following month for most field crops. Judgmental reporting of condition estimates had roughly doubled [Becker and Harlan (1939)]. The following year’s yield forecast was simply par yield multiplied by condition. The first such conference occurred in the US on 20-21 April 1923. The national annual outlook conference became a feature in most developed countries. excluding cotton. Condition summaries are still important in early-season estimation of field crop yields [USDA (1983)]. the statistician obtained the 1 July ‘par’ yield.1.2. Judgmental estimates of the state of a growing crop. It was later moved to February and is now held (more usefully for production planning) in December. USDA statisticians first calculated a yield forecast within a crop reporting district. correlation of condition and final yield replaced the par method as the means of yield forecasting [USDA (1969)]. The analysts then present forecasts and receive feedback from the audience. The following year. a 5 year moving average of 1 July ‘full’ yields.86 P. After 1929. Although acreage weighting removed some of the biases caused by the non-probability samit could not accommodate pling procedure. For example. Field crop production forecasts were calculated as the estimated acres available for harvest multiplied by the yield forecast. to ultimate yield. The ultimate development of short-term crop forecasts was based on ‘objective yield’ estimates. a politically sensitive crop. with adjustment for trend. Essentially. Agronomic studies related observable intermediate characteristics. Today’s typical conference features a number of commodityspecific sessions in which government analysts review the present situation and forces of change. the number of . Its first monthly crop report stated the condition. the USDA took over the task. the reported condition of various crops was interpreted as a forecast of yield based on the ‘par’ method [Becker and Harlan (1939)]. the monthly condition of each crop in each state was converted to ‘full’ or 100% equivalent yield. statisticians used charts of monthly condition and final yield plotted against time In 1862. as opposed to condition report. as of May 1863. condition reports for 23 crops and pasture.
was divided into estimates of final production. Each past year. plant development might be delayed compared with a normal year. the condition at time of harvest. 3. One problem was that the forecast was required for the entire United States on the same date each year. Frank Parker proposed a plan to improve cotton yield forecasts by counting the number of plants and number of bolls of cotton. In 1956. For fruit production. since it had developed the area probability sampling method and had access to studies on the effect of weather on yield of cotton. Commentators at the time [Wallace (1953)] suggested that the USDA already had the means to make improvements. Objective yield forecasts In 1925. Similar problems in relating observable characteristics of young corn and young soybeans to yield were reported by. These two patterns were charted and used first to predict the next historical par from the latest reported condition and then to predict total production in the current year. This approach was found to give forecasts as good as those relating reported condition to yield or to actual production [Palmer and Schlotzhauer (1950)]. respectively. In 1951. Farmers were estimated to have lost about $125 million in revenues [U. fruits are in different stages of development and the numbers visible exceed the final number of bolls. for the 1 August forecast. The regression of reported par on historical par showed that reporters tended to slightly underestimate par (leading to a consequent underestimate of yield) coupled with over-optimism in good years. On 1 August. the USDA Crop Reporting Service made estimates of cotton production based on these data that turned out to overstate actual production by about 15%. Also. As the original culprit. Surveys conducted in 1954 and 1955 established the relation between counts of different kinds of fruits on 1 August and final numbers of bolls. multiple regression of condition and time variables on final yield was used to make yield forecasts. The method essentially required a detailed quantitative understanding of plant development so that observable characteristics measured earlier in the season could be related to final harvest weight by regression analysis. a modified par method was used [Becker and Harlan (1939)]. ‘objective forecasting’ of crops was advancing on several fronts. Although the USDA makes cotton price forecasts for internal use. dealers were paying farmers relatively low prices for the supposed bumper crop. The result was the historical par or percentage of full production for that year. Presumably the purchasers of cotton gained the windfall $125 million once the forecast error was revealed by the end of the year. Congress (1952)]. summarized from crop reporters’ judgments. especially where yield trends were noted. Huddleston (1958) and Kelly (1957). and between fruit count per plant and average weight per fruit. By 1956. plants at the northern limit of a crop’s production area would be less developed than plants in states to the south. for example. corn and wheat. By 1 September.G. Objective yield surveys became operational for cotton and corn . Finally. However. principally the establishment of a special unit within the Bureau of Agricultural Economics to examine problems with the present methods and devise improvements. Even later. to overcome the biases and uncertainties which accompany judgmental assessments. it had to be possible to count. the final number of bolls has appeared and the boll count is a good predictor of total yield. Historical par values also trended upwards over time. Allen I International Journal of Forecasting 10 (1994) 81-135 87 to judgmentally adjust yield forecasts.3.P. Since the commodity market relies heavily on crop forecasts. weigh or measure the characteristics in a standard way.S. Such data have been regularly collected since 1928 [Becker and Harlan (1939)]. the relations were used for a state by state forecast of cotton yield in a ten state region [Hendricks and Huddleston (1957)]. the cost of the error prompted a Congressional inquiry from which several recommendations were made. cotton was the first crop to be investigated. However. Congress still prohibits their publication.
G. The USDA has continued to refine its sampling techniques to maintain sampling accuracy while reducing cost. Congress passed legislation prohibiting future intentions reports for cotton. the USDA had sufficient statistics on acres planted to be able to convert planting intentions reported by farmers into ‘prospective plantings’ [Becker and Harlan (1939)].5. usefulness and accuracy of USDA outlook reports. 3. the list frame can be expanded to represent the entire population [USDA (1983)]. The legislation was not repealed until 1958 [USDA (1969)]. the first pig crop report was issued in 1922. In the US. 3. reaching the rest of the US by 1965. At present. for wheat in 1962 and for soybeans in 1967. Probability sampling for acreage estimates started in June 1961 in 15 states [Trelogan (1963)]. the USDA published the first report on intended acreage for nine spring-sown crops. In multiple-frame sampling. Criticisms on timing made some 25 . The obvious solution of comparing planting intentions with actual planted acres had to await data on planted acreage. More accuracy can be achieved with a smaller sample and standard errors can also be computed. placed the non-probability mailings from about 1979. By knowing the inventories or intentions of each farmer in the area frame. based on a survey delivered to pig farmers by rural mail carriers. an underestimate of the actual increase. Probability sampling requires definition of a proper random sample. Evaluation of short term forecasting for outlook work Farmers and economists have criticized the timing. 3. the forecast caused activity on the cotton exchanges and some reduction in price. Allen I International Journal of Forecasting 10 (1994) 81-13. process- ing vegetables and mushrooms. Estimates of intended breeding are supplemented by inventory surveys for all classes of stock. Its disadvantages are the difficulty of expanding sample findings to the population and the inability to estimate sampling errors. Semi-annual inventory surveys are the main method of forecasting cattle production. enumeration is sometimes by interview rather than by mail. The following year. In 1923. Probability sampling Non-probability sampling by mail is cheap. Producer intentions In 1918.88 P. A large percentage increase in intentions to plant in the following year might only represent an attempt to return to the normal pattern. The series was constructed by summarizing farmers’ responses to the question: ‘Compared with the acreage of (name of crop) you harvested last year. based on a non-probability survey of individual farmers.4. and have since expanded to include potatoes. it can sometimes be much less. Multipleframe sampling supplemented the livestock mail and probability surveys entirely resurveys. on the grounds that such reports were more harmful than beneficial [Becker and Harlan (1939)]. several tree nuts and citrus fruits [USDA (1983)]. USDA administrators must have had second thoughts about the effort because they kept the results secret [Ebling (1939)]. when drought or disease results in a yield too small to harvest profitably.6. Because the sample unit may be a collection of fields and not necessarily an entire farm.~ yield forecasts in 1961. particularly when dealing with specialized types of production. with planted estimates for fresh vegetables and acreage melons [USDA (1983)]. including cotton. Breeding intentions have since been surveyed quarterly in the major producing states. One danger became apparent in focussing on a series of intentions to plant. and semi-annually elsewhere. Farmers reported that they intended to increase cotton acreage by 12%. all producers in a randomly selected area (the area frame) are identified as belonging or not belonging to a master list of names (the list frame). acreage intentions or prospective plantings are reported for all major field crops except cotton. In a response that was to be echoed almost 30 years later. how much percentage increase or decrease in acreage do you intend to plant this year?’ While the harvested acreage can never exceed planted acreage. By 1938. the USDA sent out a questionnaire in order to find out how great an acreage of spring wheat farmers intended to plant.
the carryover decision cannot be avoided. decision makers would on average choose to carryover the correct amount. farms have concluded that actual production is both inefficient in use of inputs and too conservative compared with decision-makers’ stated risk preferences. prices to farmers would be higher. for grain) at which marginal social value (benefits to consumers) equals marginal social cost (storage charges). especially for prices. Dalrymple’s (1987) survey of 134 US businesses . Using historical information only. from quarterly (for hogs and pigs) or monthly (cattle on feed) to weekly (broiler hatchery). Estimates useful for livestock producers appear frequently. Prospective plantings reports appear at the beginning of March. Point estimates of production and yield are almost never accompanied by confidence intervals or similar indications of reliability. These prove to be myths. Once crop or animal production is underway. (2) if outlook reports were not released. Bullock et al. As long as the information in a forecast causes decision makers to store a quantity closer to the optimal carryover. The need is as great today. farmers’ responses to price signals are limited. outlook reports have provided a form of probabilistic information for a long time. the agricultural outlook for 1930 stated [quoted in Kunze (1990) p. since optimal resource allocation in a risky environment depends on the decision-maker’s risk-reward tradeoff as well as knowledge of the production Studies of production on individual response. Timm (1966)]. For example.G. not consistently higher. before many farmers have begun to plant. Movements in futures prices when hogs and cattle reports appear [reviewed by Schroeder et al. With no information. government agencies could issue a price forecast and then explain the logic behind it [Freebairn (1978)]. And forecasts of acres or animal numbers need to be translated into total production and then into price. Nelson (1980) was the first to suggest how such an outlook program might be set up.P. A number of beliefs exist [Bullock et al. The third myth is more difficult to demolish. Freebairn (1978)]. The annual outlook conference has been moved forward to December. 2571: These reports are not designed to tell individual farmers what to do. A continuing problem is ensuring the usefulness of forecasts. Today. A widely accepted psychological explanation is that people explain their successes as a result of their own efforts and their failures as a result of things outside their control. Using phrases such as ‘ample supplies will likely keep prices below last year’s levels’.2). Business forecasts suffer from the same problem. but the carryover quantity will be sub-optimal. As a compromise. In a simple two-period model. Improved accuracy through the use of better data and techniques was an early concern [Botturn (1966). Some actions are possible and perhaps profitable. For example (as discussed later in section 8. such information has value. Planting or breeding intentions are reported instead. Daly (1966)] may no longer be valid. Daly (1966). Farmers want price forecasts when the planting or breeding decision is being made. (1982) show diagrammatically how perfect information (a perfect forecast) can be used to determine the inventory carryover (say. The studies described in section 8. both the US and Australian outlook programs seemed deliberately to leave the more difficult step of price forecasting to individual farmers. the discretion of the commodity analyst appears to determine whether or not outlook reports contain quantitative price forecasts in the narrative. as is the impact of their actions on forecasted prices. (3) inaccurate reports are a major cause of shortrun resource misallocation. but from year to year price would be either too high or too low. (1982)]: (1) production forecasts must be perfectly accur- ate to be of value. (1990)] suggest that market participants use the outlook information (perhaps because it provides a more accurate estimate of current situation).2 show limited gains in profitability from using forecast information. Allen I International Journal of Forecasting 10 (1994) 81-135 89 years ago [Bottum (1966). but to give them the basic facts upon which to make intelligent decisions in view of their local conditions. Although there was early recognition of the need to provide probabilistic information [Bottum (1966). forecasts can be used for crop storage and livestock rearing decisions. Historically. Matters have improved only slightly.
. number of correct one step ahead increases or decreases divided by total number of changes forecast. Sometimes the comparison is with a naive no change forecast. but generally worse than a Table 1 Outlook forecasts range of methods. As Armstrong (1985.90 P. (root) mean squared error (R)MSE. even though this was one of the purposes of reporting intentions data [USDA (1969) pp. Use of planting intentions data led to noticeable improvements in forecast accuracy. Freebairn (1975). When prospective plantings data are reported to the public. 67-681. pp. F(t . Naive is either a no change forecast or trend forecast or (in one study) the futures price. Later in this review. covering seven crops over 42 years. Freebairn (1975). Evidence ranging from capital investment.F(t ~ l)). experts and commodity analysts do seem better at determining current status than at making forecasts. Allen I International Journal of Forecasting 10 (1994) 81-135 found that only 22% of sales forecasts were ‘frequently’ or ‘usually’ accompanied by interval estimates. compared with which outlook is generally. Iv = livestock products. Turning point ratio. or the average of Then’s R. Elam and Holder (1985). compared studies with others Number 36 cr 5 Iv 1scr 6 Iv 15 3 9 10 cr Iv cr Iv series Finding Type Comparison Quant Price Number with naive 6 4 Accuracy better than naive about 70% of time Accuracy better than naive or trend about 70% of time Quant Price 3 3 Turning point better than naive about 85% of time Turning point better than naive. mean absolute percentage error MAPE. most farmers are unable to change their production plans.1) is previous forecast (naive forecast in this case) and A(t) is actual production. The top panel of Table 1 summarizes studies that compared forecasts from mechanical methods with outlook [Baker and Paarlberg (1952). that they report correctly and that they are unlikely to change [Armstrong (1985)]. a group including outlook forecasts along with other expert forecasts will be shown to be more accurate than naive and exponential smoothing methods. Foote and Weingarten (1958) made forecasts of 19 crops using production data only and compared these with forecasts based on planting intentions. Gunnelson. Jolly and Wong (1987)]. found slightly better results for (Australian) BAE forecasts. where F(t) is current forecast. who studied annual forecasts of price and output for ten crop and livestock products over 8 years. trend or random about 70% of time Econometric Quant Price Quant Price studies 3 1 1 1 25 cr 1 Iv 1 cr 6 cr 1 cr Including improved producer intentions accuracy as explanatory variable always Including producer intentions as explanatory turning point accuracy 85% of time variable improved cr = crops. though not universally. Dobson and Pamperin (1972) found that the first USDA crop production forecast of the year was better than naive 70% of the time. consumer durable purchases and political voting shows that intentions can act as good forecasts. Certain conditions must be met: the event is important. Acreage intentions data meet these conditions all too well. Accuracy. better. A value of R between 0 and 2 indicates that the current forecast is an improvement.G. In the most comprehensive study of this kind. 9296) maintains. responses can be obtained. respondents have a plan that they can fulfil. R = (F(t) ~ F(t .l))/(A(t) .
” After this pioneering effort. Lowenstein (1954) reported similar results for cotton. hog prices. a feature that is missing from many later studies. whose specifications for 6 and 12 month ahead price forecasts for hogs also relied on explanatory variables known at the time of the forecast. This section describes the dominant causal modeling approach. since the decision to sell or retain potential breeding stock can be made at several points in the year. methods that did not use intentions data explained only 20-50% of the variation. P* is expected price and t is the time period. Production response is frequently disaggregated into a two-equation recursive system. Trapp (1981) showed that intentions data forecast beef marketings more accurately than do objective indexes of growth rate. The ‘empirical formula’ approach appeared to be more accurate. based on six forecasts. Later. if one exists . presented the first econometric forecast for an agricultural commodity [Moore (1917)]. Trend analysis was a rarely reported but probably widely used forecasting method. 4. chapter 31. . dynamic supply response became the main line of time-related single equation work. about 60-80% of the actual variation in production. However. Early history Henry Moore. coarse grains and soybeans [Surls and of these Gajewski (1990)] and of production crops plus late potatoes [Gunnelson et al. crop response models are typically annual. G. Livestock intentions might appear to be more useful. Smith (1925). Ezekiel (1927) compared short-run price forecasts (l-6 months ahead) from this ‘empirical formula’ approach with forecasts based on the survey indicators described in section 3. Use of a futures price.1. These studies are summarized in the lower panel of Table 1. Other crop and input prices (or indexes) often appear as explanatory variables.P. methods. Ezekiel also recognized the value of combining. . Allen I InternationalJournal of Forecasting 10 (1994) 81-135 91 use of intentions data explained Generally. In the simplest model. Crop and livestock production and price Because most crops have an annual growing season. . noting (p. weights or inventory numbers. cattle prices). though a footnote to the paper hinted that a bigger analysis might reverse the findings. eventually the most satisfactory results may be obtained by some combination of the . His regressions of cotton yield on rainfall and temperature in selected months made better forecasts than USDA forecasts based on condition reports. 4.2. Ladd and Kongtong (1979) reached similar conclusions for within-sample predictions on six crops. agricultural statisticians estimated several true single equation forecasting models [Sarle (1925). first of acreage response. hog prices. One of the few efforts of pure forecasting was by Cox and Luby (1956). They reported average errors (probably corresponding to MAPE) of 8. All but nine of the 48 forecasts indicated the correct direction of price movement.4. (1972)]. 29) “ . cotton acreage. Errors in USDA crop forecasts decreased with each monthly revision of production and exports of wheat. The generic supply response model is QT =f(f’:) > where Q* is anticipated output. These models specified lagged explanatory variables whose values were known at the time of making the forecast. 4. These results support the advice given in the forecasting literature that it is best to use the most recent data.1% to 9.3% for 16 annual and 32 semi-annual within-sample forecasts. . Its history is almost entirely one of explanation and policy analysis [Nerlove (1958). then yield response. Ezekiel (1927). farmers’ price expectation is assumed to correspond to the naive no change model and P* is replaced by lagged price. Hopkins (1927). widely recognized as the founder of statistical economics. Single equation econometric forecasts Regression analysis and deterministic trend analysis share the common origin of correlation analysis.
Fifteen studies have examined production.T. with a few on lamb. Almost all of the studies that test forecasting performance do so within-sample. Allen I International Journal of Forecasting II) (1994) 81-13 for the commodity. though not consistently. Typically. By a suitable transformation. This will typically contain contemporaneous variables for income and other commodity prices. [See Askari and Cummings (1977) for an extensive review. The adaptive expectations model corresponds to simple exponential smoothing of observed prices. Naive price expectations were used initially to explain the existence of hog and beef cycles. -P. aside from the earliest studies cited in section 4. the technical rigidities model. there have been 39 studies since 1952. Price and quantity are usually assumed to be recursive in agriculture._. Since livestock production is year round. Since 1964. based on a simultaneous system. The single equation econometric approach has been popular. a number of weaknesses are evident that limit their usefulness. studies soon came to use quarterly or monthly data series and different methods of describing seasonal and cyclical patterns were employed.) o<Ly < 1. Livestock production has been modeled by the same partial adjustment as described for crops. This statistic is slightly misleading. The final result is that output is a function of price lagged one period and output lagged one and two periods. output is a function of lagged price and lagged output. expected price is the previous expected price plus a fraction ((Y) of the previous error in expectation. If the studies summarized can be taken to be representative of the state of the art. Most official government forecasts are. Evaluation Institutional forecasters produce many forecasts. For price forecasting. production. Slightly more sophisticated is the adaptive expectations model for price developed by Cagan and Nerlove [Nerlove (1958)]: PT . A common alternative is to estimate a single reduced form equation for price. a demand equation is added. though simultaneous specifications exist. although. more accurate than single equation forecasts. though the published record contains insufficient information to state what proportion of forecasts are produced from this approach. Since explanation or policy analysis was the usual purpose of any study. econometricians ignored the need to first forecast contemporaneous variables before the price equation could be used in prediction. 13 studies used single equation methods to forecast 17 quantities (acres. Tables 2 and 3 summarize all single equation forecasting studies located. Similarly.] While such equations could readily be used for one step ahead output forecasting. 20 investigated price and four. even though these would be unknown in a real forecasting . since it was published before the forecast date.1. actual values of explanatory variables are used (ex post forecasting). Equivalently. Over half (46 of 85 specifications) require ancillary forecasts of contemporaneous independent variables to make forecasts of the dependent variable.G. Cape1 (1968) actually produced a forecast of Canadian wheat acreage. yield and export quantities) and eight prices of crops. Livestock production and more especially prices have been popular subjects for single equation econometric studies. both.Pl. has occasionally been tried. Some algebra shows that expected price is also an exponentially decaying function of past prices. The production studies were almost equally divided between hog and beef production. in any case. = cl(P. 4. the consensus of a committee [Newell and Warrington (1962)]. but far fewer reports detailing the methods used. in contrast to crop production. milk and wool production.3.02 P. he could report no comparison with actual acreage. since a proportion of these studies only require standard macroeconomic forecasts such as disposable income or a consumer price index. this was rarely done and even more rarely published. On the livestock side. output can be expressed as previous output plus a partial adjustment of the difference between anticipated and previous output. L’Esperance (1964) found forecasts from a reduced form to be slightly.
241. quarterly. Allen I International Journal of Forecasting 10 (1994) 81-135 Table 2 Single equation Type of series 93 econometric studies. when corn production and personal income both needed to be forecast. The studies represent the state of the art in single equation econometric specification and reveal its present limitations. 49/53 means that four series had known explanatory variables for one step ahead that were unknown for several steps ahead. 1952 onwards No. Though logically defensible. Fcst. bimonthly. annual. daily. of series which Q 11 19 2 32 B _ _ _ 1 1 M D* _ 2 2 1 5 Curr. by date. _ 3 10 24 5 42 No. 38 livestock studies. *Data frequency: A. Comp. var. D. 18 7 7 4 31135 6110 25 Fcst 17 3 7 8 49 16 33 Comp. *Data frequency: A. of series which Q 1 _ _ 1 21 9 22 B _ _ 1 1 M 2 _ _ 2 12 3 10 D* _ _ 5 2 3 Curr. as will be shown in section 7.: number of series that compare forecasts with other non-econometric methods. See Appendix A for details.: number of series that compare forecasts with other non-econometric methods. In short-term forecasting. the standard error of forecast for 12 months ahead corn price increased by two and a half times. D. Q.P. Comp.: number of series with contemporaneous explanatory variables. . the most likely cause of poor performance is insufficient attention to dynamic specification. Curr. Q.: number of series that produce forecasts after the estimation period. one of six studies Table 3 Single equation Type of series on production forecasting and 12 of 16 studies on price forecasting were comparative. S. McNown (1986)]. when explanatory variables needed to be forecast. var. In an article that demonstrated a largely ignored point. Fcst. M. monthly. var. context. 31135 means that four series had known explanatory variables for one step ahead that were unknown for several steps ahead. M. daily. bimonthly.: number of series with contemporaneous explanatory variables. monthly. 1 both. Fox (1953) observed that error of forecast was greater in ex ante forecasting. by commodity.: number of series that produce forecasts after the estimation period. semi-annual. O/l 3 17118 27129 2 49153 Fcst _ 9 19 33 5 66 Comp. B. mostly with time series methods. 11 1 7 4 31 8 23 No.G.2. of series Data frequency A S 1 2 _ 3 1950s 1960s 1970s 1980s 199osl Total 2 10 29 38 6 85 1 6 7 16 _ 30 _ 9 3 2 14 ‘Different ending dates depending on source. the finding contrasts markedly with the empirical evidence from macroeconomic forecasting [Armstrong (1985) p. In particular situations. semi-annual. S. Data frequency A S 3 1 2 1952 onwards No. annual. Crop production forecasts can reasonably be based on annual series. var. Later studies have presented single equation econometric models mainly in comparison with other techniques. expl. quarterly. B. Since 1980. for example where seasonal prices of stored crops differ from the post-harvest price by only econometric studies. expl. it performs poorly against time series methods. In his example. Curr. of series All crops’ Acres/yield Export quant Price All livestock Production Price 25 10 7 8 60 22 38 22 10 7 5 8 6 2 ‘12 crop studies. Since vector autoregression is shown to be much more accurate.
the surge of interest in forecasting occurred in the 197Os. especially for prices. Agricultural series appear to accommodate fixed parameter models better than do non-agricultural series. as has been admitted by one of the few studies to put forecasting in a decision making context [Brandt (1985)]. Econometric models 5. When the commodity is storable.5 agricultural commodities and 13 non-agricultural commodities or services. even when monthly or more frequent data are available. lo-year data sets were best and 20-year data sets worst. the multisector system becomes large enough to qualify as a large scale model. wheat for human food or animal feed. to a much lesser extent. There are a large number of sectoral models in the agricultural economics literature. separate from demand for consumption. All of these complications can usually be accommodated by a ten equation system. trade can be handled in a spatial equilibrium programming model. But with such limited information. it was common in the early studies to treat the equation as a recursive system and so justify the use of ordinary least squares estimation. at which point a market clearing identity will also be needed. except where many regions are being analysed.G. 5. Sometimes. which must themselves be forecast. Alternatively. Stand-alone large scale models and those linked with large scale macromodels of the economy attempt to endogenize all variables. for example. eggs for consumption in shell or for breaking. as described in the next section. They are equipped to forecast. In each case. For non-agricultural series. Trade between countries or regions adds import supply and export demand equations. as both Canada and the US and. particularly in crop The development of sectoral models slightly preceded that of large scale econometric models. agricultural series of 20 years gave better results than did series of 15 or 10 years. At both 5-year and lo-year horizons. HildrethHouck. Allen I International Journal of Forecasting 10 (lW4) 81-13s the storage cost. sectors are so intimately linked that a multisector model is called for. 15 and 20 years ahead forecasts. an inventory demand equation is required. but at the cost of complexity. Given the lag between decision making and output. Conway. nothing definitive can be claimed for the forecasting ability of varying versus fixed parameter econometric approaches.1. At some point. Sometimes there are demands with significantly different characteristics.94 P. The process was repeated for 15year and 20year periods. Hrubovcak and LeBlanc (1990) tested six kinds of stochastically varying parameter (SVP) specifications (Swamy-Tinsley. annual crop price forecasts can be useful. For 10. Each series was broken into a succession of mutually exclusive sets of length 10 years. The greater proportion are concerned with explanation or policy analysis. production. The commonest example concerns the livestock and feed grains sectors. A further complication with livestock is that the inventory can be used as investment for further production or can be sold. Recent livestock studies seem to have settled on quarterly forecasts as being the most appropriate. or they are incorporated through contemporaneous explanatory variables. As noted from Table 4. with naive no change next and fixed parameter models last. Sectoral models A sectoral model contains. In a path-breaking study. net capital investment in agriculture was most accurately forecast by SVP (with the notable exception of Cooley-Prescott). The problem in forecasting with a sectoral model is either that linkages with the rest of the economy are ignored. We lack information on the usefulness of single equation econometric forecasts at longer horizons. Kalman filter and Cooley-Prescott) against six others (autoregressive and fixed parameter models) and the naive no change model. Gold (1974) compared annual series for 1. Gold made forecasts to different horizons. a supply equation and a demand equation for a single commodity. Australia struggled to bring a set of . at a minimum. The usefulness in decision making of such aggregate forecasts is doubtful.
