Scribd Logo
Upload

Welcome to Scribd!

  • Some Monetary Facts

    Uploaded by

    vsamoa
    5 views16 pages
    macro

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    macro

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    5 views16 pages

    Some Monetary Facts

    Uploaded by

    vsamoa
    macro

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 16
    Fel seve Bink of Monegos Carey Review ‘1,0 8, Samer 185, gp 2-11, Some Monetary Facts George T. McCandless Jr Warren €. Weber ‘Associate Professor of Economics Senior Research Oioer Universidad de Andes Research Department Buenos Aires, Argentina Federal Reserve Bark of Minneapolis Abstract ‘This article describes three long-tun monetary facts derived by examining data for 110 counties over a 30-year period, using thee definitions of a country’s money supply and two subsamples of countries: (1) Growth rates of the money supply and the general price level are highly coreated for all three money definitions, For the full sample of countries, and for both subsamples. (2) The growth rates of, ‘money and feal curpur are not correlated, except for a subsample of coumiries in the Organisation for Economic Co-operation and Development, where these growth rates are positively correlated. (3) The rate of inflaton and the growth rate of real output are essentially uncorrelated. ‘The views expres herein ae those of the euthors ond not necessarily those of the Federal Reserve Bank of Minneapolis or the Feral Reserve Syste ‘The Federal Reserve System was established in 1913 t0 provide an elastic curency, discount commercial credit, and supervise the banking system in the United States. ‘Congress changed those purposes somewhat with the Em ployment Act of 1946 and the Full Employment and Bal- anced Growth Act of 1978. In these acts, Congress in- structed the Federal Reserve to “maintain long run growth cof the monetary and credit aggregates commensurate With the economy’ long nin potential to increase production, 50 as fo promote effectively the goals of maximum employ- ‘ment, stable prices, and moderate long-term intrest rates” (FR Board 1990, p. 6). Implicit in these instructions from Congress is the assumption thatthe Fed has dhe ability, through its monetary policy, to control these economic variables. Does it? Clearly it does have a measure of con- trol over some definitions of money. But the links between ‘money and the other economic variables have yet 10 be conclusively established. The facts about those finks can help determine how well the Fed can do its job. The pur- pose ofthis study isto improve upon past attempts to de- termine what the facts are. ‘A central bank’s major instrument of monetary policy is the growth rate of the money supply, targeted either: rectly or indirectly through some nominal target like an in- terest rate or the exchange rate for the country's curency. Different central banks choose to adjust different defni- tions of money, whichever they deem appropriate for their direct instrument, The target for price stability is typically some measure of the country’s inflation rate, andthe tar- get for real economic activity, or production, is typically the growth rate of national output ‘A natural way to start to analyze the ability of changes in money growih to affect the rate of inflation or output ‘growth is to examine the statistical comelations between ‘these Variables. Todo that, we need to make some choices. Do we [ook for corelations in data over the short rum— during a quarter or a year, for example—or do we con- centrate on much longer time periods? Do we look for cor- relations within or across countries? For reasons explained below, we here examine long-run correlations over a large cross section of countries, although we do check the 10- business of our results by determining how sensitive they are tothe choice of countries included inthe eross section. Our findings about these correlations indicate that over the long nun the Fed has more ability to follow Congress’ ‘mandate about inflation than its mandate about prociuction, Specifically the comelations that we compat reveal these long-run monetary facts: ‘¢ There isa high (almost unity) correlation between the rate of growth of the money supply and the rate of in- flation. This holds across three definitions of money and across the full sample of couniries and two sub- samples, ‘¢ There is no conation between the growth rates of ‘money and real output. This holds across all defini- tions of money, but not fora subsample of countries in the Organisation for Economic Co-operation ancl Development (OECD), where the correlation seems to be positive. ‘© Theres no conelation between inflation and real out- pat growth. This finding holds across the full sample and both subsamples Studying long-tun, cross-country correlations lke those wwe consider is, of course, not new, What is new here is threefold: we consider a larger number of countries than have been used before, we consider more definitions of ‘money, and we consider how sensitive the results are 10 the choice of subsamples of countries. Methodology In this article, we examine fong-run correlations between ‘money growth ank other variables because many econo- mists and policymakers have strong reservations about the ability of monetary policy to ht short-run targets for ether inflaton or output. Milton Friediman is perhaps the best- {known exponent ofthis view. He has said, “T don’t try to forecast short-term changes inthe economy. The record of ceonomists in doing that justifies only humility” (quoted in Bennett 1995). We study across section of countries rather than just a single couniry because we want our results to be indepen- dent of policy rules. If we were to study a single country, the corelations we obtained could be an artifact ofthe par- ticular policy rule or rules being followed there. For ex- ample, suppose a central bank were to follow a constant growth rtc nile forthe money supply. If we examined the time series for the growth rate of money and the inflation rate for that economy, we would observe no correlation be- tween these two variables. If, instead, the central bank chose t follow a feedback rule, where the growth rate of _money was determined by the inflation rate, then we would observe a perfect corclation between money growth and inflation ‘We hope thatthe range of policy rules in our cross sec- tion oF countries is so broad that the correlations we ob- serve are independent of the policy rules. Even if all cen- ual banks were following a constant money growth rule, ‘ve doubt that they’d all be following the sare one, That's ttue for feedback rules too. So, by using a large cross sec- tion of countries, we hope our correlations willbe free of policy rule influences. Independence of correlations from policy rules is im- portant because we want the corelations we find to be useful for determining whether causal relationships exist ‘While correlations are not direct evidence of causality, they do lend support to causal hypotheses that yield predictions consistent with the comelations. Consider, forexample, the hypothesis that a monetary policy with'a higher growth rate of money will result in a higher inflaton rate than a policy with a lower rate of growth in an otherwise ident- cal economy. That hypothesis would be supported (ough by no means conclusively) by observed pexitive comrela- tions between money growth and inflation, This study is based on time series data for 110 coun- tries. For each country, we calculate the long