Table 5 Data frequency Type of model in sector and aggregate studies Data A frequency S Q M No. S includes models with some A equations.P. though Bliemel (1973) had already done so elsewhere). although. Theil’s U statistic was reported irregularly (frequently the incorrect U. the test has low power. The process was started. Where forecasts were made beyond the end of the estimation period they were short. monthly. typically covering four quarters. Q includes models with some A and S equations. sectoral models together into a comprehensive system [Agriculture Canada (1978). the model was regarded as suitable for use in forecasting. . ‘Numbers in parentheses There are 51 studies using single equation methods and 53 sectoral studies. I 6 85 ‘Different ending dates. the hog sector. The most popular assessment technique was validation by dynamic simulation within the sample used for estimation. Salathe et al. allows a comparison with the naive no change model. A handful. readers of the agricultural economics journals were unaware of the consequences of the different ways of computing U until Leuthold (1975) pointed them out. Of the 60 sector models summarized in Tables 4 and 5. S. somewhere in the estimation period. M. B. ten. (1982). of which 12 concern the poultry sector. Practically all the models contain contemporaneous exogenous variables. the beef or beef-feed grain sector and nine. See Appendix A for details.G. reported the values of exogenous variables used to make ex ante forecasts. . Allen Table 4 Distribution Time interval 1950-59 1960-69 1970-79 1980-89 1990-91 Total I International Journal of Forecasting 10 (1994) 81-135 95 of forecasting studies over time Single equation 2 10 29 38 (2Y (12) (34) (45) (7) (100) Single sector 2 9 34 14 1 60 (3) (15) (57) (23) (2) (100) are percentages. with the typical sample of size four. of series Single sector models Crop Livestock Mist Total Multisector or aggregate models 24 16 8 15 43 2 60 10 15 2 2 20 1 23 3 6 1 10 25 2 Data frequency: A. the majority (43 studies) are livestock models. (1980)]. Q. As long as the dependent variables predicted in this way gave reasonable forecast accuracy and turning point statistics. M includes models with some A and Q equations (five of 60 sector models are mixed). These Multi-sector 1 4 9 8 2 24 are number (4) (17) (38) (33) (8) (100) of models. Theil’s U. annual. Kingma et al. bimonthly. quarterly. semi-annual. The structural system was then solved for each successive time interval using calculated lagged endogenous and actual exogenous variables. with actual values of lagged endogenous variables. usually those relying on one or two macroeconomic variables and aggregate indexes.
Reconciliation of direct forecasts of total world supply or consumption and the sum of the individual country forecasts would require further judgment. Other approaches Ashby (1964) described a balance sheet approach that could easily be worked on a spreadsheet. Alternatively. MacAulay (1978). . supply and demand functions for each region must be updated and the spatial equilibrium found. Spatial equilibrium or interregional competition models have one advantage over econometric methods in multiregional analyses.3. when defining transport cost.2. Of competing approaches. Modeling change as a Markov process has occasionally been suggested. Martin and Zwart (1975) specified supply in each region as a function of lagged price in that region. the supply for a region could be a function of commodity price and a time trend variable. Johnson and Carter (1963) provided an early example. Dean. Dairy industry structure. Its typical use is to forecast the number of businesses of different sizes. in which enterprise budgets and a detailed specification of the structure of the beef producing sector would be used to forecast industry changes. The author provided no details 5. Crom (1975) described a systems approach. was similarly predicted in Canada [Furniss and Gustafsson (1968)] and in Britain [Colman (1967)]. being developed by the USDA for the beef sector. Countries with limited data would need to have their supplies forecast judgmentally. Colman compared the first post-sample prediction with actual census data. They had a surge of popularity in the mid 1970s especially in Canada [e. The differential can consistently exceed the cost of shipping goods between the regions. more recently.G. The calculated price in each region then updated the quantity supplied in the next time period. Quantity supplied in each region was fixed at the start of calculations. They can distribute the total quantity of a commodity among regions in an internally consistent manner. using the census data for 1950. For example. 5. nor do they offer much of a guide to forecasting ability. measured as the distribution of dairy farm sizes. In countries with good data series and existing quantitative models. Quantities produced and consumed in each region are normally specified as functions of commodity price. For accuracy. The balance sheet layout ensured internal consistency. Programming approaches Programming models are not usually thought of as being suitable for either explanation or forecasting. For forecasting. His example was a forecast of the world sugar market. No study was found that compared the forecast error of the Markovian transition probability approach with the error of other methods. The approach consists of collecting forecasts for different regions and for aggregates by whatever means available. Ashby’s study was the only example of the method discovered. only judgmental and programming methods could work with such a short series (minimally. The market clearing identity is raised to a new position of prominence. these could be used. Additional constraints prevent prices in different regions exceeding the cost of transporting the commodity between them. The matrix raised to successively higher powers provides a forecast of changes over time. close attention needs to be paid to historical price differentials between regions where trade has occurred. the time variable is collapsed into the intercept term. The first step is to estimate a transition probability matrix based on historical data.g. The functions are estimated econometrically by separate equations. as.96 P. the supply function can be dynamic. did Edwards. two periods). 1955 and 1960 to predict the size distribution of California cotton farms to 1975. Spatial equilibrium models also calculate the commodity price in each region. As a simple example. In any period. Smith and Patterson (1985). Allen I International Journal of Forecasting 10 (1994) 81-135 But such nonstochastic simulations do not adequately test dynamic specifications [Shapiro (1973)]. Martin and Zwart (1975)].
Chen and Bessler (1990). Models are either abandoned or revised in such a way that a long series of forecasts is unavailable. Ordinary least squares (OLS) is often as good as methods developed to deal with simultaneous equations bias.0025. They were developed in departments or by teams separate from those responsible for the outlook reports. Since only one official forecast is released. 5. While such models are being used by government agencies as aids to producing official forecasts. where the general belief is that they would do better. and the relation between validation results and forecast performance is unknown. The institutionalization of formal quantitative models appears to have been a struggle. the probability of 10 or less successes is P = 0. the raw forecasts are not reported. Cluff (1990) described the process at Agriculture Canada. Most of the systems contain from two to four equations. Examples include Egbert (1969). The formal modelers and the commodity analysts produced different forecasts. this result is unsurprising. stand-alone models. In year by year comparisons. the largest is the 67 equation cotton sector model of Chen and Bessler. Soliman (1971) found that three stage least squares produced the best forecasts (measured by Theil’s U) in two equations and OLS in the other two.) For one step ahead comparisons (which are the vast majority).2 where comparisons are discussed in more detail. Evaluation The forecasting performance of sectoral models remains an unknown quantity. highly aggregated. Most comparisons of sectoral models.4. Fanchon and Wendell (1992)]. two stage least squares was best in 2 years and OLS in 1. They used estimates of elasticities and rates of growth and inflation from other studies. nor on how they would actually be used to make forecasts. First generation models treat agriculture as a separate entity and often fail to link factor demands with output. Vere and Griffith (1990). Structural sectoral models were more accurate in only ten of 38 pairwise comparisons with other forecasting approaches. including those few where forecasts are made. to maintain and update it and to train a new group of people in its use were typically underestimated [Hedley and Huff (1985)]. though it is likely to be short. (1982). concern the relative performance of different econometric estimators. the different values produced within an agency must be reconciled. then dropped. Their purpose was projection of agricultural output and prices under different . The researchers who developed the models were not responsible for their maintenance and regular use. Rausser and de Gorter (1982) distinguish three classes or generations of agricultural industry models. Using the broadest definition of sectoral model. What is surprising is that sectoral models have not been compared at more distant horizons. We are left to rely on the validation process. six studies have been located that compare econometric sectoral model forecasts with other methods [Leuthold and Hartman (1981). but there are too few studies to draw any conclusions. 6. These constitute about half of the methods labelled ‘other multivariate’ in section 7. Kulshreshtha et al. Allen I International Journal of Forecasting 10 (1994) 81-135 91 on how the budgets would be updated. Using simulated data with known autocorrelated structure. (19X9). (If the models had been as accurate as the other member of the pair. Aggregate models and large scale econometric Penson and Hughes (1979) and Freebairn. These were small. Nor is the typical life of such sectoral models known. Early models apparently were updated once or twice.G. Quance and Tweeten (1972) and Yeh (1976). Park et al. Naik and Dixon (1986) found that OLS was most accurate within-sample but two-stage least squares and reduced form with autocorrelation corrected (by Durbin’s method) were better when forecasting.P. The expense and effort required to develop a forecasting model.
one of the potential policies might have been identified as ‘most likely’. only with current No current exog. + studies Macro exog.G. if necessary to remove inconsistencies. It had disposable personal income and the US general price level as exogenous variables. vars. Cromarty’s model could be solved after Klein-Goldberger to produce forecasts of output and prices for 12 agricultural commodities or commodity groups. exogenous No exog. Typically.g. The models of the macroeconomy constructed by the principal business forecasting units such as Chase Econometrics. ‘Only one after 1979. Microeconomic typically available from forecasting services. the macromodel is first used to forecast a set of variables exogenous to the agricultural sector. if linked variabies and number that perform forecasting and testing Validate by dynamic simulation? Yes No Make post-sample forecasts No Yi?S Tests?” Yes No Single Multi or Aggr. 5 and 6 summarize the features of multisector and large scale agricultural forecasting models. Data Resources Inc. however. Rausser and de Gorter (1982) simply characterize them as having better linkages between the domestic macroeconomy and the international economy or the agricultural sector. . interest rates and the consumer price index. vars. In second generation models. 43 17 6 2 5 3 3 2 19 11 38h II” 9 10 51 14 33 7 18’ 7 “Tests? refers to summary statistics of forecast accuracy and turning point performance in the post-sample period. formal models are designed from the top down (section 6. They may be informal in the sense that several models each of a single sector or interrelated sectors (such as feedlivestock) are examined in concert and re-estimated. such as personal disposable income. The agricultural components of these models are usually small to non-existent. are then transmitted back to the macromodel. population and consumer price index. vars. Macroeconomic exogenous variables are those e. such as total agricultural output. Within the first generation. endogenous in the Klein-Goldberger model of the US economy. These variables were. hThree studies provide insufficient evidence on whether or not validation was done. Kost (1981) briefly surveyed these and the individual country models of project LINK. In this situation. Tables 4. The tables also show comparisons and aggregate Macro micro. Forecasting was incidental. Examples Table 6 Number of sector Type of model include Chen (1977) and Roop and Zeitner (1977). vars. The models often fail to include explicit variables to represent sector policies. Penson and Hughes (1979) describe third generation models as having direct or indirect accounting of capital accumulation and financing. exog.98 P. The linkage is incomplete. Large scale models with from 30 to several hundred variables are intended to describe multiple sectors of the economy. a second group of studies could be characterized as large scale multisector models. An exception is Cromarty (1959). More commonly. essentially larger versions of the sector models described in section 5 [including Maki (1963) and Crom and Maki (1965a)l.1). and the Wharton Economic Forecasting Unit are the most familiar. Allen i I~~ernariona~ Journal of Forecasting 10 (2994) 81-135 policy proposals.2). generally acknowledged to be the first large scale econometric model of agriculture. a large scale representation is built from the bottom up (see section 6. Solution values. ‘TWO studies provide insufficient evidence on whether or not validation was done. such as acreage diversions or deficiency payments. variables such as capital accumulation are not transmitted back. variables would need to be forecast specifically for the study. while Freebairn. These forecasted variables are used to solve the agricultural system. personal income. A model designed to be linked to a system may have exogenous variables that are endogenous to the system.
And their creators seem reluctant to test their model’s forecasting performance. In terms of data frequency.2. Individual researchers had failed to consult on variable definitions and data sources. was named FAPSIM. agencies followed a bottom-up approach. though useful for policy analysis. It describes ten of the 13 structural commodity models under development by. demand for these commodities was a derived demand from the livestock producing sectors. and multisectoral models excessively. or the time series data access method [Bell et al. The new model. In long-term projections and policy analysis. price indexes and government outlays. Maki’s (1963) 44 equation model and the Crom and Maki (1965a) 30 equation model of the beef-pork industries were the first studies on interrelated sectors. Agriculture Canada. a use to which the larger models have rarely been put.P. both types of models typically require forecasts of exogenous variables. T-DAM. personal communication.G. the CCFS was updated. sectoral models substantially. (1976)]. Crom (1970) listed the 128 operating rules de- . the USDA uses results from a set of econometric models. In 1980-1981. since about 70% of corn. from single to multiple sectors. enlarged and given a policy analysis orientation. while most of the succeeding models were annual (see Table 5). so the individual models had to be redefined and reestimated. the Food and Agricultural Policy Simulator [Salathe et al. several milk products. 6. The persistence of the problem led to the construction in the late 1970s of a common database. The latter model was simulated within-sample to derive operating rules. Informally linked commodity sector models The philosophies of the agencies responsible for agricultural forecasting in Australia. Initially. The ultimate goal of the bottom-up approach is either a single forecasting model for all commodities or a set of consistent sectoral models with formal linkages. Publication of both sectoral and multisectoral models peaked in the 1970s. now with 360 equations. The mechanical projections from the models are moderated by judgments from a committee of analysts (Paul Westcott. with other sectors still to be added [Boutwell et al. Attempts to link sectors frequently encountered the problem of variable incompatibility. barley-grain sorghum and oats produced in the US went for animal feed. Development of separate commodity models was piecemeal.2. The development of formal quantitative models also follows similar paths and timing. often overlapping. (1978)]. the annual linked crop-livestock model known as the cross commodity forecasting system (CCFS) was made operational. Multisector models Two groups of researchers began work on the livestock-feed sectors in 1965 [Crom and Maki (1965a). respecified. They were unusual in being semi-annual models. From the early 197Os. Several of the sectors were being operated separately rather than being linked in a consistent manner. Canada and the US appear remarkably similar. or on behalf of. such as the changing of a coefficient or an equation specification when a particular variable reaches an extreme value [Crom and Maki (1965b)]. Formally (comprehensively) linked models 6. The more recent surge in single equation models is accounted for by their use in forecasting comparisons. building gradually more complex econometric or programming models. And. (1982)].1. In the 1970s the Economic Research Service of the USDA began a two-step process of developing individual sector models which were then linked together. Egbert and Reutlinger (1965)].1. 1993). They realized that there were many linkages among the prices and quantities consumed of the different meats. 6. Equations were added for cotton. At the same time. For forecasting purposes. it consisted of 133 equations for nine livestock and crop sectors. Allen I International Journal of Forecasting 10 (1994) 81-135 99 with models described earlier. Perhaps the most comprehensive description of the process is a two volume report that appeared in 1978 [Agriculture Canada (1978)]. rely on annual data-too long an interval for most private decision making purposes.
systems include the accountingtype POLYSIM [Ray and Moriak (1976)]. Improved specification and model testing are closely linked. also contains reaction functions that endogenize policy. with detailed regional and commodity breakdowns. much of the work with agricultural models was concerned with structural specification. researchers in the US had made little progress in improving the forecasting performance of international 6. at different levels in different nations. Chen (1977). Questions of usefulness and comparative performance of large scale models were rarely addressed. small policy. meant that price forecasts in nominal terms were impossible. Because the agricultural sector is a small part of the total economy. agricultural prices in the US rose dramatically. mentioned earlier. Before that time. There are four geographically distinct multiproduct agricultural industries and four type-of-farming industries for a total of ten commodities. Rapid inflation. Only occasionally did forecasting appear to be a goal.G. Linked agricultural-macro models The model of Cromarty (1959). With increasing computing power and longer post-war data series. “We are in the infancy stage of estimating the economic interrelationships among agricultural commodities”. (1982)] is a structurally detailed computable general equilibrium model that is unusual in its detailed treatment of agriculture. The Freebairn. surprising everyone and dismaying forecasters. Evaluation of large scale models In 1961. The improved system was used for forecasting. Around 1972-1973. at least in print. The ORANI model of the Australian economy [Dixon et al. A few third generation models have appeared [e. Changes in US farm combined with rapid inflation. estimation techniques and policy simulation. Cromarty could observe (1961. These are annual models with limited application to short-term forecasting. the econometrically based TECHSIM [Collins and Taylor (1983)] and its successor AGSIM [Taylor (1990)]. plus noncompeting imports for a grand total of 114 . p. in reference to both sectoral and large scale models. agricultural economists seized on the excitement of more comprehensive modeling of agriculture. 365). As well as FAPSIM. The entire model has 103 other sectors. sectors [Higgs (1986)]. converted the US farm economy from a relatively closed system to a relatively open one.2. Although the early developers of large scale macroeconomic models regarded forecasting as a major objective. The price shocks of 1972-1973 marked a watershed in the development of international commodity models.75 veloped from the simulations. is generally reckoned to be the first large scale agriculturally oriented linked model. Allen I international Journal of Forecasting 10 (1994) 81-1. He also updated the model and showed graphically how the operating rules improved within-sample performance. The USDA has used a variety of large scale model systems intended to simulate various policy options.2. crops and rising export demand. failure to feedback sector solutions to the main model has little impact. (1991)]. Rausser and de Gorter (1982) model. Forecasters concluded that their models were inadequate because they failed to consider the impact of international markets on US prices [Rausser (1982)]. Sometimes formal testing is not required.100 P. and exports are variables can be and exogenous. Imports determined endogenously and reclassified between endogenous The model has been used for although forecasting is possible. in addition to linkages with the macroeconomy and international markets.3. but no tests of forecasting performance were made. described earlier. Penson et al. policy analysis. Roop and Zeitner (1977) and Chan (1981) demonstrate second generation models where the agricultural sectors are essentially addon components to large macromodels and are solved sequentially. (1984)]. Another example is the Food and Agricultural Policy Institute’s (FAPRI) policy-oriented econometric model [Brandt et al.g. Even less often was anything more than rudimentary testing of forecasting performance considered. 6.
Fildes (1985) recommended that models be tested for ex ante performance relative to extrapolative or judgmental alternatives. The recent documentation of farm income forecasting is a first step [Dubman et al. Just (1993) in his ‘conclusions and a call for action’ asks for a reduction in the emphasis on standard statistical concepts of fit. Improvement is still needed. 581. Shapiro (1973. speaking from the viewpoint of industry. of the few modeling exercises that listed forecasting as an objective. Greater dependence on world markets had forced Canadian researchers to address the issue of international impacts. actual use of the methods does not appear to have greatly improved. Thompson and Abbott (1982) noted that. Single and multiple equation econometric models. While economic theory is some guide to specification. In the context of agricultural trade modeling. Outlook forecasting at the USDA is a complex and diffuse process without an overriding formal approach [see Bell et al. commercial forecasters and academics [Rausser (1982)]. While the last two decades have seen many new model test procedures appear. 37) “[tlhe crucial criterion for forward-looking analysis is the ability to represent out-of-sample phenomena”. The comment is still true. against which possible improvements can be tested. a series of conferences was held among representatives from the USDA.G. supply of a commodity is a function of commodity price and input prices. Agriculture Canada claims that its Food and Agriculture Resource Model . Cromarty and Myers (1975).P. Between 1976 and 1980. As Table 5 shows. questioned the usefulness of annual price forecasts. (1978)]. They found much to be dissatisfied with. (1991) p. On the other hand. it is impossible to take all previous work into account. forecasters place relatively more emphasis on econometric models (Paul Westcott. For example. 255) noted that almost all model evaluation procedures to date have employed non-stochastic simulation (with respect to both equation error term and the sampling distribution of the estimated parameters) to generate the experimental parameters . These are among the criticisms that Fildes (1985) levelled at quantitative forecasters in general.a procedure which is inadequate in testing dynamic theories. about two thirds of aggregate and large scale models are annual. Allen I International Journal of Forecasting 10 (1994) 81-135 101 commodity models [Labys (1975)]. For longer horizons. various commentators began to assess the progress that had been made. agricultural forecasters continue to suffer from two major handicaps: lack of a common maintained model and inadequate test protocols. (1993)]. They also found the complex simultaneous equation models of little use in short-term decision making. Competition forced commercial forecasters to improve the performance of their product. statistical analyses of survey data and analyst judgment are combined to produce the official forecasts. they found that an endogenous variable could be better determined within a sub-system rather than simultaneously within a system. 76. Agriculture Canada. A series of US General Accounting Office reports has recommended that the USDA document its forecasting procedures [US General Accounting Office (1988) p. Table 6 shows that most large scale models require forecasts of exogenous variables. He notes (p. both in agricultural sector models and in large scale models in general. In most cases. personal communication. p. But what about competing products? What about constraints on rate of adjustment to price changes? There is no reference specification. a view held by leading forecasters since the late 1970s. Soon after sectoral and large scale models became popular. almost none provided any forecasting measures outside the range of the data used to estimate the model. Stochastic simulation would not address the predictive performance where it depends on contemporaneous exogenous variables. The USDA lagged behind. it is silent on many practical issues. 1993). Without the feedback from striking events. Because there is no accepted cumulation of past research. although in this respect they are relatively better than sectoral models. especially dynamics. for example for corn supply.
if necessary. In three studies [Roop and Zeitner (1977). Results within-sample were much better. (1977). This process maintains internal consistency while achieving forecasts that are judged acceptable. There is every reason to believe that these findings are typical of large scale models. His 1984 article [Bessler (1984a)] provided a good explanation of the basic approach without dwelling on the overparameterization problem. Tiao’s exclusion of variables method. the existence of a cycle of about 3. 7. the era of time series analysis began in 1970 with the appearance of an article illustrating the Box-Jenkins and exponential smoothing methods [Schmitz and Watts (1970)].1. Again. In the 198Os. once a policy was enacted. US.4 years. by reporting proper post-sample forecasts the article set a standard that was not followed for many years. Articles on spectral analysis began to appear at the same time [Rausser and Cargill (1970). Hinchy (1978). Exponential smoothing produced the more accurate forecasts. Australia has had an aggregate recursive program available since 1976 and the model has been used for 5 year projections [Kingma et al. exponential smoothing has practically never since been used for agricultural forecasting. (1980)]. not for forecasting. Weiss (1970). For agricultural economists in the US. . Although intended as a demonstration. In contrast to business forecasting practice. by reverse simulation [Cluff (1990)]. But can their conclusions be trusted? Models would have greater credibility if. interest in multivariate time series analysis became evident. In the 1960s interest in the hog cycle led to the use of harmonic analysis. statistic was greater than one in more than half of the variables forecast in the post-estimation period (79 of 161). A demonstration of the use of multivariate cross spectral analysis followed [Ahlund et al. general priors. symmetric and non-symmetric random walk Bessler and Kling (1986). the forecasts produced by analysts were conditioned on the given policy variables. After the results have been reviewed. Time series models Deterministic trend extrapolation was an early form of time series analysis. the model is calibrated. Just and Rausser (1981) found that futures prices were more accurate predictors of eight agricultural commodity prices one to three quarters ahead than were four major private forecasters and the USDA (in that order). prices. A series of articles introduced various parameter reduction methods [Brandt and Bessler (1984). Shonkwiler and Spreen (1982) used a transfer function to analyse the much studied relation between the number of hogs slaughtered in the US and the hog-corn price ratio. Few comparative studies of the forecasting performance of large scale models have been located. broiler cycles. world cocoa Cargill and Rausser (1972). Allen I International Journal of Forecasting 10 (1994) 81-135 (FARM) is “one of the few being operated on a continuing basis within a successful outlook program” [Hedley and Huff (1985)]. At about this time. Use of time series methods in agricultural forecasting Jarrett’s (1965) forecast of Australian wool prices using exponential smoothing marked the first application of modern time series methods to agriculture. beef price]. Stillman (1985). Bessler and Hopkins (1986). lead-lag relation between export and Australian saleyard prices of beef].G. their interest was to confirm what had already been shown by spectral analysis. Do large scale models have a use? Their makers would argue that they are intended for policy simulations. futures prices. It is updated each quarter and used to produce forecasts for up to 6 years ahead. Theil’s U. probably widely used in institutional forecasting though little reported in the literature. 7.102 P. Westcott and Hull (1985)]. in which production and price of hogs functions [Larson were modeled as cosine (1964)]. Bessler introduced vector autoregression (VAR) to the agricultural economics profession. The intent was to explain historical data patterns rather than to forecast.
1).28 0.67 0.60 0.63 0.00 0.2). a forecast of a peak (a rise followed by a fall in the series) will be counted as correct when the actual series displays a trough (a fall followed by a rise).73 The upper triangular matrix is a better than/worse than set of pairwise comparisons.43 0.3).0). Accuracy After the pioneering (1970) no comparisons performance multvar. Bessler also pioneered the commendable practice of providing comparisons among methods based on post-sample forecasts.64 0. The example is illustrated in detail in Table 7a.00 0. For example.30 0. the first entry is interpreted as ‘the row entry method (Naive) is better than the column entry method (ARIMA) in 15 pairwise comparisons and worse than ARIMA 23 times” (in a total of 38 comparisons from various studies).65 26 35 9 29 35 0 21 0. The lower triangular matrix is the success rate. 0. (2) econometric. (1.04 0. Turning point measures have fascinated agricultural economists.11 0. (1. Allen I International Journal of Forecasting 10 (1994) 81-135 103 Bayesian priors].10 0. Naik Table 7 Success rates and Leuthold (1986) overcame the problem with a 4 x 4 contingency table that distinguished between peak and trough turning points. Kaylen (1988) provided a review of parameter reduction methods.2.39 0. Better usually means lower one step ahead forecast root mean square error. (4) different econometric.36 23 10 17 8 0 35 8 0 0 0. . the distinction is unnecessary since actual data will reveal whether the series is rising (with a potential peak about to occur) or falling. In terms of (x. or proportion of total pairwise comparison in which the column entry method is better than the row entry method. The interpretation is inverted.1).33 9 15 0 I1 4 0 6 0. including both exclusion of variables and Bayesian techniques. study of Leuthold et al. For example.00 0.3).00 0.39. though a forecaster might be concerned about the error rate of predicting peaks compared with that of predicting troughs.56 1 20 16 10 1 7 12 0 12 6 22 6 7 0 8 3 12 0. (4. The value is found by dividing the number of times with better than comparisons by the total number of comparisons for each pair of methods. Each member of a group is compared with each method not in the group but not with other members of the group.35 0.2).4).30 0.54 23 33 26 0.4) and (2.58 0. (4.46 27 10 24 24 23 13 6 0.54 0.27 0.2). then the pairwise orderings considered are (1. perhaps because microeconomic theory emphasizes direction of effect over quantity of effect. is the success rate of the column entry method (Naive) over the row entry method (ARIMA). y) pairing. (2.40 for the Naive method).2.1. of agricultural forecasts of different ARIMA methods in pairwise univar. 0.00 0. (3) ARIMA.31 23 8 15 14 9 0. The better than/worse than comparison is done as follows. (3.46 0.P. Sometimes results are only presented with RMSE over a range of forecast horizons. Evaluation: comparisons 7. A dash indicates that the methods have never been compared. 7. these are recorded as composite (3. He also introduced a new exclusion of variables approach that is similar to Hsiao’s method but allows for deletion of insignificant intermediate lags.70 0.26 0. Since the standard 2 x 2 contingency table only distinguishes a change in direction from no change. For example.76 0.3). the first entry.51 0. which is 15 divided by 38.59 0. Only for several steps ahead forecasts is the larger table really necessary [Kaylen and Brandt (1988)]. if a study ranks the methods as follows.22 0.43 0. comparisons Expert of ex ante forecasting Econ VAR Other Naive Other Composite Simole Other 1 14 2 17 3 0 0 32 26 33 17 31 20 0 4 54 Total Better 104 162 63 107 71 70 42 187 185 Worse 155 138 142 124 125 55 84 99 69 N A ou EX EC VA OM CS co Success rate 15 0. econometric (1. (3.33 _ 0. Table 7b collects Table 7a information into the form used in Table 7b. In a one step ahead forecast. ARIMA (1.l). Sometimes only MAPE is reported.36 0. (1) composite.3).65 _ _ 0.37 0.G.40 0. The bottom row is the overall success rate for the column entry in all comparisons (for example. (3.
00 (3) 0. More surprising is that complex methods of calculating weights performed relatively better than simple averaging (based on 11 series in seven studies).104 P.1 1. Simple averaging was typically better than some of the composites.o 1. The (root) mean square error ((R)MSE) was used to rank methods. perhaps because it faced relatively weak opposition most . when the study reported it. Instead. For example. and for 40 comparisons. the lower triangle has been used to report the success rate between pairs of methods.4 2.125.1 18 1.2 4. achieving a success rate of 0. Theil’s U statistic or the mean absolute percentage error (MAPE) was used to make the pairwise comparisons. The success of composite or combining methods is a result that conforms with widely held beliefs. In the 5% of studies that failed to present the MSE. Table 7 summarizes the comparative forecasting accuracy of nine methods or groups of methods. The previous statement needs some qualification.33 (3) 2.1 4. the naive method is more accurate in 104 out of a total of 259 pairwise comparisons.y) pairs: the method listed on the left of the row was more accurate than the method listed in the column heading in x comparisons and worse in y others. A notable omission is a comparison of vector autoregression with any form of composite forecast. its success rate is 0.3 3.3 3. 0. though rarely the best.3 1.1 3.0 1. since the number of observations is frequently small.3 0.2 1. while the most popular.76). ARIMA beats other univariate 26 times in 34 pairwise comparisons.l Comp ARIMA Econ 130 1.5 (2) 0. is not the most consistent criterion for choosing the best method [Armstrong and Collopy (1992)].l kl O.2. one of which was simple averaging. the 90% confidence interval for ten comparisons is 0. Next was simple composite (with weights calculated as simple averages).4 B/W 2.3 to Table 7. column j is the proportion of times the method in column j is better than the method in row i in terms of the criterion used in the study. The upper triangle shows (~.0.00 (1) 1. Since ‘other univariate’ beats ARIMA eight times in 34 comparisons. The value in row i.3. Very approximately. there are many gaps. Comparisons between pairs of methods need to be made with caution. and RMSE.l 0. For example. The comparisons are based on forecasts of 129 agricultural series reported in 49 studies.40.00 (2) 1.2 appeared until almost the 1980s. This information can also be expressed in reverse order. Allen I International Journal of Forecasting 10 (1994) 81-135 (Footnotes Table 7a Pairing 1. Even after grouping methods together.76. The final row of Table 7 is the overall success rate of forecast accuracy of the method listed at the column heading against all others.0 1 .25 (4) Composite 1. Because studies typically compare only two or three methods. But the studies provide insufficient cumulation of evidence to favor any particular composite method over simple averaging. so the lower triangle would convey no new information. a success rate of 104/259 or 0.G. The typical study compared three to five composite methods. continued) Table 7b Composite ARIMA Econometric Total B/W 38 1.o 18 ARIMA Econometric O. Symmetric off-diagonal elements of the matrix contain the same information in reverse order.O Overall 0. for 20 comparisons 0. The most accurate method was other composite (with weights calculated adaptively or from ratios of variances or by regression). Table 7 summarizes pairwise comparisons of the different methods.1 3.24 or (1 . VAR was the best single method. Choice of criterion matters.2 1.
the starting direction must be forecast and the forecast may turn out to be incorrect. Again.67. based on ten studies performed between 1974 and 1984. state space or structural equation systems) do slightly worse than the naive no change forecast. but only VAR and ARIMA are markedly better.2. a forecast peak (a move upwards then downwards) where an actual trough occurs. The methods may represent minor variations. for example. Turning points Forecast accuracy was the commonest criterion used in comparing methods. A previous price increase. For one step ahead forecasts the key difference between Brandt and Bessler (1981) and the improved method proposed by Kaylen (1986) is that Kaylen compared forecast change with preceding actual change. There. Conclusions are similar to the forecast accuracy comparisons. But 55 of the 125 pairwise VAR comparisons were with ARIMA models. First. for example. A smaller number of studies reported turning point performance. The ordering of methods is essentially unchanged. the success rate of causal methods in short-term ex ante forecasting was 0. if each comparison had consisted of the same total number of pairs? The answer is found by calculating the simple average of all the success rates for a given method. the single study should lead to the same conclusions as a number of separate studies that use the same series. whereas Brandt and Bessler used the forecast change. Three authors have examined an identical data set using the same seven forecasting methods (Table 9). A study that compares a large number of methods tends to dominate the ratios. Econometric (single equation) and other multivariate methods (transfer function.G. If so. Only other multivariate methods perform worse than naive methods. so that certain forecast possibilities are eliminated. The same hog price series (over almost the same time interval) was used by about ten of the studies. the state space approach in the competition organized by McIntosh and Dorfman (1990) performed rather badly relative to two strong methods: a VAR approach and an alternating conditional expectations approach estimated by Berck and Chalfant (1990). 7. On the other hand. For example. a competition that ARIMA won 64% of the time. (Alternating conditional expectations is a form of regression analysis that uses a smoothing algorithm to find the best transformation of each variable. because it permits more pairwise comparisons. Allen I International Journal of Forecasting 10 (1994) 81-135 105 of the time.2. the preceding actual change is known.50 (equal ability) was arbitrarily assigned where less than ten comparisons existed between a pair of methods. A number of criticisms can be levelled at the results.P. Rates for some pairs are unknown or are based on very few comparisons. But one of the criticisms levelled against it by econometricians is that VAR is atheoretical. Turning point comparisons need to be interpreted cautiously for two reasons. the even smaller number of observations than for the forecast accuracy comparisons makes their ability to rank methods even less powerful. the criterion counts as correct a forecast directional change the reverse from the actual change. The success rate of 0. different combinations of methods in a composite forecast. At the time of forecast. For horizons more than one step ahead. with other univariate methods (trend extrapolation and exponential smoothing) coming off worst. The poor showing of econometric methods is in contrast to the finding in Fildes’ (1985) survey that compared causal (econometric and transfer function) and extrapolative (ARIMA and exponential smoothing) methods. This result would be unsurprising if VAR was regarded as a causal method. Second. Few comparisons with the naive method were located. The strength of the competition obviously matters.) How would the methods have fared if the competition had been equal? That is. the choice of criterion is important here also. for exam- . Table 8 summarizes comparisons from 41 series reported by 13 studies [Freebairn (1975) contains almost half the series but makes only one pairwise comparison of each]. VAR and the composite methods perform best. The commonest criterion is number of directional changes correctly forecast.
35 00 60 7 0 4 13 0.5 29.96 _ 0.5.38 0. . 0. ECOIl.13 0.33 0.40 0. A tie between two methods receives a better than/worse than score of 0. simple Comp.5 54 3 45 42.64 Naive ARIMA Other Univar.00 0. other 1.53 0. the commonest being the ratio of turning point errors to total number of changes.5 33 26 7 59.00 VAR Other multi Camp.00 0. of ex ante forecasting ARIMA Simple 0 3 0 0 4.5 0 4 0 2. Because turning point measures for different methods often score the same value.90 0.5 6.56 0.40 0.69 0.5 1 6 13.5 41.5 23 5. The lower tnangular matrix 1s the success rate.5 0 Y 0 6 14.5 0 7.00 0. The upper triangular matrix is the number of better than/worse than comparisons.5 42 0.46 _ _ 0.5 19 Overall success rate 0 See the footnotes to Table 7.5 0 1 Expert VAR Other multi Composite turning points Total B/W Other Naive 0.5 0 7 1000 8.40 0. Expert 2 0 002 0 0 0 0 0 0 4 6.55 0.5 13.5 1 0 0 0.33 0.5 5.00 0. read in the same way as in Table 7.38 0.5.58 0.5 00 15 00 10 21.04 0 36.Table 8 Comparison performance: Other univar.58 0.s 0 5 12.5 0 0 3. Various turning point measures were used in the different studies.38 0.5 Econ 0.5 6 0.42 _ 0. those methods tie in the rankings.00 0 9.00 0.69 3 42.
is the number of upward observations. can only be followed by a further move upwards or a peak-type directional change. 1986) that when removed increases by one the number of correct forecasts for the simple composite of two methods. both forecast and actual.I. ratio of accurate forecasts.2. a turning point is forecast when (F..G. Since the first part of the expression is known at the time of the forecast. This adjustment is reflected in the rankings.3.)<O.. unrestricted VAR models with large numbers of parameters have been found to perform rather poorly._.)x(F. Kaylen McIntosh. .) eliminates certain off-diagonal elements from consideration..F. the known value of (F. Kaylen (1986) modified theformulato(F.. is the number of correct downward forecasts and nz the number of incorrect downward forecasts and n = nI + rr2.._.. Ina2~2 table of directional change and no change..P.) < 0 where F.F. For example. a previous upward price movement limits the forecast to a peak turning point or a continuing upward movement. . N = N._. .. The sum of the number of correct upward and downward forecasts is the same as the values calculated by Kaylen. trough Ratio accurate forecasts Henriksson-Merton Number “P Number down correct correct 7 4= 1 7 3= 4= 5 l= l= 4= 6 1= 7 7 7 6= 3= 4= 2 6= l= l= 3= l= 3= l= 1 3= l= 4= 5 3= 3= I= 3 1= 3= Sum of ranks Ranking Sum of ranks excluding Ranking correct 30 6= up and down 22 7 30 6= 21 6 18 4= 11 4 13 2 9 2 14 3 10 3 18 4= 12 5 12 8 All studies use the data series for hog prices (price of barrows and gilts at seven terminal markets) and forecasts from Brandt and Bessler (1981)._*) X (F.. Adaptive weights 3= Brandt. McIntosh and Dorfman (1992) use the procedure of Naik and Leuthold (1986) to construct a 4 x 4 table that separates peaks.1. ple. + N. so that for overall ranking the disaggregation adds to new information. . downward movements and troughs. maintains the rank ordering since it provides no new information. The Henriksson-Merton measure is the conditional probability of a correct forecast. In Brandt and Bessler._. Dorfman Bessler Correct/total Separate peak. An almost identical data set and results are also found in Bessler and Brandt (1981). The ratio of accurate forecasts (RAF) is the sum of the diagonal elements divided by the total number of forecasts.. The original article contains an error (noted by Naik and Leuthold.. even though broad similarities do occur. 7.. var. 1) x 67 / . Allen I International Journal of Forecasting 10 (1994) Xl -135 107 Table 9 Rankings Author of different one step ahead turning point measures on the same methods (2) ARIMA (3) Expert and data series Error measure (1) Econometric Composite Simple (1) (2) Simple (1) (2) (3) 6 I= 2 4= Min._. N. For a one step ahead forecast. is the (one step ahead) forecast for time t made at time t . is the number of downward observations. . Vector autoregression methods Six studies that compared vector autoregression methods are reported in Table 10. The variable reduction methods that use variable ordering and information criterion statistics produce the most . McIntosh and Dorfman calculate an exact confidence level based on the hypergeometric distribution as where N. n. ) < 0 and the forecast is correct when (F.F.._. The summary statistic.F. upward movements. For one step ahead forecasts.-F. As noted by Kaylen (1988). ~ F. . But the Henriksson-Merton confidence intervals reported in McIntosh and Dorfman (1992) lead to different rankings. the sum of the diagonal elements is the number of correct forecasts.-F. the RAF has the same value in both 2 x 2 and 4 X 4 contingency tables.
The values b.3 2. the values in brackets [thus] indicate the ranking of the methods according to the criterion listed.12 RMSE TPE l-8 RMSEI MAPEI RMSE4 MAPE4 RMSE I .3 d1. The method of Tiao and Box is to delete insignificant variables. followed by Bayesian models with general prior distributions. Hsiao and Schwartz methods examine different combinations of lagged variables in each equation.1. are areas where improvements need to be made.l dl. Probability Decision probability distribution. accurate forecasts (Hsiao and Schwartz methods). symmetric Bayesian prior. because of their historical emphasis on the analysis of resource allocation decisions.s 2.3 d3.o 2.x Yes Fanchon.2 5.7 Schwartz Prefilter (Parzen) Bayesian Symmetric General VAR overall best Yes Criterion and horizon Unrestricted Bessler. unrestricted.2. Agricultural econom- makers want information on the distribution of a point forecast. Khng and Bessler (1985) use both r-test and F-test statistics to delete insignificant variables. He found no difference among methods for one step ahead forecasting. ists.4 4. Allen I International Journal of‘ Forecasting 10 (1904) Xl -13. Babula Kaylen 1987 4 RMSE 1.108 Table 10 Pairwise comparison Author Date P.a.s 2.1 4. Kling. Park (1990) reports a x2 test of forecasting ability. and a number of other practices that have been criticized in earlier sections. u is undifferenced (raw) data and d is first differenced data.G. .0 No No No No Yes Park 1990 4 111 PI U2.5 5. 8.0 s.4 0. ‘In Zapata and Garcia (lY90).1 ‘Zapata. though the differences were significant for longer horizons. Both were more accurate than USDA-ERS forecasts. Bessler’ 1985 5 I.O 6. forecasting A much-needed development is to combine forecasting with decision making (the so-called decision support system).o 1.l 4.5 l.2 U2. Garcia 1990 3 Yes Yes d3 .4 3.s 2.0 [31 uo.w indicate the number of pairwise comparisons that the method listed at the head of the column is better than and worse than other methods examined.6 5. The prefiltering method (Parzen) uses the residuals from univariate ARIMA models applied separately to each series. are perhaps more likely to emphasize this area than are other forecasters. The or the statistics that summarize it.5 of vector Series autoregression methods Variable Tiao-Box reduction Hsiao. with re-estimation and repeated deletion if necessary. Parameter weighted information criteria (Akaike.3 0.O 4.6.0 u0.1 3. Elizak and Blisard (1989) found that Hsiao’s method was generally slightly more accurate than Kaylen’s for five retail meat series at one.8 I’)88 3 [41 ISI 111 131 PI n. Where this information could not be extracted from the study. Their performance was similar with one-step ahead forecasts but the t-test method was worse at four-steps ahead. This.3 1. usually after judgementally ranking the series in order of causality. Schwartz) are used to find the best set of lags.4 1.3. 8.1 4.6 RMSE 1-6 TPE l-6 MSE l-58 7. four and eight quarters ahead. Current developments and conclusions 8. Wendell I992 4 ‘Bessler and Kling (1976) used an almost identical data set (the forecast period was extended a further 12 quarters) and found that the ordering of accuracy was general Bayesian prior.
A comment on the paper suggested that such a rule could lead to strategic behavior. forecast evaluation and dissemination of results. Their confidence intervals on corn price. No calibration tests were performed. contain verbal descriptors such as ‘expected to return to nor‘should be around’. In a practical setting. Grisley and Kellogg’s (1983) study is an exception. Nelson (1980) outlined a specific proposal that included a survey of user needs. shortage of data is the problem. Respondents could maximize their expected reward by stating more concentrated probability distributions than they actually believed. the linear scoring rule was improper (so that the reward structure gave participants incentives to misrepresent to the researcher their true subjective probabilities). When actual rewards are used to motivate individuals. since the amounts of money were relatively large (up to one day’s pay) and the respondents were risk averse. from the earliest to the present day.e. basing payments on an improper scoring rule ultimately leads to biased assessments. The linear scoring rule used by Grisley and Kellogg (1983) is easy to interpret and explain to participants. Little follow-up work appeared. Making and testing probabilistic forecasts requires more data than does point forecasting. Many studies have reported subjective probability distributions elicited from farmers for future prices and yields. Outlook reports.P. mal levels later in the year’ and so on. no actual payoffs are made for forecast accuracy. However. Payment of cash rewards might matter much more than choice of scoring rule. Most likely. The respondents who were paid according to an improper scoring rule initially stated the same probability forecasts as those paid according to a proper scoring rule. in the manner of weather forecasts. While a reward encourages participants to take the task seriously. Knowledge of the option premium and an assumed lognormal distribution enabled Fackler and King to calculate an artificial probability distribution on the futures price 4 weeks and 8 weeks ahead. the goal is to obtain subjective probabilities which are as accurate as possible. Bottum (1966) and Timm (1966) both recommended that probabilistic outlook forecasts be developed.b). the typical economic price series of monthly. Allen I International Journal of Forecasting 10 (1994) 81-135 109 provides information on the precision the forecaster attaches to the point forecast. The problem was recognized in agricultural forecasting in the late 1970s [Teigen and Bell (1978a. despite the fact that the need for it was recognized almost 30 years ago? Outside of weather and sports forecasting. Why has probabilistic forecasting received so little attention. A handful of other agricultural applications have appeared since: in a policy context [Bessler and Kling (1989)] and comparing univariate and multivariate [Bessler and Kling (1990)]. More common in the general forecasting literature are attempts to mechanically calculate correct (i. well calibrated) forecast probability distributions. The first testing of this kind to take place in an agricultural context was by Bessler (1984b). a decision maker can see the risk that attends taking the point forecast as true. quarterly or even annual frequency offers many fewer opportunities. the same question could be asked about any kind of forecasting. did recognize uncertainty in the values of both the parameters and the explanatory variables. Spriggs (1978)]. but after about 20 successive forecasts the two groups diverged. Compared with daily or more frequent weather forecasts made at many locations. training. Prescott and Stengos (1987) demonstrated the bootstrapping approach by constructing confidence intervals for pork production forecasts. development of elicitation procedures. Such descriptors convey only part of the information possessed by the commodity analysts. based on option prices of commodity futures [Fackler and King (1990)]. Nelson and Bessler (1989) provided the same historical information to a sample of participants and asked each to forecast a series of 40 probability distributions. A reviewer suggested that probabilistic forecasts are more challenging to present . based on a sector model. Equally. choice of scoring rule might be important. Analysts responsible for outlook forecasts have recognized the uncertainty of the forecast in a qualitative way.G. Generally.
The highest expected profit might carry too much risk for a risk averse producer who might elect to sell calves more frequently than the expected profit maximizer. or rearing for sale as a feeder steer (in the second quarter of the following year) (Spreen and Arnade (1984)]. Park et al. though there were exceptions [Lave (1963). then sell. For a moderately risk averse decision maker. the producer of a (fourth quarter) beef calf who must choose between immediate sale.2. Brandt (1985) and Park. To attain maximum profits. perhaps surprisingly. otherwise doing nothing. (1989) used monthly data. the producer (buyer) sells (purchases) a futures contract when the forecast price is below (above) the futures price. A risk averse producer would want to know how precise each of the forecasts was. Forecast information was of little value. This is perhaps indirectly saying that a forecast needs to make sense to the recipient in a decision making context. Whether outlook price information was available or not. the forecast from an exponential smoothing model gave the fewest wrong decisions. Brandt’s results suggest that the adaptively weighted composite forecast is better than the simple average composite. otherwise it should be sold. The difference between having no predictive information about the outcome as opposed to some predictive information was not often put in the forecasting context. . Peck (1975) calculated the optimal hedge for egg producers using the forecast error variance of three forecasting methods (one of which was the futures price itself). the strategy with highest expected return and acceptable risk required outlook information. 8. although with quarterly data the analysis is probably too aggregative for useful decision making. a second area of current development. In this study. including one that forecast the probability of feeder steer price exceeding the break-even price. a risk neutral decision maker would pay nothing for outlook information. the outcome using perfect information and the outcome using a frequency distribution from a historical series were compared. Eidman et al. The difference gave the value of perfect information. a producer’s hedging decision can be formulated as a portfolio analysis problem. if the feeder steer price exceeds a break-even value. Skees and Black (1984) examined a similar situation. Even studies that compare decision makers with different risk preferences usually ignore the forecast error distribution. DeCanio ( 198O)j. the correct decision is to retain the calf. Requiring a payment from the farmer of $450 for the outlook information removed that strategy from the non-dominated set. (1967). Bullock and Logan ( 1970). When futures markets exist for a commodity.than are point forecasts. where a grain producer with a known cost of storage had to choose each month between storage or immediate sale. Under these conditions. Such a person would be unconcerned about the probability distribution around a forecast. since a fully hedged strategy gave almost the same returns and risk reduction as partial hedges based on the forecasts. Spreen and Arnade compared forecasts of feeder steer prices by five different methods. Early researchers simply used historical data as risk measures. Some studies focussed explicitly or implicitly on a risk neutral decision maker. Consider. Garcia and Leuthold (1989) considered how producers and buyers of hogs might execute selective hedging strategies based on prices forecast by different methods. that ARIMA forecasts performed well at all horizons and for all stochastic dominance criteria against simple composite and econometric model competitors. They found. the correct decision (the one that gave the greater profit) could be seen. providing an indication of the value of outlook information to that producer. In a selective hedge. econometric and ARIMA models. for example. When each season had passed. Forecasting and decision making Quite early in the development of the theory of decision making under uncertainty. the strategy with highest expected return and highest risk was to store for 3 months. Rister. except for bias.
Conclusions Agricultural forecasting uses a wide range of techniques in a wide variety of situations. papers rarely quantify the fragility of their estimates. The conclusions are drawn from published studies and there are many unpublished forecasts. although (as noted in section 7. a test which should be carried out stochastically anyway. The largest group are outlook forecasts. On the other hand. vector autoregression is replacing simultaneous equations systems. It stems. from the desire and the training of agricultural economists to explain phenomena rather that predict them. subjecting systems of simultaneous equations to within-sample validation has become common practice.3. They have a long history and a detailed and specialized development of leading indicator analysis unique in forecasting. but lagged endogenous variables are usually the . no significant improvement is revealed in the success rate for short-term forecasts of econometric models against other methods. It is usually carried out deterministically to assure the stability of the estimated model over time. But there are too few comparisons of variable reduction and Bayesian approaches to favor one over the other. on on econometric modeling of ever increasing complexity has been a hallmark of agricultural forecasting. perhaps. If this review appears to have concentrated on short-term forecasting.76 for medium and long-term ex ante forecasts (11 studies) and of 0. Exogenous variables are taken as given. analysts make conditional predictions or projections based on assumptions that are oversimplifications of any policy that might arise. Often. More studies are needed on the relative success of causal models for long-term forecasts. There is an impression that econometric practices have improved over the years though concrete evidence is hard to produce. it is because few published long-term forecasts were located. Dividing the econometric models in Table 7 somewhat arbitrarily into the 12 studies (22 series) before 1985 and the 11 studies (23 series) since 1985. For short-term production forecasts. Composite forecasting is best. However. Although the types of forecasts required are similar to those found in business. Results found here conform generally to the beliefs held by forecasters. perhaps because in most of the comparisons their specifications are not developed in any systematic way.67 for short-term forecasts (ten studies). While the sensitivity of parameter estimates to choice of specification is now widely acknowledged.G. mainly of production. A typical econometric modeller performs a limited amount of testing of dynamic structure (often only the Durbin-Watson test. Where conclusions are based on relatively few results. The limited evidence of a number of careful VAR studies supports the generally held view that unrestricted VAR models are not efficient. Exceptions are the relative handful of studies that use VAR or that investigate cointegration among variables where more detailed testing is a necessity. The best that can be said is that. indeed one might say overemphasis. agricultural forecasters have made little use of univariate time series methods. Emphasis. Structural econometric methods do less well than their proponents might have expected. at different national and regional aggregations. Allen I International Journal qf Forecasting IO (1994) Xl -13 111 8. the difference is not statistically significant. producer intentions are good indicators.2) the case for using simple averaging over other composite methods is less clear. Fildes (1985) reported success rates for causal models in ex ante forecasts of 0. though matters are improving).P. Nor has much systematic investigation dynamic specification into occurred. the addition of a small number of studies might reverse the findings. Fascination with the subtleties of different econometric methods has produced numerous articles but has had little influence on performance. The ratio of policy analysis to long-term forecasting appears higher in agricultural applications than elsewhere. There is also no widely agreed belief about the best specification to build from. for pure forecasting applications. There are too few comparisons to single out a particular composite method as best. since about the late 197Os. where they would be expected to do better.
but using actual values of exogenous variables that would be unknown at the time of forecast (or conditional forecasting) defeats the objective of finding the best forecasting model. arguing for using all available data and for conforming with economic theory. Progress depends on researchers following a proper validation procedure. Post-sample forecasting is better than within-sample. A current unresolved question is that of whether to build more or less structure into a model. over 1 year ahead of the American Economic Review (September. How well does the model perform out-of-sample (in a holdout sample for cross section analysis. A counter-argument is that data and the relationships of different reliabilities are mechanically aggregated. One suspects that Theil’s statistic is often not reported because of the poor light it would cast on the model’s forecasting performance. details that can affect the outcome of a comparison must be given. as a matter of course. I am particularly grateful to Scott Armstrong and Bill Tomek. Doing so would shift the focus from the point estimate. to avoid adding layers of variability to the forecast calculation. All applied economists (and not just agricultural forecasters) should ask themselves three questions. if only with the performance of a naive model. help and inspiration. creating larger forecast errors than would suitable (and probably judgmental) weighting. the editor. Since there are no standard procedures for testing forecasting ability. which will be wrong anyway. two anonymous reviewers and other colleagues for comments. this may truly reflect the level of ignorance about the future.e. Reporting of empirical studies in economics journals is improving. forecast rather than actual current exogenous variables. Aside from the need to examine dynamic properties. 9. (The International Journal of Forecasting was a pioneer in requiring disclosure. 1990)). Large scale econometric models seem particularly remiss in making comparisons. If current exogenous variables were treated as unknown.G. An argument in favor of more structure is that although forecasts may have higher variance. we should admit that we know less than we claim. there is no reason to use calculated values of lagged endogenous variables that would be known at the time the forecast is made (i. revising its guidelines in Fall. A good comparison would be with a single reduced form equation with. in both cases. but it would not have been as good. Few go as far as the Journal of Agricultural and Resource Economics in demanding documentation of model specifications estimated but not reported in the submitted manuscript. Is the model re-estimated? In what way? How often? How many steps ahead? (Does one take the omission of this detail to indicate one step ahead?) Are forecast accuracy statistics based on aggregation of all the h steps ahead forecasts or the one through h steps ahead forecast? More comparisons are needed. Econometricians generally favor more structure. Forecasters take the opposite view. Without their help. to the information content of the forecast. Mark Nerlove. make quantitative probability statements. or in the post-estimation period in time series analysis)? How does the model perform compared with the one it is intended to replace? And why the difference? Finally.112 P. at least gives a comparison against a naive no change model. a much better assessment of forecasting ability would result. Theil’s U. RMSE in particular provides little information on relative performance. Forecasters should. variables whose lags are longer than the forecast horizon). Allen I International Journal of Forecasting 10 (1994) 81-135 calculated values. . I would have finished this paper a long time ago. Many journals now require as part of the guidelines for manuscript submission that the data be available and clearly documented and that details of computations sufficient to permit replication be provided. Acknowledgments I thank David Bessler. 1989) and the American Journal of Agricultural Economics (February. 1988.
1985. International Journal of Forecasting. Journal of Marketing Research. 9-75: 4)..A. Journal of the Operational Research Society. 1970-1993 (Vols. 1982-1991 (Vols.12. 7110.S. 1970-1993 (Vols.1 to 1972. Quarterly Review of Agricultural Economics. feed grains. time series (set 7). a clarification. publication nos. 1987-1993 (Vols. . markets and marketing) and the keyword root ‘forecast’ over titles and abstracts for the years 1969-1992. 1992. PR. 1987. l9: 2). 1). ST. 1844: 2). Fildes. article is referenced in main paper. 38-61: 1).. No forecasting. l-10). comparison of forecasting methods.P. Decision Science. Canadian Journal of Agricultural Economics. October and December 1978. 36. Articles in the bibliography text are denoted by R. Rausser. TS. 2nd edn. coordinated by Z. Policy Sciences.S. R TS Illustrates method with monthly farm. (Wiley. eggs. This produced 245 citations of which six had already been identified through searching the above journals. 1993. 1-41: 1).. International Journal of Forecasting. 8. 7344752. Review of Agricultural Economics (formerly North Cen- Bibliography key: CO. and G. ‘predict’ or ‘projection’ for the years 1970-1992.. supply and demand. Long-range Forecasting: From Crystal Ball to Computer. 1978. 379-391. 1977-1993 (Vols. Australian Journal of Agricultural Economics. wheat and farm inputs. Journal of Agricultural and Resource Economics (formerly Western Journal of Agricultural Economics). tral Journal of Agricultural Economics). Commodity Forecasting Models for Canadian Agriculture.” (H. Huff.. Journal of Agricultural Economics.C.. Just. l-44: 2). 359-380. The governance structure of agricultural science and agricultural economics: a call to arms. dairy. J. single equation econometric (set 4). Bliemel.A. MK. Ahlund. Huff (Ottawa. 19701993 (Vols. large scale (set 6). 1948-1978 (Vols. D. Barksdale and J. 11. Canada). Allen I International Journal of Forecasting 10 (1994) 81-135 113 10. 1986. single sector model (set 5). Sixty-six citations were identified as possible additional references. hog marketing. Dalrymple. American Journal of Agricultural Economics (formerly Journal of Farm Economics). 13-22). American Journal of Agricultural Economics. EV. 10. situation and outlook. Journal of Forecasting. F. also includes value of information and Bayesian decision making (set 8). Includes hog supply. References Only references not found in the below. 75 (October). New York). LA. 14-37: I). probability.J. l-25: 1). 3. Quantitative forecasting-the state of the art: Econometric models. retail beef prices from 1949. R. McNown. outlook (set 3). Appendix A: Publications searched 12. OU. Collopy. Error measures for generalizing about forecasting methods: Empirical comparisons. 444-446. “The primary goal of the research program is to improve the forecasts of basic market relationships to complement the work of commodity specialists. wholesale. Multivariate spectral analysis: an illustration. R. Review of Marketing and Agricultural Economics. broiler. 1984-1993 (Vols. 1952-1993 (Vols. bibliography are listed referenced in the main Armstrong. Hassan and H. 2 Vols.W.L. 1949-1993 (Vols. R. 7812 and 7813. R. ‘Introduction and Overview’. PG. 1-18: 1). p. Theil’s forecast accuracy coeffcient. evaluative or critical assessments. On the use of econometric models: A guide for policy makers. Several of the models are projections rather than forecasting models and model simulation is illustrated rather than forecasting. 1927-1993 (Vols. 69-83.G. 1977. beef. The Dialog database files 10 and 110 (Agricola) were searched over titles and descriptors using the keyword roots ‘economic’ and ‘agricultur’ and either ‘forecast’. market efficiency. 7150 (agriculture: general.C. 3-33) (continues with different subject matter emphasis as Agriculture and Resources Quarterly).B. and F. 1985. SN. 69-80. 1979-1993 (Vols. H.B. J. 1969-1993 (Vols. Hilliard. M. 7120. 19. 1-15: 2). Armstrong. 549-580. 1973. Journal of Agricultural Economics Research (formerly Agricultural Economics Research). 8. Agriculture Canada.E. pork. A further 21 citations were identified as possible references and eight were added to the bibliography. Two articles out of the 245 were added to the bibliography. R SN ST Describes ten of the 13 structural commodity models under development by or on behalf of Agriculture Canada. Sales forecasting practices. although most were cost of production projections or regional crop and livestock supply-demand projections published by various USDA branches. This produced 70 citations of which 27 had already been identified through searching the above journals.. programming model.B. Appendix B: Bibliography Agricultural and Resource Economics Review (formerly Northeastern Journal of Agricultural and Resource Economics). The Dialog database file 139 (Journal of Economic Literature) was searched over descriptor codes 7100. Journal of Agricultural and Applied Economics (formerly Southern Journal of Agricultural Economics). International Journal of Forecasting.
788-792. 38. 31. 1939. R PR Review and possible applications in agricultural economics. A. AR model more accurate than naive no change in one step ahead post-sample forecasting. 72. Baker. 1973. Chalfant.. Statistics and Cooperative Service (ESCS) using the outlook and situation information system (OASIS).S. J. R TS Good explanatory article on VAR. Alternative estimates of fed beef supply response to risk. R TS CO For beef. Forecasts from a nonparametric approach: ACE. 32. in P. Antonovitz. Western Journal of Agricultural Economics.. 13. Chapter 4. T. R OU Early history of outlook work in the US.J. Barr.R. A quarterly econometric model of the United States livestock and feed grain markets and some of its policy implications. Applied Economics. TS CO Compares restricted and unrestricted four-variable VAR and univariatc autoregressive mod&. Brandt. and J. and M. Turning point accuracy and error reduction accuracy scores. Bessler. 13. D. Journal of Business and Economic Statistics. American Journal of Agricultural Economics. ST CO Compares ordinary least squares. 5. Babula. Journal of Agricultural Economics. An analysis of dynamic economic relationships: an application to the U.A. 264-267. Composite Forecasting of Livestock Prices: An Analysis of Combining Alternative Forecasting Methods Purdue University. TS Compares within-sample cranberry yield forecasts from univariate ARIMA and from econometric models with various combinations of actual and forecast monthly average temperatures. 1088. Compares three individual and four composite methods for hog.T. OASIS-an overview.J. Green. Ames. and J. Becker. 6333643. and W. American Journal of Agricultural Economics. Developments in crop and livestock reporting since 1920. Aradhyula.N. 43-52. 109-124. 1988. Canadian Journal of Agricultural Economics. pigs and corn sectors with quarterly. 21. S. D. Harlan. 22-31. Outlook evaluationmethods and results. Annual. Bessler.A. 72.. Forecasting multiple series with little prior information. OLS predicts cotton and rice imports with least MAPE. R SN A review (190 references) containing about 550 short-run and 550 long-run price elasticity estimates categorized by commodity. TS CO See McIntosh and Dorfman (1990) competition. IA). R. ST A six equation semirecursive model.G. H. R OU Describes the process of generating and disseminating forecasts in USDA-Economics. Agricultural Economic Research. P. OU. naive is best and OLS worst. adaptive and rational expectations. Journal of Agricultural Economics Research. Babula. Journal of Economic Behavior . 365-374. MSE decomposition shows that wheat price export shipments are slightly related to past exchange rates. ARIMA worst MSE. Canadian Journal of Agricultural Economics. Outlook 1Y92. 7999803.A.A..L. but econometric with forecast temperatures worse than trend regression. 33-38. lY84b. Estimating agricultural supply response with the Nerlove models: a survey.A. Askari. Arzac. Contemporaneous correlation and modeling Canada’s imports of U. 41. et al. Medium term outlook for tree fruit. P. Beenstock. 397-406.M. and J. T. Bhansali. 1990. P = 2 was best fit according to several final prediction error criteria. naive. Bessler. 53-72. and C. Cummings.. R ST Applied to the world sugar market. 18.V. carryover and price forecasts for wheat. 799-827.W. Forecasting livestock prices with individual and composite methods. 19X7. F. OU Recent history. Uses VAR. A quarterly forecasting model for the consumer price index for food. Holt. Wilkinson. Ashby. and D. 1990. including forecasting.. followed by SUR. hog market. 27. Based on 1983-1985. and R. Rasmussen. Bessler.D. Australian Bureau of Agricultural and Resource Economics. 13. 105-l 14.A. Bell. LA CO A 42 equation model including beef. cattle and broiler prices. 237-242. l-14. Agricultural Economic Research. R.A.. Also references other ABARE publications. Brandt. 46. Subjective probability. 30.. Bessler. Forecasting wheat exports: do exchange rates matter‘?.F. 1992. Paarlberg. R TS See McIntosh and Dorfman (1990) competition. 475-487. Agricultural Economic Research. Uses alternating conditional expectations. D.A. D. Baker. American Journal of Agricultural Economics. while for soybeans. D. Gale. On forecasting commodity prices by the balance sheet approach.L. Allen I International Journal of Forecasting 10 (1994) 81-13 Allen. and J. R OU EV Production. 1975.G. 1964. 1-7. OU A politically oriented historical review with very little on forecasting or outlook.A. J. 1977. GARCH time-series models: an application to retail livestock prices. Agricultural Economic Research. and M. 1981.A. 1979. crops. International Economic Review. Berck.A. Journal of Farm Economics. Journal of Farm Economics.T. 25. 1980.A. Barry (editor). 1990. G. A note on forecasting with econometric models Northeastern Journal of Agricultural and Resource Economics. American Journal of Agricultural Economics. and H. 61.114 P. 257-292. 591l 602. Economic research in the Department of Agriculture: a historical perspective. E. 513-522. An analysis of forecasts of livestock prices. pork and chicken prices a within-sample test showed that a GARCH model fitted better than an AR model with time trend. R SN TS CO A condensation of Bessler and Brandt (1079). Risk Management in Agriculture (Iowa State University Press. location and data range. 4. 1984a. 1984. Analysis of cocoa price series by autoregressive model fitting techniques. Bessler. Barichello. semi-annual and annual equations.A. pp. 1979. D. TS CO AR(p) model fitted to monthly data. 1952. and R. M. and R. Brandt. and J. 72. IN.A. TS CO The study is condensed and reported in Bessler and Brandt (1981). seemingly unrelated regression and naive post-sample forecasts of annual imports. 1990. D. 1978.D. Agricultural Experiment Station Bulletin 265.S. No consistent ordering for forecasting ability. Bessler. West Lafayette. SN CO PR Compares forecasts of econometric.
ST Monthly three equation econometric model of US and Australian prices and US exports validated-within-sample and used for both forecasts and simulations of different policies. Bessler.S. Moore. Little difference in the means and standard deviations of the prices resulting from each strategy. Kling. W.L. TS Illustrates the Box-Jenkins (B-J) approach.. Use of probability assessments and scoring rules for agricultural forecasts. Review of Marketing and Agricultural Economics. Edwards. 11-4 70-81. 1986. 2. 1982. J. Restricted VAR has lowest RMSE in post-sample forecasting at all horizons up to ten days.A. Time series modeling of economic phenomena. D.L. Forecasting vector autoregressions with Bayesian priors. Brandt.E. 751-758. Policy analysis usually included forcing variables (exogenous variables set by policy). 11. 1986. J.W. 249-263. 1990. Covey. 68. 1979. Agricultural Economic Research. Blond. Western Journal of Agricultural Economics. Brandt. American Journal of Agricultural Economics. 1970. Bessler. The forecast and policy analysis. R TS CO Quarterly hog prices forecast better by ARIMA model than by four equation VAR (using Tiao’s method of variable reduction) with first-differenced data. TS CO Compares futures price and expert as forecasts of one quarter ahead cash prices of steer cattle and hogs. Bessler. Forecasting an agricultural system with random walk priors. Forecasting and hedging: an illustration of risk reduction in the hog industry. R SN TS CO compares single equation. D. Brandt. J. Bessler. estimated by VAR. Distributions cannot be recalibrated to later time periods. VAR has significantly smaller RMSE for out-of-sample one day ahead cash price forecasts. 1966. Comprehensive forecasting and projection models in the economic research service. 1976. VAR has tighter distribution and is well calibrated.. Agricultural Economic Research. J. 9. US hog prices. Business Economics. and from expert and constant. 1984. 1154-1159. 95-106. Journal of Applied Meterology. Generates by simulation daily forecast distributions. R TS Assessment of predictive performance: reliability (calibration) and ability to sort. Penn. Bessler. Composite forecasting: an application with U. R. Comparison of four methods. Applied Statistician. Changing functions of outlook in the U. 503-506. Quance.A. New Mexico. Journal of Forecasting. M. D.A. 1984. and P. 1987. Hogg. Univariate cash price forecasts are not. ST A four equation sector model. American Journal of Agricultural Economics. Bottum. 52. No forecasting. Based on the work of R. from expert and random walk. Winkler and A. and C.L.A.. SN TS CO Quarterly. 1991.L. probabilistic predictions. American Journal of Agricultural Economics.H. 1968.S.A. Young.J. Roop and L.A. and J. 1979. 47. two stages. Brandt. 1983. judgment. 18.. hog slaughter. Focus is on use of forecast by producer or buyer to make hedging decision.G.A.E. H. 461-474. Compares composite hog price forecasts formed from two series of university expert forecasts. Forecasting with a dynamic regression model: a heuristic approach . 21. Selective hedging based on a price forecast is some advantage over routine hedging or no hedging. 67. Boutwell. Chamberlain.A. and T. 7. hog prices. D. 28. 1991.. Comparing the Box-Jenkins and econometric techniques for forecasting beef prices. Modeling the impact of two agricultural policies on the US livestock sector: a systems approach.V. 41-51. R. 805-813.A. Bessler. 195-199. J. and D. and J. Agricultural Systems. J. (Hsiao) restricted VAR and univariate models in forecasting daily cash prices of slaughter steers. Forecasting with vector autoregressions versus a univariate ARIMA process: an empirical example with U.A.P. TS CO Compares US hog prices forecast by ARIMA and bivariate regression. Murphy. External effects and U.L. TS CO Compares accuracy of VAR form of cointegrating model. Price forecasting and evaluation: an application in agriculture. 237-268. J. 4. 43-48. 71. Bessler. corn and hog price. North Central Journal of Agricultural Economics. 39. Brandt. On Bayesian composite forecasting. 27-37. 24-31. Allen I International Journal of Forecasting 10 (1994) 81-135 115 and Organization. Kling. B-J the more accurate. 6. D. comparison of three types of priors and ARIMA. Bourke.J. and D. 135-140. Bessler.B. 59-67. 1985.J. 95-106. 1976. 1981. R TS CO PR Compares univariate and two-variable VAR models of daily cash and futures prices of live cattle. and A. ARIMA.A. Also referenced in Rausser (1982) p. Bieri. Schmitz. Bessler.M. Journal of Futures Markets. Kling.A. R TS CO Quarterly sow farrowings. Blake. Clevenger. 48. R TS CO A five variable system. D. 144-151. W. R OU Describes near-term outlook and long-range projection models in use by USDA-ERS in 1976. hog prices . but not for cattle price where expert forecast contains additional information. J. Kost. . Omega. but see Revel1 (1981). D. J.. PR Demonstrates the use of the logarithmic scoring rule (a proper rule that gives the expert assessor the highest payoff for setting stated beliefs equal to true beliefs). Journal of Farm Economics.A. SN TS CO Compares seven approaches: time series.A. Fits three variable VAR model using Hsiao’s method to find best specification (according to Akaike final prediction error criterion). and J. TS CO Forms Bayesian composites with beta prior distributions at various tightness parameters. II and A. North Central Journal of Agricultural Economics. and D. 63.A. 29-36. 829. and J. Futures price contains all information for forecasting hog price. R SN TS CO Compares same methods as Brandt and Bessler (1981). Cointegration: some results on U. and T. Hopkins. US shrimp market. 31. American Journal of Agricultural Economics. wheat prices.S. D. J. cattle prices. “Good probability assessors”.C. Example with VAR model. J. econometric. 44-47. 1989. Prequential analysis of cattle prices. expert opinion and four composite methods. C. A linked annual and monthly model for forecasting alfalfa hay prices. Bessler.S.A. American Journal of Agricultural Economics. Womack.C.A. Bessler.S.. Agricultural Systems. Brandt. more timely and effective dissemination. monthly US cow beef wholesale price. and D. Haidacher. R OU EV Need improved accuracy. 15.
B. SN Annual single equation model based on inventory values known by mid-February of the forecast year. coordinated by Z.W. Illustrates the calculation of par yield. pork and poultry components of the FAPRI model. No tests of predictive ability. PR LP used to compute Markovian transition probability matrix for six farm sizes (measured by standard man days) plus entry/exit class in each of eight UK regions. 392-398. Rausser. 131-143. 29.A. A. Review of Economic Studies. Chen. Journal of Agricultural Economics.F. C. D.A. 234-241. M.12). Hassan and H. and D. pp. 103113. Projections for 1965-1970 compared with actual. Spearin. usually over 1965-1968 by means of graph only. An application of statistical decision theory to cattle feedlot marketing.. 615-629. annual macroeconomic-type model. Current period MLC forecast more accurate than autoregression or no change model. J. and C. D. Agriculture Canada’s medium term outlook: 1990-95. 5. MK A hypothetical futures trader was assumed to receive the hogs and pigs report one day in advance and buy or sell a futures contract if the report contained unanticipated information.8-1984.8-1986. Shicksmith. 129-155. R OU Describes process. and D. 52. and M.C. Thabet. TS Compares AR and GARCH models to show that the pork cycle is better captured by a non-linear dynamic process than by a linear one. 34.. Huff. 1990. LA LP formulation with dynamics included through capital stock adjustment. specification and some simulation results. Forecasting annual cattle slaughter. 1970. 1971.A. W. 16. 216-225. Updating brings little benefits.M. 1990. Predicting wheat acreage in the prairie provinces. 1978. and G. Peel. OU Explains the meaning of ‘condition’ in crop yield forecasting. H. A ‘combined’ model with structural exogenous variables predicted by VAR rather than naive no change was worse than the structural model alone.A. Becker. summarizes forecasts . Chin. 235-241.E. 30.A. Allen I International Journal of Forecasring 10 (1994) XI-135 35. 1952.A.A.. with and without parameter updating. No forecasting. American Journal of Agricultural Economics.B. 265-282. 87-89. Capel. 73.. 1972. T. Bullock. Journal of the American Statistical Association. 6. 59. Policy simulations for 1986-1995 based on tabulated macroeconomic variable values. single equation. 38. No forecasts. The use of ‘pars’ and ‘normal’ in forecasting crop production. 1970. 107-116. Within-sample validation includes historical simulation. American Journal of Agricultural Economics. The Wharton agricultural model: structure. 5-14.T. Forecasting monthly cotton price. 4.. 1979.D. R. Chavas. and D. 1923.-P.. 65-76. 1987. 38. American Journal of Agricultural Economics.T.T. R SN Uses Cagan’s adaptive expectations model.. 63-94.F. publication no. Byers. International Journal of Forecasting. 13-19.G. R PR Single equations for monthly beef price and for number marketed used to forecast prices and to examine various feed/sell strategies. However. 75.116 P. Journal of Farm Economics. Canadian Journal of Agricultural Economics. 1977. Carter. R LA Second generation model (recursive or feedback approach).. 1981. Bullock. individual farm commodities (pork as example). American Journal of Agricultural Economics. Chen. Ray and 9. 185-197. Time and frequency domain representations of futures prices as a stochastic process. Cargill. Cluff. 1968. Informational content of government hogs and pigs reports. and S. 819-828. An econometric model of the Canadian agricultural economy. A. 1982. Production and investment response to changing market conditions.F. Agricultural Economics Research. Value of forecasts about 1% of gross value of cattle sold. R TS CO Compares 67 equation sectoral and five equation VAR models both in a time of ‘policy shock’ (1986. American Journal of Agricultural Economics.12) and in an ‘ordinary time’ (1984. Journal of Agricultural Economics.P. Italy. Forecasting livestock slaughter: an empirical assessment of MLC [Meat Livestock Commission] forecasts. J. Logan. Holt. Canadian Journal of Agricultural Economics. PR A single loss prediction equation and Bayesian measures used to arrive at optimum pesticide levels in controlling brown rot in cling peaches in California. 23-30. M. 52. Annual data and projections for 1975-1980-1985-2000. D. G.H. Chan. On non-linear dynamics: the case of the pork cycle. adaptation to risk and technical change applied to forecast five crop and one (?) livestock outputs and prices for N. S. Canadian Journal of Agricultural Economics. R OU EV Dispels some farmer beliefs about livestock reports: need for accuracy. price influence and resource allocation impacts. 67. VAR does well in ordinary time period but badly in policy shock interval. J. a first generation annual model of about 250 equations that covers six livestock and eight held crop sectors. technical know-how and government policies. Suggests intentions reports more valuable than late season crop size estimates. Southern Journal of Agricultural Economics. 1. Journal of Farm Economics. Projecting farm structural change. Ottawa. 3X. Buckwell. Canada. A decision-theoretic approach to crop disease prediction and control. R LA Describes the beef. OU CO Forecasts zero to four quarters ahead of cattle and sheep slaughter in Britain tested for unbiasedness and efficiency in revision and use of public information. R LA Large scale econometric. J. Forecasting the demand for agricultural products. 7812. aggregate agricultural income and prices.E. and M. 1991. An analysis of quarterly provincial and regional hog supply functions. Only a risk-neutral trader with low trading expenses would be willing to pay for such advance information. Cigno. Bessler. the information content of the report is low. Valuation of crop and livestock reports: methodological issues and questions. SN An appraisal of the forecasting method used by the USDA: level of economic activity. 711-718. Commodity Forecasting Models for Canadian Agriculture vol.. 1993.L. 4. and J. Galopin. SN Set of nine Nerlove-type partial adjustment single equations. Carlson. Cavin. R TS Spectral analysis. Breimeyer.B.E. in Agriculture Canada. Also tabulates actual annual changes and forecast changes for 5 years. 1952. Callander. J.
J.J. TECHSIM: a regional field crop and national livestock econometric simulation model. R LA ST Traces development of systems approach in USDA-ERS meat division.. using standard error and turning points. 1990. 1975. Adjusting dynamic models to improve their predictive ability. Agricultural Economics Research. 1987b.R. R LA A 30 equation semi-annual model of the beef-pork sectors.J. 1961. 117 International Journal of Forecasting. 72. R. Myers. Ingram and C. Journal of the American Statistical Association. Varying parameter regression best in two of the three series. A Dynamic Price-Output Model of the Beef and Pork Sectors USDA-ERS Technical Bulletin Number 1426. C. J. American Journal of Agricultural Economics. R LA Precursor of AGSIM.J. 21.4-2&l% compared with actual. Cromarty. TS CO See McIntosh and Dorfman (1990) competition. R. 172-77. American Journal of Agricultural Economics.G. Hildreth-Houck. R OU CO MK Compares survey of market analysts with USDA report. and S... Forecast distribution of producers in 1970-1971 and 1975-1976. exports with fixed and stochastic coefficients estimators. Journal of Agricultural Economics. LeBlanc. R LA Development of operating rules through simulation.. 1990. Hallahan. R. 1990.G. Supports efficient market hypothesis. Havenner. Coiling. C. Lkerd and A. AR(l). including macroeconomic and trade links). R SN CO Compares stochastically varying parameter regression (Swamy-Tinsley. Cromarty. 1972.R.) SN Fixed coefficient models generally superior to time-varying parameter models. D. pork and chicken retail prices from four econometric and one generalized ARIMA models using actual values of explanatory variables. which had error range of 3. 1967. Stillman and P. R. Womack for other articles in same issue. Hrubovcak and M. J. R LA EV Simple and non-simultaneous models forecast better. Colman. 6. Vancouver..B. R LA EV Forecasts from wheat and feed grains/ livestock sector models. 1956. Luby. and W. 1970. Prentice.E. 509-512. University of Manchester. D. American Journal of Agricultural Economics. 1970. Canada. Leech. Conway. 57.. Cox. Free market projections based on a formal econometric model. R. A dynamic model of a simulated livestock-meat economy.. 1959.H. EV Review of agricultural forecasting.R. 43. 1987a.K. W. 931-939. Actual 1960 population used to predict 1965 distribution. 509-519. Cornelius. 556-574.J. Taylor. Bulletin No.T. Maki. 17. W.C. Development of a systems approach for livestock research in ERS. 18. and W. SN CO Compares post-sample quarterly forecasts of beef. 54.J. A forecast evaluation of capital investment in agriculture USDA-ERS Technical Bulletin Number 1732. Forecasting wheat. Colman. and P. . corn and soybean U. J. 1965a. Kalman filter. Downey. A forecast of milk supply in England and Wales.A. G. 1990.S. 73-83. 1981.L. Journal of Farm Economics. Evaluated within-sample. (Abstract in American Journal of Agricultural Economics. 151. Agricultural Economics Research. Forecasting and Projection in the Agricultural Sector Department of Agricultural Economics. AR(2). Journal of Farm Economics. Economic projections using a behavioral model. Agricultural Economics Research. 47. Required two exogenous forecasts (income). (1990). American Journal of Agricultural Economics. 1965b. Crom. CooleyPrescott). 788-792. Conway. and C.K.. Childs. LeBlanc. 24. See Barichello. and A. 35-2. 63. Conway. PR Markov transition probability matrices with six size classes (based on milk output) plus entry/exit class (assumed three times the population) computed from annual permanent producer sample of dairy farms for each of 11 regions. An assessment (p. 365-378.P. 28 pp. R SN True single equation forecasting models 6-12 months ahead. CO Survey of agricultural economist forecasts for five agricultural products. R. Journal of Agricultural Economics. 1975. Forecasting Livestock Prices: Fixed and Stochastic Coefficients Estimation USDA-ERS Technical Bulletin Number 1725.P. 38. Collins. D. Crom. Crom. Hrubovcak and M.K. 13 region model of eight crop and four livestock products estimated in regional blocks by generalized least squares. SN CO Earlier version of Conway et al. 365) “We are in the infancy stage of estimating the economic interrelationships among agricultural commodities. Crom. Colman. Journal of Farm Economics.” Cromarty. N. R. An econometric model of United States agriculture. 712-714. R PR 19581965 sample of 236 farms (not for all years) classed into five sizes (number of cows) plus entry/exit group. American Journal of Agricultural Economics. 9-15. No validation or forecasting described. poster session at the American Agricultural Economics Association meetings. and W.S. 253-265.M. Conway.. Predicting hog prices. R. Owen. A forecast evaluation of capital investment in agriculture. R. 1975. 72.R. 72 (December 1990): 1384. Annual and semi-annual models. 963-972..M. P. R LA Annual 12 product agricultural sector model linked to Klein-Goldberger macromodel with no feedbacks (first generation model). 25 pp. R. 1-18. LA The model of Crom/Maki modified by 126 operating rules. The application of Markov chain analysis to structural change in the northwest dairy industry. 351-361. 1983. fixed coefficient (Lucas flexible accelerator and same function as Swamy-Tinsley with and without added independent variables) and random walk (total of 13). R ST Lists 128 operating rules for improving predictions from the CromiMaki model. Forecasts from a state-space multivariate time series model.J.. W.A. Allen I International Journal of Forecasting 10 (1994) 81-135 (based on Food and Agricultural Regional Model-large scale econometric. The reaction of live hog futures prices to USDA Hogs and Pigs Reports. 84-94. Irwin. Arnade. A preliminary evaluation of price forecasting performance by agricultural economists.B.R. Needed improvements in application of models for agriculture commodity price forecasting.A. Criddle. Nelson. Crom. Maki. K. Jones.. and D. K. et al. 57.
Patterson. Current questions on national agricultural outlook. McElroy and C. California. J. Engel. eight size groups plus entry/exit group. Eales. Edwards. Dorfman. 28.C. TS CO Neither true model nor any of three methods dominated.W. Why the government entered the field of crop reporting and forecasting. Most forecasts show small downward bias. M. R PR Assumes that farmers are profit maximizers and their observed production is based on an incorrectly predicted product price ratio. and A. C. 93-99. and A. 1982. American Journal of Agricultural Economics. R.L. Agricultural Economics Research.S. DeCanio. Thompson. See McIntosh and Dorfman (1990) for references to methods. Dalton. a finding of overconfidence consistent with the psychology literature.F. soybean exports. 15. Martin. yield and production of soybeans for different months of the year from 1965-1991. 718-734. Quarterly. J. Statistical vs judgement and audience considerations in the formulation and use of cconometric models. SN Annual single equation econometric model that measures only within-sample forecast accuracy. 48. American Journal of Agricultural Economics. Gutierrez. and L. 54. 1990. Downey. 1939. American Journal of Agricultural Economics. 37. Journal of Agricultural Economics Research. Dean. OU Description of 199(&1991 prospects. Projected 1974-197881982 and compared with actual.B.W.P. 1980. 116881174. I-14. Smith and R. K. log-linear and Box-Cox functions and actual exogenous variable values. Journal of Farm Economics. C. Hauser and S.E. SN CO Compares fixed and varying coefficent models. pp. 1982. 1963. Arnade and C. l-16. 1993. Dalton. Further developments in a model projecting short-term wool price movements. and L. 1990. OU CO Compares different government agency forecasts by decomposing MSE. B. Soybean production estimates: a journey through the last 27 years. Parnenter. CCC loans for nine crops and values of inventory change for 17 crops and four livestock commodities. volatilities implied by option premia usually overestimate the subjective variances of farmers and merchants.. Grain price expectations of Illinois farmers and grain merchandisers. B. 209-222. S. Dubman. Grains and oilseeds outlook.S. It is updated monthly and published quarterly using expected prices and production provided by USDA analysts.K. W. 575-576. SN Quarterly single equation econometric model that makes price projections. 1990.. A. ORANI: A Multi-sectoral Model of the Australian Economy (North-Holland. and E. OU Calculates average forecast (as percentage of final value) and 95% confidence interval for planted acres.E.S. Johnson and H.H. 64. pork production using a random coefficient model. Davison. J. 1992.. compares them with actual prices after market intervention by government agency and forecasts intervention purchase quantities. G.. 2344 258. harvested acres.N. 13-20. R OU EV Discusses timing of reports and outlook conference.. Sutton and D.F. M.H.S. R LA Description of a Johansen-type computable general equilibrium model with a detailed agriculture sector. 73. No post-sample comparisons. 88. Using an assumed product-transformation curve the difference between gross revenue based on the incorrectly predicted prices and that which would result from using perfectly forecast prices is the value of a perfect price forecast. Naive forecasts generally had smallest MAPE.. S. American Journal of Agricultural Economics. Short term wool price movementssa projection model. R. Quarterly Review of Agricultural Economics. Quarterly Review of Agricultural Economics. Journal of Farm Economics. followed by the linear model. 1985. J. McIntosh. Results of a price forecasting competition. Ebling. R PR Used lY7441978 longitudinal records of farms. Daly.D. 225-239. Projecting sheep numbers shorn-an economic model. Lee. B. State-space modeling of cyclical supply. 1975. Box-Cox estimation of U. See Tegene (1091). Used to calculate optimal carryover inventories. Tabulated forecast and final estimate values permit calculation of standard forecast accuracy statistics. 1973. C. 1989. Vincent. American Journal of Agricultural Economics. Hallahan.S. Amsterdam). Canadian Journal of Agricultural Economics. J. Havenner. 1975.T.J. Quarterly Review of Agricultural Economics.O. 196’). M..135 Crowder. M. 829-X40. Dixon. 1972. 28.H. Validated by Henriksson-Merton confidence interval test. LA ST EV Describes the forecasting situation in the commodity industry. 804-808. Dixon. R.R. American Journal of Agricultural Economics.E.. 1991.. 195-219.B. R LA The accounting-type model of approximately 1000 equations (listed in the appendix) forecasts cash receipts for 21 crops and 11 livestock commodities. seasonal demand and agricultural inventories. SN Estimates annual supply and monthly demand of five sizes of canned olives in California. R..G.. 72. R OU Early history. and C. 1966. analytical and data bases for outlook. Oil Crops Situation and Outlook Report USDA-ERS OCS-35. 55. 72. 8-16.. 41. R.A. 701~-708. However. Carter. Dorfman. 1974. Dalton. G. Journal of Political Economy. An aggregative model of agriculture: . Economic losses from forecasting error in agriculture. The changing distribution of farms by size: a Markov analysis.R. PR In most instances the futures price (of soybeans and corn) is an appropriate proxy for expected price. Allen I International Journal of’ Forecasting 10 (1994) XI.. SN CO Postsample forecasts (3 years) of soybean imports from the US into nine markets using linear. J. No comparisons with other forecast methods. 27. 21. 779-783. Forecasting Farm Income: Documenting USDA’s Economic Model USDA-ERS Technical Bulletin Number 1825. Forecasting U. 38.118 P. Douvelis. Supply functions for cotton in Imperial Valley.. Dodson. Dietrich. Egbert. need for improved accuracy. 48 pp. Agricultural Economics Research. R PR Uses 1950-19551960 census data to obtain Markovian transition probability matrix for five farm sizes (acres) plus entry/exit group (assumed at 100 000) to predict size distribution for 19551960-1965-1970-1975.G.J. SN Econometric model that is revised in Dalton and Taylor (1975). An evaluation of short-term forecasts of coffee and cocoa. Also projected 1990-2000. R. P. Taylor. 530-538.
10. V. Journal of Farm Economics. 1968. Two methods of forecasting hog prices..E. 1972. No significant difference found.N. Eidman. Feather.J. 1965. A. Kaylen. Allen I International Journal of Forecasting 10 (1994) Rl-135 empirical estimates and some policy implications. ‘empirical’ regression (using lagged explanatory variables known at the time of the forecast) was more accurate than a ‘synthetic’ method of setting expected supply (estimated from leading indicators such as pig crop survey) against a supply function. A dynamic long-run model of the livestock-feed sector. Applied Economics. 1954.R. restricted VAR has least MSE. 71 pp. R LA Annual four equation model treating agriculture as a separate sector with no feedbacks (first generation model). R PR Option price premia used to calculate probability distribution around the closing price of the futures contract 119 of the same commodity 4 and 8 weeks ahead. Southern Journal of Agricultural Economics. and W. 37-66. 1992. R. 518. 62) “. 1990.L. An evaluation of the rice outlook and situation price forecasts. Journal of Farm Economics. with VEC better only at the longer horizons. Quarterly forecasting of meat retail prices: a vector autoregression approach USDA-ERS Staff Report AGES 89-27.A. 14 pp. J. Fanchon. use of intentions data explains about 60-800/o of actual variation in production. Ezekiel. Tested on three hold-out samples with correct predictions about 70% of the time. 1967. R. 54. LISE and Cochrane-Orcutt methods. 24. Agricultural situation and outlook work. 72. 71. Over 1 to 58 months ahead. Holder. Williams. Quarterly models to predict cash prices of pork bellies. then VEC. 1288-1305. farm size more than or less than 400 labor hours) to discriminate between early and late adopters on a calibration sample of farmers. Overall 76% are in the correct direction.” Foote. Fisher.. Epperson. that a number of naive models would be able to do better..J. Reutlinger. Elizak. 22-30. guaranteed payment per turkey (contract A) and payment per pound (contract B). M. 33.R. P. 1988) with ERS forecasts worst. 71-86.. Elam. Williams.S. ST Annual model converted to semi-annual to make a 25 year projection. 1927. national and international. R OU CO Compares forecasting methods that use intentions to plant data with methods that use only past or projected acreage and yield. EW. Poor price prediction in backcasts to 1913-1914 suggested (p. American Journal of Agricultural Economics. 207-217. American Journal of Agricultural Economics. Jr.S. 1989. cattle. and S. A four-equation model of the feedlivestock economy and its endogenous mechanism.. Egbert. Rural Sociology.603-610. Table 1 reports directional accuracy of 219 FAO outlook forecasts for 13 commodity groups. and R. Ezekiel. Foote. American Journal of Agricultural Economics.O.. 26. Validated withinsample and used for long-term projections. G. and M. 1973.. Agricultural Economics Research.J. .g. ST A three equation recursive sectoral model. and S. Alternative methods for estimating changes in production from data on acreage and condition. Econometrica. 1958.. Findlay. 375-385. Fletcher. An application of statistical decision theory to commercial turkey production.S. P.P. Calibration of optionbased probability assessments in agricultural commodity markets.. and J. Wendell. R PR Bayesian decision theory used to aid turkey farmer deciding among: independent production. and H. J. Blisard. Calibration tests performed on corn. and S. Journal of Farm Economics. Carter. 1958. Canadian Journal of Agricultural Economics. For quarterly hog price forecasts at various horizons up to 44 steps ahead. Dean and H. Value of perfect price forecast $2700 per year. 852-868. Value of price predictor (from single econometric equation) $600 per year on expected return of approximately $4500. econometric worst. R. R. 1985. 33. 20-26. 195-201. Conditional qualitative forecasting. Quarterly and Shorter-term Price Forecasting Models Relating to Cash and Futures Quotations for Pork Bellies USDAERS Technical Bulletin Number 1482. 51.P. R TS CO Compares restricted VAR and VEC. 47. Foote. and J.P.R. Generally. Foote. Farm practice adoption: a predictive model. methods that do not use intentions data explain about 20-50% of variation. Tandem forecasting of price and probability-the case of watermelon. 1989.M.M. Craven. SN ST CO Single equations for liveweight and number of barrow/gilts and sows both by quarter and by month (each time period estimated separately). H.C. change of direction of price) constructs a composite based on Bayesian updating of probabilities. Jr. R. R LA Annual 57 equatton recursive model covering seven livestock products. 29) $6 eventually the most satisfactory results may be obtained by some combination of the methods”. FAO Monthly Bulletin of Agricultural Economics and Statistics. unrestricted and Bayesian VAR and univariate models of three cattle prices (for different animal weights) and corn price. Hsiao restricted VAR has generally lower RMSE than Kaylen’s method (Kaylen. A sector model-the poultry industry of the U. Journal of the American Statistical Association.G. 3. soybeans and hogs contracts. expert and ARIMA methods made more correct turning point forecasts than composite. P. M. 22. Craven and R. estimated by 2SL.H.J. Weingarten.W. Forecasts prepared 12 months later suggest opposite conclusion (based on footnoted actual values). J. Sought to determine demand-supply simultaneity. SN TS CO For qualitative forecasting (e. 1953.A. ST Annual 12 or 11 equation models estimated from 1915-1940 data by OLS. R OU CO Compares agency forecasts with ARIMA model. Notes (p. King. 155-162. OU Reviews history and accuracy of outlook work in US and Canada. American Journal of Agricultural Economics.R. Uses four observable binary farm or farmer characteristics (e. made between 1949 and 1952. Estimating VAR models under non-stationarity and cointegration: alternative approaches to forecasting cattle prices. 44-61. PR A segmentation (classification or configural) method. SN Specifies two equation OLS and probit model. 17. 18-28. 73-83. M. 49. TS CO Compares two VAR approaches with ERS forecasts of CPls for 5 meat groups. R SN Based on graphs of forecasts 1-6 months ahead. 35.g. Forecast criteria include probability prediction accuracy interval. Fackler.. 1985.
The pricing efficiency of agricultural futures markets: an analysis of previous research results. Gellatly. J. New Directions in Econometric Modeling and Forecasting in U. naive. Forecasting N.. Journal of Farm Economics. 123-129. Chapter 17. ture and features of ten large scale macroeconomic models. Garcia. two absorbing) constructed from 1061 and 1966 censuses for Canada. Grain market outlook (Economic Council of Canada.. 1968. Fortenbery and G. 198X. Bayri.. . Roy and G. lOYY122.. simulated trading in the market using the best-to-date model gave small profit relative to variance. 1978.S. J. Gallagher. L. R SN Discusses difference hetween standard error of estimate and standard error of forecast. Lcuthold. G. P. An evaluation of outlook information for Australian agricultural commodities. Fuller.A. Frccbairn. No forecasts. Garcia.W.. Storey. P. J. 1953. J. 19X6. 503-545. LA A 15 equation price and farm income determination model of livestock and crop food products. S. with updating of models. Sadler. Kuh (editors).L.. Furniss. Pricing efficiency in the live cattle futures market: further interpretation and measurement. pp. Effects of changes in the level of U. 1989. Fox. 1975. Trend models of feeder. Expert and its combinations more accurate but see Revel1 (IYXI) and reply by Gellatly. G. C.R. DC. Uses logit analysis. 64-78. 54.. 2944314. U. ou. Walker. J. R TS OU CO Reviews techniques.W. Atkinson.R.R. North Central Journal of Agricultural Economics. Review of Marketing and Agricultural Economics.M. current USC. LA Annual 1X equation model. 676-68X. beef imports. New York).S. Review of Marketing and Agricultural Economics. Quebec and Ontario.. P. MK A meta-analysis (50 referenccs) of 3X studies.W. generally has lowest MSE and futures price generally highest. Freebairn. 16. 47.A. Duesenberry. Review of Marketing and Agricultural Economics. Forecasting for Australian agriculture. and L. 1079.G. slaughter and wholesale beef prices. 101 pp. 22 pp. 19. Gertel. Australian Journal of Agricultural Economics. Hudson and M. Structured Models and Automated Alternatives For Forecasting Farmland Prices USDA-ERS Technical Bulletin Nunber 1824. Rausser and H. Fromm. A submodel of the agricultural sector. Projecting Canadian dairy farm structure using Markov processes. 323-340. Compares naive or univariate forecasts with I year ahead Bureau of Agricultural Economics (BAE) outlook forecasts for ten Australian agricultural commodity prices and ten outputs over 8 years.. Freebairn. R PR Transition probability matrix for number of cows per farm (eight states. Washington.G. No forecasts. in J. Furtan. R. Food and agriculture sector linkages to the international and domestic macroeconomies. The Brookings Quarterly Econometric Model of the United States (Rand-McNally. 1973. No forecasts. Implications to and from economic theory in models of complex systems. Fox. ST A four equation sectoral model. 797-812. Gellatly. W. LA Annual 20 simultaneous equation model of the fed and non-fed beef. SN TS CO Compares single equation. R. 8.K.. American Journal of Agricultural Economics. A dynamic quarterly model of the beef and pork economy. future prospects and economic benefits from more accurate outlook.S. 53-90. Compares standard errors in 30 prices with actual for years 1947-1952. 41. EV. 1976. American Journal of Agricultural Economics. Ottawa). Klein and E. Southern Journal of Agricultural Economics. K. Quarterly prediction models for live hog prices.. Rausser.A. 8. 1973. 1073. Gustafsson. Fox. Factors affecting the accuracy of price forecasts. 1965. 46.J.. Review of Marketing and Agricultural Economics. TS Single equation spectral analysis.F. K. 70. ARIMA. and B. T.W.W. 43. Canadian Journal of Agricultural Economics. Simple average composite of ARIMA and econometric with latest 72 observations. pork and chicken sectors.. lY82. W.C. C.A. each quarter separately estimated.Y. Rausser (editor).S. pp. 259-271. Southern Journal of Agricultural Economics. BAE more accurate for seven prices and nine quantities. 57. Referenced by Arzac and Wilkinson ( 1979) as the “most complete econometric study of the livestock sector”. M. Journal of Farm Economics. composite and futures price forecasts 1 to 6 months ahead. Freebairn. Franzmann. 1544 174. ST Specifies eight equations estimated by single equation methods. and G. Revised ARIMA is best. 1993. Chicago). 35. However. X1-94. Gray and G. T.S.. K. Chapter 12.C.H. Discussions by Christ and Rausser. l-19-130. American Journal of Agricultural Economics. de Gorter. 5077512.W. R LA Reviews the three generations of large scale econometric models.F. Some estimates of supply and inventory response functions for the cattle and sheep sector of New South Wales. difference between independent variables known and forecast. K. 1. G. and G. 127-130. J. 409-464. R OU EV Information required. Describes a quarterly third generation model with X7 equations covering three crop and six livestock groups. R. Allen I International Journal of Forecusting 10 (1944) 81-135 Fromm. Ladd. in G. heef production: an evaluation of alternative techniques. expert and pairwise combinations. C. TS CO Estimates and compares two new AROMA models with those in Gellatly (1979) in response to comment by Revel1 (1981). Forecasting NSW beef production: a reply. 40. Freebairn. lY61. ARIMA. 1988.120 P. LA Review of strut.A. Corn yield capacity and probability: estimation and forecasting with nonsymmetric disturbances. Wailer. Quarterly.L. SN Shows difference in point and interval forecasts when nonsymmetric (y) distribution used compared with normal. 1972. Council of Economic Advisors.. Agriculture (North-Holland. American Journal of Agricultural Economics.W. SN TS CO Compares econometric. 1622169. 1975. TS CO Compares ability of OLS and univariate methods to detect trend reversals and the performance of several univariate methods at I year and 2 year ahead forecasts in the 1Y73- Foote. and R. 1981. Sarassoro. 20. 55. An appraisal of deficiencies in food price forecasting for 1973 and recommendations for improvement.
. H. LA. Makes successive pairwise comparisons of naive no change forecast. (Uses all pairs and all three models. 1974.. 1991. 1976. Allen i International Journal of Forecasting 10 (1994) 81-13. 1962. Used the visual impact (visual counter) method on a sample of 39 small-scale farmers to get distributions on future prices and yields of their rice. Irwin. and S. and E. American Journal of Agricultural Economics. 54. K. and L. 65. 74-82. and R. B. Australian Journal of Agricultural Economics. 197-204. Uses three criteria: accuracy improvement (Theil’s revision ratio statistic. W. 44. Harris. For the hold/sell investment decision from various points in 1973-1987 to a 1990 horizon.M. Granger. multivariate state space and trend autoregression on OLS residuals. Kellogg. Quebec City. Southern Journal of Agricultural Economics. 1992. G. Most price analysis models not used for forecasting. Journal of Econometrics. Composite of three models best. Journal of Agricutural Economics. 1-13. For both RMSE and MAPE. 1986.M. R). 65. Haidacher.E. soybeans. Model selection among alternative steer price forecasting techniques. 139-149..S. Can economists accurately predict commodity prices. R OU CO Analyses 1100 crop production forecasts for barley.. Harlow. 15 or 20 years).L. 1967. oats. K.) Ridge regression composite generally has slightly lower RMSE and MAPE than simple averaging.E. absolute forecasting error and bias.H. 87-97). corn.D.. Composite rorecasting methods: an application to Spanish maize prices.) SN TS CO Same data and results as Harris and Leuthold (1985). 264-271. Gunnelson. and B. A spatial model for analysis of the Canadian livestock-feed and livestock product sector. American Journal of Agricultural Economics. Eight post-sample one step ahead forecasts had ratio of unexplained to total variation of from 0. 86-99. ST EV A critique. and G.. A comparison of . TS Uses Bessler and Brandt (1981) hog data to 121 show that composite formed by unrestricted regression is most accurate. but outlook reports have significant ability to forecast large price changes. 1981. M. Department of Economics. compares 15 agricultural and 13 non-agricultural annual series for best length of data series (10.5 1987 period. and R. Graham. spring wheat and winter wheat for 1929-1970. then trend autoregression. 22. 8. Attempted validation. 25. West Lafayette. Jr. Proceedings of the 1974 CAES Annual Meeting. pp. For turning point forecasts. 3. 1985. 1985. Harris. Leuthold. A.C.J. Giles. Winter. Farmers’ subjective probabilities in Northern Thailand: reply. Gerlow. 1972. Journal of Business and Economic Statistics. initial and revised USDA forecasts. 1987. 52. Farmers’ subjective probabilities in Northern Thailand: an elicitation analysis. The future role of the agricultural economist in outlook preparation.. OU EV Based on survey of 15 agricultural economics departments. 1983.. 149-152. MK Similar to US findings of Tomek and Gray (1970) and Leuthold (1974). 24. 1977. 1974.G. Gruen. ST PG A ten region interregional programming model.D. 67. OLS.K.A.P.A. Albisu. advantages and disadvantages.D. then OLS. 814-819. Haidacher. 235-253) had lower forecast RMSE of quarterly cattle prices than univariate ARIMA which was better than a (random walk) time-varying VAR model (see Wolff. 1983. Analysis of the accuracy of the USDA crop forecasts. B. C.M. OU CO Compares forecasts of hog and cattle prices 1974-1989 by USDA and three land-grant universities.C. (Abstract in American Journal of Agricultural Economics. soybean and peanut crops. J. R SN Using linear or exponential trend regression. ARIMA and econometric model forecasts. but for the others the reverse is true. 31 July-3 August 1983.08-0. R PR Believed to be the first economic study to use substantial monetary rewards (up to 1 day’s pay) as incentives to elicit accurate subjective probability assessments.M. and E. W. Australia 1965 to 1980. 209-223. Southern Journal of Agricultural Economics. TS CO A six-variable gradual switching VAR model (see Tsurumi et al. 1970. F. American Journal of Agricultural Economics. Australia. E.. Kellogg. Futures prices as forecasts of commodity spot pnces: live cattle and wool. American Journal of Agricultural Economics. SN TS CO Compares five forms of composite based on exponential smoothing. Gold. MSS. and J. Grisley. Grisley. 21. 1985.93. 11-22.R. Some suggestions for developing new models from existing models. Ramanathan. Clayton. OU EV Describes types of forecasting. et al. Harns. Improved methods of combining forecasts.. presented at the International Symposium of Forecasting. fewest wrong indicators are issued by VPR. 639-645.S. 1993. Goodwin. Canadian Journal of Agricultural Economics. Matthews. Journal of Forecasting. R. ST A six equation quarterly model of US hog industry. EV. Forecasting cattle prices in the presence of structural change. PR In reply to Knight et al. Long Term Projections of Agricultural Supply and Demand. From backcasting towards forecasting. 2. 1185. Monash University. IN.. Goss. Dobson and S. 13 pp. New York. Goddard. University outlook programs: a review and some suggestions. longer series are better. not forecasting. Naive random walk model produced lower RMSE errors than any outlook forecast. admits that the linear scoring rule used can be improper but (1) is easy to communicate (2) not necessarily improper for risk averse individuals (as these were) and (3) is less important than getting the individuals to take the elicitation process seriously. For agriculture series. 85 pp. Omega.D. J. Gil. D.W. Pamperin. simple averaging always as good as or better than any other composite. varying parameter regression ranks best. potatoes.A. Factors Affecting the Price and Supply of Hogs USDA-ERS Technical Bulletin Number 1274. W. 1984. paper presented at the American Agricultural Economics Association meetings. 5. R. tobacco. 15 or 20 years) to make long-term forecasts (10. 51-74. Review of Forecasting in the Economic Research Service USDA-ERS.
38-45. Hendricks. Hacklander. Apple price and production forecasts for Maine and the United States. 45. 2977309. in G. Quarterly Review of Agricultural Economics.5.. 0. Jr.F.A. Journal of Farm Economics. Naive based on present price was next best. 4333446. ST A five equation annual model that uses both dynamic and static simulation within-sample and assumed values of exogenous variables to make forecasts.. Proceedings of the 1974 CAES Annual Meeting. Hayenga. Hedley.K. G..G.. Reports mean forecast. Andersen. D. 46. 1957. 1977. 60-69.122 P.. W. 9. and W. Risk neutral and risk averse producers are better off from using the forecasts to create or liquidate selective hedges compared with routine and no hedging strategies. 40-50.A. ST Annual 17 equation model of the broiler and turkey industries. Hoffman. 1927.G. ST Annual six equation model used to produce one forecast. 1963. Heicn. Comparison of livestock price forecasting using simple techniques. ARIMA. lY76.S. The effect of quarterly livestock reports on cattle and hog prices. and between fruit count per plant and average weight per fruit based on surveys in 1954 and 1955. poultry sector. SN. and outlook information. W. R LA Describes ORANI model simulations. D. OU Forecasts of eight commodity prices 5 to 12 months ahead (depending on commodity) collected from about 200 Iowa farmers. Objective forecasts of cotton yield. ST A five equation system. R. M. lY7. 5X. Huddleston. Western Journal of Agricultural Economics. forward pricing.. A linear programming livestock-feed grains model. 69. Voss (editors). Hinchy. pp. and D. . and J. D. 1970. Quebec City. R SN Uses multiple rcgression to forecast prices of fat cattle 6 months ahead. G... 311-316. then reduction in average forecast error from 3% to 1% would increase aggregate economic surplus about 6% of gross value of crops and about 1% of gross value of livestock products. Journal of Political Economy. 1508-1513. Hendricks. with four quarters of total egg production forecasts.R. R OU Describes relation between counts of different kinds of fruits surveyed on 1 August and final numbers of bolls. Houck. A statistical model of the demand for soybeans. pork economy. Canadian Journal of Agricultural Economics. 62. North Central Journal of Agricultural Economics. 156-168. Utilization of institutional and quantitative analysis in outlook preparation: some management considerations. Hopkins. actual price. OU EV Cash prices affected by information in report but not futures prices. No forecasting. Readings in Egg Pricing University of Missouri-Columbia. and D. Hoffman. and H. 366-374.S. M. farm commodities. SN TS CO Similar to Brandt (1985) but with monthly data. composite. 370-375. Journal of Farm Economics. No validation. 5.S. Relations used for a state by state forecast of cotton yield in 1956 in a ten state region. Australia). R TS Spectral techniques for lead-lag relationships. 268-276. 34-47.P. American Economic Review.. D. An economic analysis of the U. An econometric model of the U. 1985. pp. Rogers and L.. 1978. Discusses problems of running a forecasting program. 1064. 57. Helmers. 1974. 1972. ARIMA.A.J. Huff. Skinner. Hayward. Hauser.T. Heady. ST Annual five equation potato model. American Journal of Agricultural Economics. current position. Holt. 1987. Forecasting cattle prices. Brandt.D. 31. American Journal of Agricultural Economics. mean error and distribution of individual forecasts. TS CO Compares three forms of naive forecast. Hayami. X3-105. 535-544. Tests for predictability of statistical models. Combining price forecasting with hedging of hogs: an evaluation using alternativ. 1970. USDA outlook has generally smallest bias and least MSE. 145%150. Irend regression.. 7.A. R OU Describes forecasting in Agriculture Canada. North Central Journal of Agricultural Economics. MP 240. Also probability forecasts for corn and hog prices. Journal of Futures Markets.J. G.B. I lY-130. Allen I lnternutional Journal of Forecusting 10 (1994) Xl -135 alternative forecasting techniques for livestock pnces: a case study.B.A. I. 2. Compares econometric.J. 2. and H.A. with 4 years of forecasts. R. Journal of Farm Economics. 1980. 1971. futures price and two expert forecasts 4 months ahead as guides for fattening steers and hogs. 1479-1484.P. Southern Journal of Agricultural Economics. Journal of Farm Economics. P. 1984. Henson. Peterson. 1986. Held.O. Hce. and L. Social returns to public information services: statistical reporting of U. TS CO Comparison of forecasts of monthly variance of daily November soybean futures prices finds ARIMA most accurate then econometric then naive model.e measures of risk. ST PG A seven region static feed grain-beef interregional programming model. 75-87. E. 1985. PR Uses inventory adjustment model. If decision makers adopted USDA production forecasts. Hedging with options under variance uncertainty: an illustration of pricing new crop soybeans. Quarterly egg production estimators. Melbourne. Expectations and errors in forecasting agricultural prices. and H. Adaptation and Survival in Australian Agriculture (Oxford University Press. with 4 years of forecasts. 9. J.. ST Annual seven equation model.B.L. M. Monthly supplydemand relationships for fed cattle and hogs. 1577160. Review of Economics and Statistics. Kaldor. J. American Journal of Agricultural Economics. l<S-160. OU Definition. 1954. composite and seasonal index forecasts of hog price 2 to 10 months ahead. Harvey. Use of predictive equations to forecast monthly average New York egg prices. H&n. 32.. 48. and D. Forecasting yields with objective measurements.R. GLS and multivariate ARMA models of cattle and hog prices. Y. 62. 13. Huff. moving average.A. 20-25. models. Northeastern Journal of Agricultural and Resource Econorrucs. Higgs. 320 pp. Criner and S. Agricultural Economics Research. W.K. 1966. The relationship between beef prices in export markets and Australian saleyard prices. with 4 years of forecasts. SN TS CO Compares single equation OLS. R. 52. ST An eight equation quarterly US model.
Jaffrelot. Hussey. D. in Agriculture Canada. OU CO Compares four different hog price outlooks and three cattle price outlooks one. A Short Term Projection Model for Wool Prices-An Explanatory Analysis Bureau of Agricultural Economics Occasional Paper Number 7.12%. Huff and G. coordinated by Z. 55. Discovering production and supply relationships: present status and future opportunities. Rausser.S. 1978. Huddleston. Demand for Food USDA-ERS Technical Bulletin Number 1821. Allen I International Journul of Forecasting House.C. 31 l-348. D. In post-sample forecasts. (I) worst for wheat. in L. Australian Economic Paper. Recommends structured representation: incorporating economic theory. Applied Economics.and cross-price and expenditure elasticities from a first order differential linear-inparameters system of 39 food categories and non-food using 1953-1990 annual data. 38. Jarrett. R LA Annual third generation model with many linkages. H.. 11-40. Hassan and H. Description and Use of a Macroeconomic Model of the U. Japanese and American Agriculture: Tradition and Progress (Westview Press). median 4. Journal of Farm Economics. 1983-1986 or 1987. 1185)..C. L. Chapter 16. number of ears forecast from stalk count. Johnson. Objective methods in forecasting components of corn yield. R OU On 1 August. ARIMA. R OU Origins. objectives.38%-9. Commodity Forecasting Models for Canadian Agriculture. 1993.A. Most of emphasis is on ways to improve estimation of supply response.G. Commodity Forecasting Models for Canadian Agriculture. 1972.J. implemented from 1977 onwards. Ingco. and M. ST National 16 equation model. median 3. 1965. Institutionalizing a large scale econometric model: the case of Agriculture Canada. An evaluation of a combination of quarterly and annual models in predicting cattle and hog prices. A forecasting model for food and other expenditures. S. A model for forecasting provincial hog marketings. 1979.0X%-7. 1989. 1990. Ottawa. two and three quarters ahead for ability to predict turning points using regression-based Merton test. R. 61. New York). LA. Review of Marketing and Agricultural Economics. 1992. S. Huang..C.O. coordinated by Z.-R. 1980. F. Chapter 23. 1983. The market timing value of outlook price forecasts. K. 65 (December 1983).1995 forecasts.. Vol.R. Short-term forecasting of Australian wool prices. S. Irwin.60%.B.. and G.A.B.H. K. Huff. 933102. Departmental Technical Report Number DTR 80-5. Tweeten et al. EV Contains comparison of annual acreage forecasts for US wheat and feed grain from (I ) acreage response equation. Hudson. presented at the American Agricultural Economics Association annual meetings. presented at the annual meeting of the American Agricultural Economics Association. Huang.D. Agriculture (North-Holland. (Abstract in American Journal of Agricultural Economics. but only the hog price outlooks are significantly better than the forecast from a trend plus seasonal dummy variable regression. 21. 18 pp. imposing in csti- . Three hog and one cattle price outlooks at one quarter ahead 10 (lW4) 81-135 123 have value in directional indication. (I) worst for feed grains. American Journal of Agricultural Economics.R.S. (editors). 70 pp. pp. A Complete System of U.B. Elward. 43-60. Gerlow and T. Hughes. Manhattan.F. For 39 food groups.. OU Uses indicator analysis. Rausser (editor). Canada. Economy which Et?$hasizes Agriculture Texas A&M University. Hassan and H. 49-53. Composite forecasting: some empirical results using BAE short-term forecasts. 1235-1246. These were more accurate than standard quarterly models and ARIMA models. 603-614. 68 pp. 69.S. (3) is most accurate.. and J. pp..W. Canadian Journal of Agricultural Economics. M. 1993. Johnson. Use of outlook in Canadian agriculture. LA Estimates by a constrained maximum likelihood method own. Discussion. R EV The profession has been too occupied with flexible functional forms and duality analysis. 7812. J. 51-73.P. Penson. New Directions in Econometric Modeling and Forecasting in U. West Lafayette. Just. Percent RMSE and MAPE calculated from withinsample static simulation over the whole estimation period. publication no. (2) best. Furniss. MAPE 1.E.B. Kansas. Wong. A quarterly forecasting model of the Canadian egg industry. 7812. and J. Canada. Ferris.S.E..28%. 133-136.E. Agricultural Economics Research. R LA Notes that the model. 1. Huff. American agricultural supply. Canadian Farm Income: Medium term outlook and value added accounts. 1977. RMSE values range 1. (2) acreage response equation with exogenous program participation rate variable and (3) system of acreage response equation and logistic participation rate equation. H. 1991. pp. R TS Monthly forecasts of six grades of wool using Winter’s multiplicative model compared with simple moving average.. Referenced by Labys and Granger (1970. and I. OU Gives 1990. 61-75. SN Set of nine cobweb-type annual single equations.C. CO.. IN. R. Huff. Ottawa. pp. Liu.06%. p. 48. 1958. Weight of grain forecast from number of ears in 60 feet of row. Department of Agricultural Economics.G. 1978. SN TY CO Annual demand equations for cattle and hogs combined with quarterly ratio models to produce quarterly models. naive and four composite methods. 220) as the only published paper “in which exponential smoothing is applied directly to a commodity price series”. LA EV Reviews the second generation models of Chen (1977) and Roop and Zeitner (1977). 801-830. Jolly. C. in G. was intended for both forecasting and policy analysis.S. and M. in Agriculture Canada. H. 10. 11601167. S.B. M.F. Just. Review of Marketing and Agricultural Economics. 4. Peckett. Vol. 1087. R ‘I’S CO Compares citrus and sugar production forecasts from BAE ([Australian] Bureau of Agricultural Economics). Jones. 1982.. 1. W. publication no. 1966. Forecasting corn yields: a comparison study using 1977 Missouri data USDA-ESCS.
37. Becker. Just. poultry) sectors. Preliminary report on objective procedures for soybean yield forecasts. Kelly.G. lY80. compares with standard turning point definition. 2544277. A model of the feed grain sector of Canadian agriculture. 5 pp. 22-32. policy-oriented model of the feed grain. Zilberman. OU Summarizes outlook. but no equation estimates of a highly structured. The key criterion for SUCCC~S is the ability to represent out-of-sample phenomena. evaluates results. Eyvindson. G.W. Vector autoregression forecasting models: recent developments applied to the US hog market. pp. Schmitz (editors). A review of three research prograrns in quantitative modeling in the Bureau of Agricultural Economics. 224-247. and G. National Agricultural and Resources Outlook Conference. Stoeckel. Approximately 90 references. A comparison of multivariate forecasting procedures for economic time series. Kingma. describes methods. Forecasts from state-space and single equation partial adjustment models. M. and K. North Central Journal of Agricultural Economics. Konyar.S. For three previous studies. A framework for comparing the efficiency of futures markets. Chapter 18.5 mation all of the economic principles and practical information that is otherwise considered in evaluating plausibility of results.. R OU LA CO Compares futures prices of eight commodities with USDA and four commercial forccasts. Kelly. 1. 1973. structural GE systems: the ORANI module of the IMPACI project. depending on number of steps ahead forecast. mid 1970s (wool.B.C. A state-space approach to perennial crop supply analysis. Farmers’ subjective probabilities in Northern Thailand: Comment. EV Notes failure of agricultural economists to detect price changes in early 1970s and early 1980s.. J. Dcscribes structure. 73. P. 147-148. 1981. American Journal of Agricultural Economics. Dynamic regional analysis of the California alfalfa market with government policy impacts. Shonkwiler. Quebec City. Proceedings of the 1974 CAES Annual Meeting.12‘4 P. Rausser and D. B. 1992. 415-416. S. also for a macroeconomic model and an oil model. soybean.L. Kofi. 1985. Western Journal of Agricultural Economics. K. Supply and demand projections for grapes. 1973). Kling. 343-3. in T. 841-849. 31-50..W.. 63. national and international. Canberra 1992. 55. quarterly. 1974.S. Australian Journal of Agricultural Economics. R. Allen I International Journal of’ Forecasting 10 (1994) XI -13. 197-208. 70. 8.S. Gellatly (1979)): modeling macroeconomic systems. J. Kerr. 1515-1520. pork. An assessment of the agricultural economics profession. and J. 1990. 1963. 74. 1177-l 190. Proctor. 1089. 1992. A note on qualitative forecast evaluation: comment. Just. M. Proposes new approach and compares six VAR methods. 71. Finley. K. and R. R TS A more precise definition of a turning point criterion than Naik and Leuthold ( IYSh). 5-24. 67. 1986. R. Knapp. T.E. American Journal of Agricultural Economics. Kaylcn. N. medium term (5 year) projections. ST 25 county/region demand and acreage regressions used to forecast directly and in spatial equilibrium model. Knapp. R TS CO Reviews the exclusion-of-parameters and Bayesian approaches to reducing VAR parameters. Model is estimated with annual 1962 to 19866lY87 data. Perennial crop supply response: a Kalman filter approach.C. Agricultural Economics Research. OU Describes Ezekiel (1953). TS CO Comparts six VAR methods with exponential smoothing and autoregressive methods for the hog market. single equation aggregate State or national models. T. S84-594. American Journal of Agricultural Economics. Gray and A.R. Just. ST PG Interregional programming.. 1955. Policy analysis. validated by dynamic simulation within sample. 24. 70. MK Cash price regressed on previous futures price.S. A note on the forecasting of turning points. Brandt. Konyar. Journal of Farm Economics. and welfare surpluses presented for 1985-1994 baseline (no policy change).C. Probability sampling in collecting farm data.. R EV Covers modeling production systems (LP). 1988. 45. Kaylen. 368-370. and G. 1985.A. and D. 157. modeling commodity markets: annual. R OU Study in which objective was solely to predict number of pods. Knight. O. R TS CO Modifies turning point criterion to differentiate peaks from troughs.A. multi-enterprise system. Kohn. OU Describes USDA-Statistical Research Service methods. R. 701-712. M. . and K. American Journal of Agricultural Economics.. International Journal of Forecasting. PR Criticizes Grisley and Kellogg (1983) for use of an improper scoring rule in eliciting subjective probability distributions. 9. American Journal of Agricultural Economics.C. 139-141. forecasting ability (Freebairn (1975). American Journal of Agricultural Economics.A. livestock models) multiequation.E. 1992. progress: 1960s. 1957.G. (Freebairn. No consistency of results across data sets.T.. The result is models with globally plausible functional forms and implications. IYSX. Germany).K.L. 1991. Rausser. B. TS State space model for alfalfa supply.S. wheat and livestock (beef. and J.C. Agricultural outlook work. Johnson and R. Bourke (1079). EV LA Contains same comparison of acreage forecasts as Just (1992). and W.E. Rausser. Kiel.. TS CO Graphical comparison only of within-sample one step ahead forecasts of annual total acreage of Florida grapefruits. uncoupled direct payments (phaseout of target and support price payments) and environmental protection (pesticide restriction) scenarios. J. Improving Agricultural Trade Performance Under the GATT (Wissenschaftsverlag Vauk. Environmental and Agricultural policy linkages and reforms in the United States under the GATT. pp. Kaylen. Journal of Farm Economics. Kelly. Commodily price forecasting with large scale econometric models and the futures market. no forecasting. Australian Bureau of Agricultural and Resource Economics. 156-158. Bcssler. American Journal of Agricultural Economics. American Journal of Agricultural Economics. T. R. Hypothesis test.52. Kalaitzandonakes. Longmire and A.
N.C.C. 90-97.C. R OU SN EV CO For six grain crops.M. Larson. 24. Kulshreshtha. An integrated econometric model of the Canadian livestock-feed sector. S.. MK Rational price formation (where prices for contracts reflect average cost of production) is generally supported by distant live cattle and live hog futures. MK. Journal of Farm Economics. time series methods and naive for six agricultural commodities. 1981. Purcell. 233-249. not for information transfer. Makes forecasts for 19811985. 1973. Livestock futures markets and rational price formation: evidence for live cattle and live hogs.. 28. Agricultural History. in Agriculture Canada. Aggegate Food Demand and the Supply of Agricultural Products USDA-ESS Technical Bulletin Number 1656. R. Wilson.N. S.. 1984. A Comparison of Alternative Approaches to Forecasting Cattle Prices in Canada. UK). R. 57. 12 and 36 months) when variables updated. S. The Bureau of Agricultural Economics outlook program in the 1920’s as a pedagogical device. The U. 54. The agricultural component in macroeconomic models. A.A. F. 12 region mathematical programming model and documentation of its code (written in GAMS and fully self-documented with 1990 data embedded). and A. Labys. Medium term world wheat forecasting model. W.. Predicting regional net marketings of beef cattle in Saskatchewan. and Y.G. Hughes.R. Huff. Virginia Polytechnic Institute and State University.N.W.. 1971. Lakshminarayan. 1975. 1. Kulshreshtha. 25. Hassan and H. The hog cycle as harmonic motion. ST An eight equation sectoral model. Koontz. 97-104. R LA ST EV Naive or no change models superior to econometric. J. No results. Journal of Farm Economics. K.A.. Ottawa.E. TS Theory. The Impact of Hog and Pig Reports on Live Hog Futures Prices: an Event Study of Market Efficiency Department of Agncultural Economics Staff Paper SP-84-11.N. 252-261.A. 1990. 46. 1972. 1955. Wilson. Jr.J. Koontz. Kulshreshtha. Lakshmanan. Fisher. 69 pp. I10. 23. 1977. 1987. ST A ten equation annual sectoral model..N.A. and A. Composite method most accurate and has best turning point performance for all horizons (6. 1964.C. Kutish. Use of planting intentions to predict actual plantings. Aldershot. 18 pp. Chapter 9 compares forecasts from econometric. Kulshreshtha. 13-32. pp. Department of Agricultural Economics Technical Bulletin 82-01. and C. Canadian Journal of Agricultural Economics. Chapters 3-8 develop a general model of price fluctuations for both cash and futures markets. 68-72.D. McCormick and T.. Also reviews the international models of several commercial macroeconomic forecasters. 290 pp. 375-386. 1970. Kulshreshtha. 1993. R OU Outlook as a teaching device. But after feeding commitments are made. American Journal of Agricultural Economics. Canadian Journal of Agricultural Economics. W. 41-62. S. 1979. R.G. Kongtong.. North Central Journal of Agricultural Economics. 37. LA An 82 equation annual model of seven livestock products and eight feeds. Commodity Forecasting Models for Canadian Agriculture. OU To increase accuracy of livestock slaughter forecasts. R TS CO Compares five approaches each with and without variable updating for forecasting monthly slaughter steer and feeder steer prices. 87-106. LA Describes the nine crop (plus conservation reserve program). W. S. ST A nine equation annual model. ST. Agricultural Economics Research. coordinated by Z. Lattimore. market prices reflect expected market conditions. R LA A survey of the individual country models in project LINK. Hudson and W. Canada. and A. Hedging and Commodity Price Forecasts (Heath Lexington Books. 873-878. Kost. including a tabulation of the number and type of equations and variables in the 25 overall models and in the agricultural sectors. Primary commodity markets and models: an international bibliography (Gower Press. Allen I International Journal of Forecasting 10 (1994) 81-135 125 Konyar. August 1984. Southern Journal of Agricultural Economics. S. American Journal of Agricultural Economics. 64. Spriggs and A. Rochette. 33. The problems and challenges for international commodity models and model builders. A monthly price forecasting model for cattle and calves.W.P. 21.N. 25. TS Monthly model of the Canadian broiler industry. Reimer. Box-Jenkins model for the broiler chicken industry. Canadian Journal of Agricultural Economics. Canadian Journal of Agricultural Economics. Labys.. SN Single equation monthly estimations for prices at five locations. Kunze.B. Rosaasen. S. S. publication no. 1975. S. but generally poor when variables not updated.G. 1978.A.G. Speculation. Akinfemiwa. Canadian Journal of Agricultural Economics. Kulshreshtha. Lexington. Kulshreshtha. Agricultural Resources Model (USARM) USDA-ERS Staff Report AGES 9317. 20. ST A 26 equation monthly model of the Canadian cattle sector. S. A short-run model for forecasting monthly egg production in Canada. ST Econometric annual .R. 36-46. Papineau and R. 1982. J. Hudson and M. Canadian Journal of Agricultural Economics. slaughter at national level and in seven regions/provinces. and R. 1972.. quarterly reports should contain breeding intentions data by month. 1-13. 34-45. Zwart.. and E. Vol. MA).. G. no empirical analysis. TS CO Chapter 2 introduces spectral analysis.W.B. R. 10501053. M.. 321 pp. 7813. S. I. Ladd. 23 pp. Labys. Osborn. 1981. Kulshreshtha. 1992. 1980. Granger. M. ST A 12 equation quarterly sector model. Intended for policy analysis not forecasting. Lamm. and C.L.S.W. University of Saskatchewan.N. use of intentions data combined with ‘objective’ data (price.N. Needed changes in state and local crop and livestock reports. 84-91. An open econometric model of the Canadian beef cattle sector. LA Highly aggregate annual third generation four equation model. Canadian Journal of Agricultural Economics. Ng. expected yield of crop and competing crops) better than either alone.D. and K. 1977. 2. An econometric analysis of the Canadian egg market. A harmonic analysis of cattle and hog cycles in Canada. W.-F. 19.
On the use of Theil’s inequality coefficient. Journal of Farm Economics.M. R LA A 44 equation recursive system. S.R. F. Biases statistically insignificant. Journal of the American Statistical Association. 1954. 31. 61. L’Esperance. Watts. Calculates MSE treating the delivery price as actual and the futures price or cash price up to 36 weeks prior as forecast. lowering the per acre value of information.. R ST PG Recursive three region (US.euthold. MacCormick. American Journal of Agricultural Economics. R OU Reports distribution of US production forecast error over 3X years. lY62.M. pp.Ci. Extensive test battery and dynamic (one step ahead) forecast validation applied within-sample. ST PG Similar to Martin and Zwart for hogs. MacLaren. P. 1063. A case study in prediction: the market for watermelons. Japan). R. 36. SN CO Compares the predictive performance of five econometric methods for five wholesale meat prices. Loyns. 151-164. H.M. Errors decrease with successive seasonal forecasts. Chin. YO-107.B.. Hartman. An analysis of daily fluctuations in the hog economy. 65. American Journal of Agricultural Economics. Econometrica. Huff. 1070.. 731-748. 612-624. 1983. North Central Journal of Agricultural Economics. OU Reports MAPE for forecasts of annual value of exports of 1. Forecasting wholesale price of meats in the United Kingdom: an exploratory statement of some alternative econometric models. The price performance on the futures markets of a nonstoreablc commodity: live beef cattle.A. Withinsample policy simulations. Use of errors of prediction in improving forecast accuracy: an application to wool in Australia.. see Foote ct al. l.S.. Decomposition of the beef and pork cycles. 1974.A. LIU. EEC3. Lcuthold. 4X2-4X9. 1978. 22. slaughter and price. 71-X0. R. 12X.. Australian Journal of Agricultural Economics. D. turkey and chicken export quantities and domestic retail prices linked recursively (i. A cross sectional and time-series analysis of Canadian egg demand.M. T.e. 1982. 57. East and West Canada) spatial equilibrium model with consumption.M. . OU A perfect weather forecast over the drying interval (two periods. Canada. 49-62. 32. Providing a forecast may raise total raisin production. IYhY..M. 196. 65. Huff and S. T. 55. lY74. MacAulay.A. A rational distributed lag model of quarterly live cattle prices. R. Variations in crop forecasts for cotton. 3. W. lY73). Lee.. August forecast accounts for X4% of annual change in production. Quebec City. MK Teats efficiency and bias of futures price as forecast of closing price by rcgrcssion using monthly data. Ottawa. Maki. and A. American Journal of Agricultural Economics.model with IX supply and demand regions. EECB. publication no. A. LA Quarterly eight equation VAR model of beef.B. 1993... Lave.3 commodity groups and volume of nine groups. 271-275). Leuthold. SN National monthly single equation model. Lowenstein. American Journal of Agricultural Economics. Commodity Forecasting Models for Canadian Agriculture. meat exports. 2. with no feedback) to a nine equation VAR model of the US and rest of world macroeconomy.A. pork hcllies (three equation. cattlc (two equation). 7X/3. Leuthold. 46 pp. J. Hartman.N. The value of better weather information to the raisin industry. A recursive spatial equilibrium model of the North American beef industry. 1975. Leuthold. The Accuracy of USDA’s Export Forecasts USDA-ERS-CED Staff Report Number AGES 9224.G. Forecasting livestock supplies and prices with an econometric model. MacAulay.. and P. Journal of Farm Economics. closing stock demand and supply for each region and exogenous rest of world trading. OU CUSUM analysis to analyse BAE forecasts of price and production. R TS CO Econometric model outperforms ARIMA and random walk models for one step ahead forecasting.G.A. Chang and W. Agricultural and Resource Economics Review. Canadian Journal of Agricultural Economics.F. R. Lu. 44. 163-173. Leuthold. Forecasting daily hog prices and quantities: a study of alternative forecasting techniques.. pp. Canada. and P.L. Econometrica..M. S3Y-547.7. SN ST CO Forecast accuracy and turning point performance better from three simultaneous equation system than in reduced form single equations. B.M. 21.Vol. IYYl. 210-221. R ST CO MK Compares forecasting performance of quarterly econometric model. R TS Points out misuse of Theil’s U.. OY%II I. SN Single equation annual or semi-annual estimations with predictions for number. Maki. 2X. 1964.. futures price and composite for live hogs (three equation). about 3 weeks) could increase a California raisin grape grower‘s expected profit by $91 per acre compared with the hest harvest strategy followed in the absence of a weather forecast. X4%X65. A semi-strong form evaluation of the efficiency of the hog futures market. R. 1977. Bui-Lan. SN Single equation reduced form price relation. Proceedings of the 1974 CAES Annual Meeting.S. American Journal of Agricultural Economics.B. Journal of Farm Economics.R. inventory dcmands for six major countries and policy intervention modelled explicitly for five regions (US. D. R. 19X1. Regressions of change in actual as function of change in forecast find some significant bias (mostly upwards) and \omc inconsistency (slope signilicantly different from one) in various commodities and regions. MSE futures larger than MSE cash for ahout IS or more weeks prior to delivery date. 1979. Naive no change model generally performs better. Schmitz and D.. 45.J. Meyers. A. l-15. A forecasting model for the Canadian and US pork sectors. and W.. Hassan and H. 51. Journal of Agricultural Economics. The impact of domestic and foreign macroeconomic variables on U. coordinated by Z.J.M. 674-680.B. An evaluation of the forward pricing efficiency of livestock futures markets. pork. 1973. Marsh. W.J. 6. in Agriculture Canada. 344-346. MK Compares futures price prediction with that from two equation (price and quantity) econometric model. W. L. R. MacDonald. SN Single equation demand (price) and supply (quantity) forecasts one day ahcad. X8-101).
L. K. Ottawa. Tabulated forecast and final estimate values permit calculation of standard forecast accuracy statistics.B. Journal of Agricultural Economics.S..A. 1987. de Gorter.G. P. C.A. American Journal of Agricultural Economics. 1982. futures market. A quarterly econometric model of the North American feed grain industry. B. Miller. 1978.J. Forecasts of pig slaughter two and three quarters ahead based on actual or forecast breeding stocks are better than naive. publication no. See Dorfman and McIntosh (1990) for results. 69.M.D.M. Projection models for U. and M. and H. Criddle and Havenner (1990) for individual methods. 1984. 18-28. Meilke. 7. Miller. MK Milonas. spill-over effects into other sectors and functional relations between labor and output. 6. LA CO Compares forecasts 3 months ahead for October prices of corn. Bessler. Canada. Zwart.D. 1977. naive. CO MK Hog futures market provides better forecast than lagged cash price.P. L. 104-117. and A. mill use. (3) three equation wheat. Garcia. The effects of USDA crop announcements on commodity prices. ARIMA. pp.T. criterion. No significant difference in forecasting performance according to Mann-Whitney (I test on either RMSE. and J. 57. 1. Econometric forecasts of pig supply. (5) 20 equation annual cattle and (6) six equation sheep models. and D. cattle and hogs. 1073. SN Single equation distributed lag models forecasts compared. 67-70. 101-105. Huff. 72. SN Weekly single equation models for east and west US for three egg sizes displayed univeral downward bias in post sample forecasts. Menkaus. 4844489.. 1993. soybeans. C. 44. 2. Southern Journal of Agricultural Economics. McClements. The response of futures prices to new market information: the case of live hogs.. A price forecasting competition: introduction. 1992. Simulation 1963-1973. 15-42.M. and R. Miller. ST Spatial equilibrium model. Dorfman. Compares US hog price forecasts using expert. Forecasting price movements: an application of discriminant analysis.H. See Berck and Chalfant (1990). and A. OU Calculates average forecast (as percentage of final) and 95% confidence interval for annual production. American Journal of Agricultural Economics. 20 (1969): 241-250. 1975. Cotton and Wool Situation and Outlook Report USDA-ERS CWS-67. Estimating slaughter supply response for U. Proceedings of the 1974 CAES Annual Meeting.K. Journal of Agricultural Economics. Hassan and H. LA Accuracy of forecasts of the Welsh agricultural sector from eight input-output tables rapidly declines as horizon increases because of insufficient attention to final demand forecasts. 73-80. J. 1981. but not cattle. Zwart. Bessler (1990). L. 216-219. 55-66. SN Quarterly single equation estimation and forecasting of three categories of cattle and two of hogs. Applied Economics. Southern Journal of Agricultural Economics. Qualitative forecast evaluation: a comparison of two performance measures. American Journal of Agricultural Economics. Journal of Futures Markets. American Journal of Agricultural Economics. Hog sector models. export.A. McIntosh. west and east Canada. 6. Another look at the hog-corn ratio. Wharton and OASIS). McIntosh. 63.135 127 Marsh. 2. B.S. 55. 1974.C. and R. McFarquhar. Midmore. A. McIntosh. . and G. Commodity Forecasting Models for Canadian Agriculture. ST Six models: (1) consumer expenditure (linear expenditure system) for 27 food plus nonfood used as final demands in (2) input-output model for 39 agricultural and non-agricultural products. 7812. R TS CO Describes their ‘little-Mak’ competition. Quebec City. 1970. pp. Jelinek. L. 1981. ST A three equation model based on author’s four equation model in ‘A model of pig supply’ Journal of Agricuhral Economics. K. University of Georgia College of Agriculture Experiment Station Research Bulletin Number 286. 2299238. 20. Martin. 1988. simple average composite of the above and four large scale econometric models (Chase. Masters.S. feeder and steer cattle and hogs. Martin. Forecasts by experts (farmer survey). 284-300. compares turning point rankings using both the ratio of accurate to worst forecasts (Naik and Leuthold. C. bayesian composite and three other composite approaches. S.. TS CO Describes composite approach using matrix of pairwise beta distributions with Dirichlet diagonal conjugate priors.R. An assessment of USDA’s cotton supply and demand estimates. N. Adams. 22. 786-787.. 1971. total use and ending stock quantities of cotton each month of the crop year (August-July) from 1980-1090 for US domestic and rest of world. No comparisons with other forecast methods. and J. Western Journal of Agricultural Economics. 209-215. ST An 18 equation system of US. Meyer. D. Meilke. R TS CO For the hog price series and seven methods of Brandt and Bessler (1981).. 321-345. and R. Martin. food and agriculture. 33 pp. The price forecasting performance of futures markets for live cattle and hogs: a disaggregated analysis.C. 1986) and the Henriksson-Merton confidence level. coordinated by Z. Forecasting agricultural prices using a bayesian composite approach. (4) two equation barley. 1992. American Journal of Agricultural Economics.H.S.D. 1990. MAPE or Theil’s li. 1979. A spatial and temporal model of the North American pork sector for the evaluation of policy alternatives. pp. North Central Journal of Agricultural Economics. 27-34. Evans. Relative accuracy of price expectations held by Georgia farmers and by other forecast sources in 1980. 209214. Vol. Skinner. Allen I International Journal of Forecasting 10 (1994) 81. and P. futures price. in Agriculture Canada. L. DRI. American Journal of Agricultural Economics. A short run price prediction model for eggs. Dorfman.R. but use of contemporaneous explanatory price variables not accounted for. Input-output forecasting of regional agricultural policy impacts. R ST PG Quarterly recursive quadratic programming model. 74.C. SN CO Annual single equation binary dependent variable model of feeder/yearling cattle price movement as function of other prices and inventory change.C.M.
Nelson. Myer. OU CO Presents MAPE of USDA outlook forecasts made one to three quarters ahead for 26 livestock and two crop commodities (15 quantities. American Journal of Agricultural Economics. 1991. R ST CO For a three equation monthly sector model with assumed autocorrelations of 0 to 0. 274-279. After about 19 forecasts. July forecasts) and also on August temperature (for August forecast) for 1892-1914 data for Georgia. beef sector. Nyankori. Myers. West Lafayette. 1972. SN CO ARIMA model used to get quarterly weights of alfalfa price (Petaluma. those subjects paid according to the linear (improper) scoring rule behaved strategically and no longer stated their believed subjective distributions. R SN Regression of cotton yield on May rainfall (for May forecast).G. Grether and J. R. pork. Western Journal of Agricultural Economics. Moore. J.S. and B. Each subject made a succession of 40 probabilistic forecasts.R. 1983.. Kogan. A Monte-Carlo comparison of alternative estimators of autocorrelated simultaneous systems using a U. Compares six methods of making annual forecasts of beef. Western Journal of Agricultural Economics. ST EV A four equation recursive model based on spectral analysis... 1986. SN Single equation. presented at the International Symposium on Forecasting. on June temperature (for June. 1953. 54. SN ST TS Page 260 is referenced in Brandt and Bessler (1984) concerning the cattle market: multiple time series models perform marginally worse than ordinary univariate ARIMA. 1962. Subjective probabilities and scoring rules: experimental evidence. SN Simple extrapolation and elasticity uses.8.. chicken quantities and prices (total six). move to probability sampling of cotton farmers’ intentions to plant and cultivated acres estimates. R PR Tests the importance of a proper versus improper scoring rule (see comment by Knight et al. Okyere. New York). A. Leuthold. 1987. G. Planning within agricultural estimates for a workable modernization program.G. American Journal of Agricultural Economics. 5. R OU EVA proposal covering the requirements for a probabilistic outlook prosurvey of user needs. Hammig. 185-193. 35. Forecasting the Yield and Price of Cotton (MacMillan. development of elicitation gram.E. presented at the American Agricultural Economics Association annual meetings. training.128 P. Western Journal of Agricultural Economics. Nelson.A. Spreen. 1980. 1917. Journal of Farm Economics. Naik. in USDA Yearbook of Agriculture (Government Printing Office. For example. Nerlove. and D. 71. accuracy and comparative merit. 17 pp.) SN CO Varying parameter models outperformed fixed.7 cents per acre. Stillman and M. pork sector model as the true structure. and S. Farnsworth. procedures. 68. No forecast performance measures. L. 615-672. Nelson. Carvalho. 1958. G. M. Variability in forecasts in a nonlinear model of the U.L. 19. but 2SLS and autocorrelation correction techniques better postsample. Journal of Farm Economics. Facts for decisions. Referenced in Harris (1976) as source of history of US outlook. scouting will continue to be used.. 68. The first forecasts of the crop year are more important than later ones..J. 855-864. R TS Describes 4 x 4 contingency table. Combining annual econometric forecasts with quarterly ARIMA forecasts: a heuristic approach. H. 60. Bessler.R. 35. K.L. 721-726. ST PR Monte Carlo simulation 14 steps ahead of an 18 equation quarterly model. Simple average composite has lowest RMSE with USDA outlook best single method and econometric worst. (Abstract in American Journal of Agricultural Economics.C. R OU Mainly about the process of determining and releasing USDA reports. M.G. corn and the soybean complex (beans.. 1979. 363-369. L.A.S. and S. 11. 200-206. and T. Alien I International Journal of Forecasting 10 (1994) 81p13S 571-589. and J. OU Need for improvements in speed and accuracy in outlook work.O. Newell. Nerlove. and spline function VP model slightly superior to Cooley-Prescott. The use of long-run price forecasts in farm planning. evaluation and dissemination.M. 1953. 530-535. W. 268 pp. Economic impact of public pest information: soybean insect forecasts in Illinois.D. DC). and M.M. 1978. W.R. 31 July-3 August 1983. 1986. 134-145. R. Alabama and South Carolina had smaller RMSE than USDA forecasts based on condition indexes made 1 or 2 months later. Nelson. New York. Unless the forecast leads to the same decision as information from scouting at least 90% of the time.R. J. Chapter 3 (pp. Warrington. MK For wheat. American Journal of Agricultural Economics. Analysis of Economic Time Series. price expectation as trend extrapolation. D. Yanagida. 117-125. 1985). Johnson. S. USDA livestock and poultry forecasts: process. Monthly steer and heifer supply. Forecast reliability of Yl% is worth 2 cents per acre and a perfect forecast is worth 66.. 9. 397-406. American Journal of Agricultural Economics. meal) results showed that market participants did await USDA announcements to make trading decisions. oil.L. . No single model consistently superior across commodities in either levels or turning point predictions. each from the same series. Naik. G. pp. Weimar. and R. IN. The case for and components of a probabilistic agricultural outlook program. A Synthesis (Academic Press. OLS reduced form estimation is more accurate within-sample. Relative forecasting performance of fixed and varying parameter demand models. CA) that are combined with forecasts from econometric model. S. 13 prices). 1984.T. Washington.. The Dynamics of Supply: Estimation of Farmers’ Response to Price (Johns Hopkins Press. PR Proposed system for forecasting pest damage compared with existing commercial scouting system. Neilson. 1989. Newell. A note on qualitative forecast evaluation. Applied Economics. Baltimore). 65 (1983): 1185. R SN Describes the partial adjustment schemes for price and quantity that Nerlove and others developed in the 1950s. Dixon. Zavaleta and M.F. R. G. Moffitt.. 66-86) is a history of dynamic supply response analysis in agriculture from 1917. Combining statistical techniques with economic theory for commodity forecasting. New York). American Journal of Agricultural Economics. 1986. 784-789.M.
ARIMA is always in dominant set and for risk neutral decision makers is always best. Texas A&M University.. C. Stengos.M. 9.B. assumptions and results of free market projections for the livestock and feed economy. An appraisal of composite forecasting methods. TS CO Revised version is Park and Tomek (1988). ST Quarterly recursive five region quadratic programming model of the pork sector. 577-589. second and third generation models (agricultural output not linked to macro variables. Tucker.. ST Combines annual. Western Journal of Agricultural Economics.D. R LA. Allen I I~terrzafio~laiJournal of Forecasting 10 (1994) 81. P. and W. A. risk neutral) used. Western Journal of Agricultural Economics. A. R PR Calculates and compares hedging strategies for egg producers using regression. 1974. Bootstrapping conftdcnce intervals: an application to forecasting the supply of pork. 233-242. T.5. LA Annual 15 equation submodel for 15 farm expense categories estimated by OLS or (where necessary) by maximum likelihood to correct for first order autocorrelation. L. Meilke and T. Hedging and income stability: concepts. North Central Journal of Agricultural Economics. United States. D. Seeley and C. 55. Tomek. Methods of forecasting production of fruit. Smith. University of Minnesota. Agricultural Economics Research 2-1. 1980-1990 and description of 1991 outlook. 1972. A. R ST TS CO PR Compares hog price forecasts from econometric (two equation). 1977.K.135 129 Oiiveira. J..45-5. 1984.. 57-66. 17 pp. 357-364. 1979.P. G. OU Recent history.W. D. and D. Owen. Actual and predicted 1969-1971 import quantities tabulated. Leuthold.. TX. J.M.O.. Agricultural Economics Research. All require an estimate of condition of the crop by crop reporters and use various correlations (or bivariate regressions).W. E-V. C. Generally.G. 11. American Journal of Agricultural Economics. Bessler. Penson. Tomek. 1982). Forecast distributions have positive skew.W. Pearson.G. ARIMA. linked and account for agricultural capital accumulation). Multivariate better.. R TS Demonstrates bootstrapping for both fixed and random explanatory variables at four forecast dates.W.D. 1992. 1991. 1961. Owen. soybeans and wheat. Methods.. Canadian Journai of Agricultural Economics. OU. 1990. Re-estimation of same specification after dropping last year of data allows limited ‘post-sample’ testing. North American-Japanese pork trade: an application of quadratic programming. and T. C. 86-92. American Journal of Agricultural Economics. Paulsen. 266273. 490-494. K.J. agriculture. 24.. W.B. Theil’s Vz (none over one) and relative turning point error. 1975. An appraisal of composite forecasting methods.E. R LA EV Discussion of multisector WdCrOeconomic model emphasizing agriculture foltowed by brief review of first. Hughes and R. Schlozhauer. 1988. Fabricated cut beef prices as leading indicators of fed cattle price. Park. A. with little difference between total hedging and partiai hedging based on the different forecasts. ST Annual simultaneous equation model. Houck.E. Park. expert and futures prices as forecasts 1-S months ahead. 25. not much worse than best single method. MacAulay. Outlook for the North American livestock sector. Canadian Journal of Agricultural Economics. Within-sample validation using mean absolute relative error (MAPEIlOO).. 10. Park.G. 61. Short-run forecasting models of beef prices. R TS CO Compares five VAR approaches using a four variable system. 1950. A modet of the world coffee economy: 1950-1968. Sporleder and D. implications and an example.A. 10-19. buyers or speculators in trading futures contracts. American Journal of Agricultural Economics. Prescott. The Food and Agricultural Policy Simulator. North Central Journal of Agricultural Economics. R LA Aggregate annual supply . SN TS CO Compares univariate and multivariate distributed lag models with random walk for 14 beef cuts. Forecast evaluation for mult~variate timeseries models: the U.S. R..A. Incorporation of general economic outcomes in economic projection models for agriculture. Excess capacity and adjustn~ent potential in U. Department of Agricultural and Applied Economics.. Kaldor. 6. composite and naive models when forecasts are used to guide produers. Journal of Farm Economics. J. Research X6-12. D. T. quarterly and monthly data in an 11 or 12 equation sectoral econometric model. 43. O’Connor and G. l-11. Garcia and R. 32 pp. D. DP-12. Hughes. Applied Economics. 4. Peck. Price.. European and World demand functions for coffee. College Station. and E. and W. Only part is described. R. 61-79. Cornell University Department of Agricultural Economics Paper A. 38. Pieri. D. J. the futures price was the more accurate forecast.S. R OU Compares standard ‘par method with two other methods.W. 410-419.W. Using a decision support framework to evaluate forecasts. Hedging reduced risk exposure. 23-43. 6. Quance.L.. Parikh. and L.P. 133-143. 1987. LA Crop yield trends (1940-1958). linked through identities. 1973. Penson. R. SSD. Price impacts of SRS production reports: corn. but conducts no tests. 1979. 1989. Restricted VAR (Schwartz) appears most accurate overall. 151-157. Risk efficiency criteria (FSD. An Overview of COMGEN: A Macroeconomic Model Emphasizing Agriculture Departrnent of Agricultural Economics Information Report DIR 84-l. and D. Western Journal of Agricultural Economics. Estimation of Farm Production Expenses USDA-ERS Technical Bulletin Number 1803. SN TS CO For monthly slaughter steer prices and soybean oil prices. Palmer. 1990.I. Park.F.G. 1977. compares six or five time series and econometric methods (including naive no change) with nine forms of composite (averages of twoforecast or three-forecast).J. 1986. TS CO ARIMA models of six cattle cash prices and cattle futures price compared with naive no change model. “The only thoroughly validated spatial equilibrium model found in the literature” (Thompson and Abbott.M. American Journal of Agricultural Economics. Jr. Tweeten. cattle market. Parikh. 57. Simple average composite best. Jr. 15. and J. Romain. published elasticities and stated assumptions used for 1959 projection of 1960-1963 annual values of 18 crop and livestock classes.
F. A. Aberdeen. 52.M. agricultural sector. Rausser. Johnson.E. Proceedings of the American Statistical 24. Roy. lY71.E. Roy. Methods. 469-478. Rpt 75. S. prices and uses of commodities.C. R. 5133516. 1975. in G. Prices account for predictable errors in farrowing intentions numbers.E. 127-130. UK). MP 240. B. Methods of forecasting hog production and marketing. Readings in Egg Pricing University of Missouri-Columbia. American Journal of Agricultural Economics. Rausscr (editor).G. Are farrowing intentions rational‘?. Short-term forecasts of U. J. and T. C.B. lY70. American Journal of Agricultural Economics. New York). (World demand prospects for grain in 1980 with emphasis on trade by the less developed countries Econ. 16YYl87.S.C. and the world grain model of Rojko et al. Agr. ARIMA and transfer function models with futures prices for soybeans. D. and R. R LA A nine equation model to link with Wharton EFA model (second generation). TS CO Compares two univariate ARIMA. and M. 54. 1-14.N. J.S. LA PG Mathematical programming model of 11 commodities and 27 regions. G. oil and meal. Runklc.R. issues: model construction strategy (depends on purpose). Do futures markets react efficiently to predictable error in government announcements?. in Gordon C. Cargill. 32 pp. 12. American Journal of Agricultural Economics.. The model is described in Konyar et al. Quiroga. R TS Results do not support existence of cycles. 28. A.J.. 43. Rojko. American Journal of Agricultural Economics. Agricultural Resources Model: Data Construction and Updating Procedures USDA-ERS Staff Report AGES 9304. Heady. Southern Journal of Agricultural Economics. pp. 3. 1977. Journal of Futures Markets. TS CO For both Bourke (1979) and Gellatly (lY79). No forecasting.J.C. 1980. Randall.J. The outlook for beef in the United Kingdom (School of Agriculture.. Moderately risk averse decision makers use outlook information and would be willing to pay for it.. Voss (editors)..A.E. Econ.1980. MK Not rational in the sense of being biased and inefficient. (1993). Econometric Models of Cash and Futures Prices of Shell Eggs USDA-ERS Technical Bulletin Number 1502. S. Allen I International Journal of Forecasting 10 (1994) SI-135 and demand equation first generation model. Moriak. a random walk would be the best ARIMA model. 289-313. Chapter 1.. Within-sample testing using forecasted exogenous variables if actual values would not be known at time of forecast. monthly fat cattle slaughterings. 1981. and C.F. Ireland.E. Compares within-sample prediction by reduced forms from 2SLS and 3SLS. The existence of broiler cycles: a spectral analysis. Henson.. POLYSIM: A national agricultural policy simulator. Schwartz. Futures market efficiency in the soybean complex. and R. Dept. 1976. Young. two transfer function type and econometric (two single equation) models. B. R OU PR Risk neutral decision maker has same storage and selling strategy when outlook information is available as when it is not. Runklc. M. 2255233. D.130 P. For. Rister. and P.S. policy impact analysis (little). aggregation (over commodities. 1961. Rausser. lY72. TS Revell.R.H.K. Journal of Farm Economics. Russell.K. 117-125. 1929. intention and pig crop survey values charted against time and judgmentally combined to give final outlook on hog slaughter. R ST EVA summary: history 1920. based on work by Takayama and Judge (Econometrica. McCormick. The U.S. Rayner. Roop. and commodity supply and demand elasticities to examine the impact of different policies. K. 1971. Developed at Oklahoma State University in 1970-1072. lYY2. 65. and M. 1983. 16.E. LA Recursive annual submodels for six commodity groups.R. 635-643. and W. An econometric analysis of the sorghum market. 1974. D.O. 32 (1064): 510-524).L. 14-21. Reveil. Konyar and I. Zeitner. Black. estimation methods. Rojko. Rausser. assumptions and results of the price and income projections of the United States Department of Agriculture. Ray. American Journal of Agricultural Economics. Review of Economics and Statistics. 1984. Southern Journal of Agricultural Economics. European Review of Agricultural Economics. G.J. 49. S. 175179. MK Live hog futures prices arc efficient with respect to farrowing announcements in USDA Hogs and Pigs report. Information.W. 151-158. 1993. Skees and T. 1974. 19X2. and A. 594-600. Agriculture (North-Holland. Econometric mod& for predicting weekly and quarterly egg prices.S. Also by MSE decomposition. Roy. Prediction of shell egg price: a short run model. Agricultural activity and tne general economy: some macroeconomic experiments. 28. December 1971). ST A five equation annual model. Modeling the World grain-oilseeds-livestock economy to assess world food prospects.K. and T. 109-121. hierarchical model structures and forecasting: a case study of dairy cows in England and Wales. Rogers and L. S. 80-98. Res. 73.S. New Directions in Econometric Modeling and Forecasting in U. ST A four equation quarterly model of the US egg sector. pp. Agr.. farrowing Association.. Elasticities and adjustment rates taken from various sources. G. MK CO Efficiency assessed by comparing post-sample forecasts of naive.W. 1976. Government farm programs and commodity interaction: a simulation analysis. 7. S. Review of Marketing and Agricultural Economics.K. Their sensitivity tested. ST Roy. Serv. Allen (editor). OU Hog-corn ratio. Evaluating the use of outlook information in grain sorghum storage decisions. beef and veal production and producer returns for fat cattle: an application of Box-Jenkins forecasting. Ray. Limited links among them.. space and time). 19Y1. Agricultural Economics Research. Projections to 1980 under three policy assumptions. 57. 5788590. Box-Jenkins forecasting models: com- ment. U. and E. Carter. ST Quarterly seven equation and monthly six equation models of the US egg sector. Agricultural Economics Research. R LA An accounting-type model that uses USDA-ERS 5-year projections of supplies. 9. in G. 34X-356. New conceptual developments and measurements for modeling the U. . OU EV Describes ‘projections not forecasts’. D.K.
American Journal of Agricultural Economics. Regression to predict monthly change in hog price (seasonally adjusted) as a function of detrended lagged prices of industrial stocks. ST Annual 15 equation pork sector submodel validated within-sample.E. Dean. 12. Outlook reports do provide new information. R TS CO Illustrates Box-Jenkins and exponential smoothing approaches using annual data. 88-102. J. Agricultural Economics Research. LA Gross farm product predicted from gnp and output per dollar of final demand for each of ten agricultural sectors. Schluter. 34. White. 1523-1536.E. Price and K. 26. A. Lists goodness of tit measures. The food and agricultural policy simulator: the poultry and eggsector submodel. Forecasting wheat yield: an application of parametric time series modeling. North Central Journal of Agricultural Economics.J. 27-44. 1987. Gadson. farm program allotments and diversions. Gadson. Eggert.E.M. 95 pp. G. Finds few significant abnormal returns in futures markets. Combining input-output and regression analysis in projection models: an application to agriculture. The food and agricultural policy simulator. Price and K. 247-254. A dynamic regression model of the U.E. Shafer. C. LA PG Annual linear program with flexibility constraints for cotton and 11 alternative crops in Fresno. but less information available for hogs than for cattle. An ARIMA analysis of tea prices. The food and agricultural policy simulator: the dairy sector submodel. 1989. 1989. Makes comparison of actual and projected acres for six crops using actual 1968 farm program provisions. 37-48.5. OU Developed scoring system to rate qualitative forecasts. J.G. Price and vafue effects of pecan crop forecast 1971-87. 30. G.P. 551563. 342-345. 1983b.N. and G.G.. American Economic Review 15 Number 3 Supplement Number 2.E. 97-103. 1949. AR(2) and ARIMA forecasts of monthly tea prices. Predicting shortrun aggregate adjustment to policy alternatives. L. 50. 1974.E. Journal of Agricultural Economics. 31.C. 1-15. Excludes livestock.T. Skinner. Price and K. corn and soybean acreages. 38. R SN The essay awarded the Babson prize by the AEA. American Journal of Agricultural Economics. 1991. hog market. Price and K. Rivista Internazionale di Scienze Economiche e Commerciali. T. Varoufakis. 1990. 52. Agricultural Economics Research. K. Journal of Farm Economics. Selzer. SN CO Single equation estimation comparing six different forms of expected price in acreage response functions. Western Journal of Agricultural Economics. Includes irrigation constraints. Canadian Journal of AgriculturaI Economics. and T. Sapsford. pp. CF. Gadson. Forecasting coffee prices: ARIMA vs. Schaller. Shapiro. Spreen.M. J. and R. Review of Marketing and Agricultural Economics. 1965.. 1990.E. 1982. Salathe. 23-34.. Compares naive. R LA Review of USDA-ERS modelling from 1970.M. corn (annual average) and hogs. 14.A. and F. TS Quadratic spline and ARIMA models applied to weekly data on sugar content of cane sugar. Flexibility constraints most critical determinants of solutions. 281-292. and D. Is verification possible? The evaluation of large econometric models. TS CO Derives ARIMA model from sectoral econometric model. Cotton and Wool Situation and Outlook Report USDA-ERS CWS-65. and L.H. Schmitz. Sapsford. harvested acres. l-22. Schatler. Predicts hog prices (equivalent to three steps ahead forecasts) both within and post-sample. D.H. Meyer.. R. Sharptes. 5. R LA EV One can corroborate but not verify theory. OU CO Pecan prices are more accurately explained by early season crop forecasts than by post-season final estimates. yield . Discussion of FAPSIM.. 1982. 21. labor constraints and capital constraints. Watts. Varoufakis. R. Alternative forms of price expectation in supply analysis for U. Monthly forecasts of hog prices 192551940 were 64% accurate and cattle prices were 62. 1973. J. L. econometric approaches. Allen I Intermtionul Journal of Forecasting 10 (i994) 81-135 131 Ryland.A. 35-1.. ratio of farm output to input and time trend. 1970. 34. 1399146. L.E. Forecasting crop quality. D.7% accurate. Salathe. SN TS CO Monthly econometric model with all lagged explanatory variables more accurate than seasonal ARIMA for 36 (one step ahead?) ex ante forecasts. 43. Southern Journal of Agricultural Economics. 55. H. Accuracy of livestock price forecasts at Kansas State College. and Y. 250~-258. Cotton production estimates: a historical review. 197.S. Blair and J. ST Annual 48 equation econometric model. J.M. R MK EV Reviews previous literature on impacts of outlook information on prices. according to the accuracy of within-sample predictions of five different specifications of average farm price.S.J.. LA Annual recursive programming of crop response in five regions of the US each divided into 8-12 subregions. J.W.293-304. Agricultural Economics Research. Cannot be compared with standard accuracy measures.. which improved the post-sample forecast. and W. 329334. CA. North Central Journal of Agricultural Economics. ST Annual 30 equation econometric model. 1982. 1925. Gadson. W. based on 1963 US I-O table. American Journal of Agricultural Economics. Predicting Regional Crop Production: An Application of Recursive Programrning USDA-ERS Technical Bulletin Number 1329. 1983a.B. Salathe. Agricultural Economics Research. L. Mintert.. The forecasting of the price of hogs. 37. ARIMA most accurate. a 360 equation model of nine crop and 12 livestock products.E. R TS Illustrates transfer function estimation of hog slaughter as function of the hog-corn price ratio. Schroeder..E. Shonkwiler. J.S.N. Prediction error regressed on ratio of gfp (gross farm product) and gnp implicit price deflators. 1968. and Y. Shideed. Sarle. Salathe. The response of the hog industry to a reduction in corn production: an application of the food and agricultural policy simulator. 95-105. Abnormal returns in livestock futures prices around USDA inventory report releases. OU Calculates average forecast (as percentage of final value) and 95% confidence interval for planted acres. I-14.
20.132 P. forecast RMSE is smaller in the bigger sample. F..P. 3SLS and LISE. Canada. Mayer.. Harrington and R. Swamy. Stillman. Sahi. R OU Compares average forecasting error 41 1981/1982-198911990 by month of US and foreign production and exports for wheat. Effect of public price forecasts on market price variation: a stochastic cobweb example. North Central Journal of Agricultural Economics.. Agricultural Outlook. Arnade. 7812. Use of forecasts in decisionmaking: the case of stocker cattle in Florida. pp. Hassan and H. 1981. 1971.P. 32-33. with some identities functioning as operating rules in forecasting/projecting. Smyth.T. 1985. ST Basic 42 equation annual model. 442-459. 145-150. Econometric model of the turkey industry in the United States. 31-47..G. and J.K. 1978. Allen I International Journal of Forecasting 10 (1994) 81-135 and production of upland and American Pima cotton for different months of the crop year (AugusttJuly) from 19651990. Subotnik. 2.H. Mueller. futures price and four composites. plus other accounting equations to calculate producer revenues. R. Tabulated forecast and final estimate values permit calculation of standard forecast accuracy statistics. 1927.. D. M. No comparisons with other forecast methods. Jr. but except for one step ahead. SN CO Table 1 contains 20 out-of-sample comparisons of root mean square errors from fixed coefficient and stochastic coefficient models. An overview of the USDA crop and livestock information system. SN Monte Carlo analysis of known three variable Box-Cox equation with different values of h parameter and sample sizes of 30 and 60. LeBlanc. Agricultural Economics Research. 1925. Conway and M.K. Agriculture (North-Holland. 22. D. Forecasting monthly slaughter cow prices with a subset autoregression model. 225-255.. R SN Uses correlation analysis on eight variables (including time and production) to forecast cotton price. Southern Journal of Agricultural Economics. 23 pp.S. Stonehouse. Southern Journal of Agricultural Economics. 167-176. A quarterly forecasting model of the U. 127-131. Chapter 8. hog and chicken model. 1982. A quarterly econometric model for corn: a simultaneous approach to cash and futures markets. 19.B.H. 55. W. Journal of Futures Markets. Smith.H. American Journal of Agricultural Economics. 1-8.. 1986. 41. 1990. 38. Smith. coarse grains and soybeans. and J. Blalock. R. A. Forecasting performance of models using the Box-Cox transformation. 3. part III: estimation.A.A. 3.B. 71. futures price is best.B. AO164. New York). and G. Canadian Journal of Agricultural Economics. 81-87. pp. 1983. Spriggs.A. September and October announcements appear to have a stronger impact than July and November. while this is not true with A less than zero.A. ZSLS. PR Variance shown to be theoretically always less with a forecast than without.R. Forecasting the volume and value of the cotton crop. Ottawa. J. ST A seven equation plus annual production equation model validated within-sample by static and dynamic simulation. For A greater than zero. 1. Spriggs. Sumner. OU Lists and discusses the USDA reports of various crop and livestock inventories and production in process that act as leading indicators of agricultural production and prices. in Agriculture Canada. B.. Are harvest forecasts news?. R ST A 29 equation (eight annual) cattle. 1979. totalling 72 expressions in all. 1984.V. Suds. Journal of the American Statistical Association. R SN TS CO PR Compares five methods (including naive no change) of forecasting second quarter feeder steer price as guide to producer of fourth quarter calves when considering whether or not to raise them to feeders. Journal of Agricultural Economics Research. B. 11. 1989. 4-5. but exponential smoothing most useful for decision making. Probable error (within-sample SO% confidence interval) was 2%. Spreen. T. A part of the model described by Westcott and Hull (1985). Spilka. Farm price is a function of futures price and quarterly stock carryover. TS CO Compares ARIMA. 83-88. 40 pp. ST A ten equation model validated by dynamic simulation both within and post-sample. Houck. Forecasting the acreage of cotton. Equal weights combinina most accurate for composite. all grades. J. R.A. Stillman..B. August. Soliman. T.. 47-60. egg sector USDA-ERS Technical Bulletin Number 1729. R. P.E.E. Agricultural Economics Research. 30. 77-110. 1978. Rausser (editor). coordinated by Z. pp. TS ARIMA model of monthly Florida cow prices.. Spreen. Smallwood. D. 1987. MK The relative change of corn and soybean futures prices on the day of a USDA announcement is significantly different from the change 5 days before and after.R. Vol.. Forecasts of Indiana monthly farm prices using univariate Box-Jenkins analysis and corn futures prices.P. McClave.S. R ST CO A five equation annual sector model estimated by OLS.R. J. D. R PR Shows that the standard errors of forecast from Goldberger method are about ten times larger than those from the approximate method of Teigen and Bell (1978). A quarterly model of the livestock industry USDA-ERS Technical Bulletin Number 1711. Single equation regression most accurate. stability testing and prediction. USDA-ERS. 16. 1989. American Journal of Agricultural Economics. Gajewski.. Simpson and J. in G.M. An econometric forecasting and policy analysis model of the Canadian dairy industry. A note on confidence intervals for corn price and utilization forecasts. 1973. 14-24.A. including three livestock price models . How accurate are USDA’s forecasts?.C. D. New Directions in Econometric Modeling and Forecasting in U. Huff. government support costs. publication no.P. No method emerged as clearly the most accurate forecaster. The stochastic coefficients approach to econometric modeling. SN First differences of acreage regressed on first differences of average spot prices for 5 separate months (known at time of forecast) and time trend using 1907-1921 data. Journal of the American Statistical Association. and C. and R. Commodity Forecasting Models for Canadian Agriculture. producer and consumer surpluses. 4420.
D. 345-387. Notes (p. But with parameter updating. 371) “Of the few modeling exercises that did list forecasting as an objective.137 pp. Results of a price forecasting competition: comment. No comparative tests against actual variability. Trapp.M. A quarterly econometric model of the Australian beef industry. 1974. 75 pp.P. Journal of Farm Economics. Journal of Farm Economics. U. 199-217. OU Compares accuracy 42 (using MAPE) and improvements (using Theil’s R statistic) of first and second revisions (in quarters two and three) to original estimates (in quarter one) for non-probability mail (rural carrier. price and exports and 12 months of forecasts. New developments in agricultural trade analysis and forecasting. Agricultural Economics Research. Confidence intervals for corn price and utilization forecasts: a reply. U. Congress. 50. 1500-1506.G. 13. 1982. Hallahan. corn and soybean futures had best price forecasting ability and potato futures best price stabilizing ability. H. 181-202. Department of Agriculture. R OU Describes mail surveys and probability sampling. 63. and T.S. Thompson.M. Tomek. Agricultural Situation. 1969. R. Stillman and Prentice (1987). OU Price change in spot corn and wheat market following monthly Crop pruductinn and quarterly Grain Stocks reports 1 day and 1 week after release of report. Abbott. X931.J. 34-35. Non-probability was more accurate.. Agricultural Sector Models: Description and Selected Policy Applications (lowa State University Press. R OU Source of the $125 million cost to fanners of overestimation of the US cotton crop. P. ST A ten equation model with beef supply.G.R. 1978b. and R. 1990. U. 1974. See comment by Spriggs (1978) and reply). 1179-1184.S. EV A critique of the methods of estimating and forecasting commodity prices. an econometric model that includes growth rate. TS Corrects Henriksson-Merton test confidence interval used in Dorfman and Mclntosh (1990). Which way profits after crop reports. The changing world for agricultural statistics. Agricultural Economics Research. T. Review of Agricultural Economics.. Crop Estimating and Reporting Service of the Depa~ment of Agriculture Report and Recommendations of a Special Subcommittee. 97-102. American Journal of Agricuiturai Economics.D. LA EV Compares equation estimates and descriptions of several large scale econometric mode&.” Thomson.. 3. Proposals for improvement of the agricultural outlook program of the United States.S. and W.R. Rausser (editor). 523-534. 48. Bell. Department of Agriculture. 40 states produced outlook reports. 45. Perhaps surprisingly. New York). 30. 457-465. U. 1213-1217. Temporal relationships among prices on commodity futures markets: their allocative and stabilizing roles.M. Referenced in Hoffman (1980).G. Journal of the Northeastern Agricultural Economics Council. 1966.N. almost none provide any forecasting performance measures out of the range of data used to estimate the model.. . Chapter 12.S. demand. is one of the two where the iixed coefficient model is better. in Gordon C. J. and P. C. Timm. Reports and meetings viewed by the Extension Service as ‘wedges’ (foot in the door) to get farmers to think about economic issues in making their plans. L.R. Forecasting the basis for corn in western New York. Referenced in Kunze (1990). Washington. Teigen. A. MK Compares futures price behavior for continuously storabte commodities (corn. The Story of Agricultural Estimates fvfisceifaneous Publication Number 1088. Myers.S. 1978a. 52. 1970. 2nd Sess. Suggests that economists’ preoccupation with highly detailed models to explain markets should be replaced by efforts to forecast and anaiyse policy using robust models. Taylor. Gray. American Journal of Agricultural Economics. Government Printing Office. IA). 1993. Taylor. pp. Throsby.. U. soybeans) where inventory hedging is important and discontinuously storable commodities (Maine potatoes) where forward pricing is important. Allen / in~ern~i~on~l Journal of Forecasting 10 (1994) N-135 I33 from Conway. R OU History of the development of crop and livestock data collection and analysis at the USDA. American Journal of Agricultural Economics. 82d Congress.D. 1981. as shown by a single within-sample prediction. Statistical Reporting Service.G. DC). 30. Bell.W. Trelogan.. Committee Printing (U. Toliey. R OU SN CO USDA intentions survey more accurate than three econometric models.C. starting and slaughter weight indexes is best. SN Basis (futures price minus cash price) is predicted from an annual single equation.S. C. Tomek. OU In 1929. Teigen.D. These lead to large forecast errors in basis. 1963. 1991. W. Confidence intervals for corn price and utilization forecasts. R PR Forecast variance is approximated by standard error of estimate squared (often obtained judgmental~y) times a correction factor (n + k)in.C. Auxiliary equations needed to predict values of explanatory variables show large forecast errors. New Directions in Econometric Modeling and Forecasting in U.S. 56. The history and objectives of outlook work. 1274-1276. Agriculture (North-Hoar land. Ames. House Committee on Agriculture. Journal of Farm Economics. 15. and T. 372-380. The pork model of Conway et al. 1952. L. 23-29.. Tegene. though perhaps it refers to a period of limited variability.. Analysis of the accuracy of USDA hog farrowings statistics. Their approximation is closer to USDA values than Goldberger is (and smaller). R OU EV Makes nine proposals including making probabilistic outlook statements and projecting probable actions of government programs. 1976. 73. where k is the average number of parameters per equation and n is the average sample size of each estimated equation. H. R PR Compares USDA forecast RMSE with standard error from Goldberger method and authors’ approximation. 13. Forecasting short-run fed beef supplies with estimated data. Tomek. probability mail (March 1963-March 1970) and multiple frame (list frame and probability area frame March 1~0-March 19’73) surveys. W. and R. American Journal of Agricultural Economics. Empirical analysis of agricultural commodity prices: a viewpoint. Statistical Reporting Service. September 1959-March 1963). Economic Record. R LA Review with over 100 references. 1984.L.
Short-Term Forecasting: Accuracy of USDA’s Meat Forecasts and Estimates GAOIPEMD-91-16.C. ARIMA.S. 1990. combining. 1983. H. P. P. 1965.C. 103-117.G. R SN TS CO Compares single equation. U. 19Ylb. July 1975. Westcott.. Estimating the United States cotton crop. 1981-lY86. G. General Accounting Office.. R OU Summary of the survey and analysis methods described in USDA. 12. with practically no difference among models.. SN Shows use of annual single equation models. 109-112. U.C. 1992. Griffith.. hogs and broilers. Regression equations for retail prices of 14 fertilizer products based on retail prices of the three main constituents. 307-327.S. A quarterly model of the U. 5.134 P.S.. system. forecasts.R. Vroomen. 933107. Von Massow. J. 1988. Agricultural Economics Research. Major impact from combination of technological advance and improvements in human capital. naive. Statistical Reporting Service. SN For CPI food at home/away from home indexes. Walker. revised September 19X3. Detailed tabulation of timing and nature (estimates. Scope and Methods of the Statistical Reporting Service Miscellaneous Publication Number 1308. Wells.. Turvey. A. TS Model that forecasts both cash and futures prices (of daily soybean meal prices) slightly more accurate than model of futures price alone or cash price plus basis. IS6 pp. Department of Agriculture. X8 pp. Walters. Compares USDA. General Accounting Office. 1985.. Agricultural Economics Research. T.5 U. SN Monthly wholesale price and transportation cost used as explanatory variables in regression equation for retail price of each fertilizer constituent (anhydrous ammonia. Westcott. 300-304. Forecasting retail fertilizer prices: a combined time series regression analysis approach USDAERS Technical Bulletin Number 1789. Using any model as signal is more profitable than either routine hedge or no hedge over a 3 month period. 32 pp. SN Annual single equation. Economic Research Service/ National Agncultural Statistics Service. D. 128 pp. exports and stocks of four major crops and dairy products. 70-76..R. 1986. Wallace. Errors were similar to those made by private analysts. Same nethods used to relate yield of same ears of corn to the characteristics. Forecasting South Carolina tomato prices prior to planting. 1923. J.S. R OU Detailed description of current survey and analysis methods by crop and livestock commodity.J. Department of Agriculture. OU Over a 4 year period. General Accounting Office. R TS Annual and monthly.S. During 10X1-1988. An early (probably the first) effort to construct an expert system. 122-126. OU Correlation analysis used to measure weights assigned to six observable characteristics of ears of corn by experienced judges. Hedging with forecasting: a state-space approach to modeling vector-valued time series. What is in the corn judge’s mind? Agronomy Journal. A spectral analysis of world cocoa Weiss. Westcott. R OU CO Reports error of USDA forecasts of aggregate farm program budget requirements. intentions) of the various reports. 27-30. 76 pp. Weersink and C. OU EV Clients need not served by a policy that separates market news from market outlook. R OU CO Calculates bias error and total error (MAPE) of annual USDA forecasts of production and price of beef.S. U. price. USDA Commodity Forecasts: Inaccuracies May Lead to Underestimates of Budget Outlays GAOIPEMD-91-24. U. 1977. 1986. Comparative forecast accuracy in the New South Wales prime lamb market. 15. 10-18..S. 33-3. 1970. SN CPIs for 21 food groups estimated by OLS and 3SLS. but little related to yield. Aggregate naive error always less over the 5 year period. expert. USDA/private are similar and usually better than naive. Agricultural Economics Research. 28-33. Southern Journal of Agricultural Economics. Crop and Livestock Estimates Agriculture Handbook Number 671. F. Agricultural Situation... PR Both stationary and non-stationary transition probability models predict about 60%~ less hog producers in Ontario by the year 2000. Statistical Reporting Service. 34. Critique of current outlook practices. weekly average price after Hog and Pigs released was about equally higher and lower than weekly average price in week when report was released. prices. Referenced in Hoffman (1980). U. Department of Agriculture. 12. Department of Agriculture. Allen I International Journal of Forecasting IO (1994) X1-13. based on data available in February. Monthly food price forecasts. 1991a. 1983. Australian Jourral of Agricultural Economics. all prices and most other variables were overestimated (negative bias). 14 pp. Hog reports and market prices. 1992. American Journal of Agricultural Economics. Suggests improvements to USDA’s forecasting process. Journal of Futures Markets. Volume 7. Length of ear most important to the judges. 52. private analysts and naive forecast errors for various marketing variables for corn and wheat. Vukina. Wallace. Six steps ahead forecasts of wholesale prices and transport costs from ARIMA models and autoregression respectively. 1953.J. Aggregate indicators in the quarterly agricultural forecasting model: retail food prices USDA-ERS Staff Report AGES 860916. OU Examines USDA forecasts 1-5 years ahead for production.. USDA’s Commodity Program: The Accuracy of Budget Forecasts GAO/ PEMD-88-8. Vere. and G. R OU EV Comments on the problems with crop forecasting raised by the Congressional inquiry. Predicting the beef cattle inventory. D. Dynamics of structural change in the Ontario hog industry. 17.G.A. M. Canadian Journal of Agricultural Economics. 32.S. No forecast. Major Statistical Series of the U. 197221986. 40. H. Total errors averaged less than 6% with small underestimates of production and broiler price and small overestimates for beef and hog prices in the 1983-1989 period. 1989. P. Timing of the various kinds of reports also detailed. dairy sector and some of its policy implications USDA-ERS . 1980. wheat and dairy programs. 1981.S. 1991. phosphoric acid and potassium chloride). Compares USDA and naive forecast errors for budget requirements for corn. Canadian Journal of Agricultural Economics. 28 pp.
P. Qrazem. 1-7. Garcia. hog and poultry sectors (the livestock sectors based on Stillman (1985)). TS CO Compares ARIMA..D.C. 38. White. ST A nine equation model tested by dynamic simulation (using actual exogenous variables) one to four steps ahead both within-sample and post-sample. unpublished Ph. Hull..C. production P. error correction (~ointegrati~~n) model and asymmetric Bayesian VAR (raw and differenced data) for forecasting monthly slaughter steer prices. Vancouver. 123-132.F. Zepp. RP and regression. Wcstcott. ST Annual forecasts based on assumed inventory changes. Geoffrey ALLEN is a Professor of Resource at the University of Massachusetts. soybean complex. dissertation. H. Within-sample and post-sample dynamic simulations one to four steps ahead using actual values of exogenous variables. B. Differenced VAR was most accurate for one to six steps ahead forecasts. iie in agricultural economics from the University of Davis.O. 15. 642-649. R LA Aggregate annual three equation (plus accounting equations) first generation model.H. Supply projections for the Australian beef industry. Agricultural Economics Research 37-1. 58. American Journal of Agricultural Economics. 1990. A Quarterly Forecasting Model for U. Biography: Economics has a Ph. Hull and R. 1987.B. Most base their expected price on recent past actual price. F. Projections to 1985 of 12 scenarios of demand.C. Milk stocks. The rationality and value of USDA crop forecasts. 1985. D. Ear& forecasts more valuable than later for market supply information. H. and income projections under alternative assumed demand and supply conditions. and P. VAR (raw and differenced data). B. corn.. 703-711.S. Journal of Agricultural Economics. Yu. Price forecasting with time-series methods and nonstationary data: an application to monthly U. G. cattle prices. California.O. TS Zapata. Bayesian and Nonbayesian Techniques for Forecasting Monthly Cattle Prices. 1976. beef. 1969. X5-370. Yeh. OU Survey of 80 milk producers in a small area of Illinois. 34 pp.-C. SN CO PC Compares LP. Williams. Predicting aggregate milk production: an empirical study. 40. TS Monthly ARIMA model fitted and used with discrete adjustment (filter) to forecast post-quota milk production. supply and inflation rates. An empirical study of price expectations and production plans.J. and D.B. 3-14. Agriculture USDA-ERS Technical Bulletin Number 1700.. 3953. They expect prices of all outputs and inputs to move together. wheat. Review of Marketing and Agricultural Economics. soybeans and spring wheat are found to be inefficient and/or biased. 72 (December 1990): 1353. 223-234. Canada. 1990. Zapata. Green. Prices. 51. Forecasting milk output in England and Wales. University of Illinois. McAlexander. presented at the American Agricultural Economics Association annual meetings. 1985. 35.. R LA A 160 equation model of the corn. P. Amherst. and P. Regression best. Westcott. White. American Journal of Agricultural Economics.D.. Parameters (elasticities) from other studies. 1972. and R. W.) MK Several forecasts of planted acreage and harvest size of barley. SN CO Compares forecast performance of two models of corn price. farm outputs. Relationships between quarterly corn prices and stocks. oats. (Abstract in: American Journal of Agricultural Economics.J. imports. Western Journal of Agricultural Economics.S.Technical Bulletin Number 1717. . C. support price and price deductions are all contemporaneous exogenous variables. Journal of Farm Economics. 1987. His research interests are in analysing decisions under uncertainty.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.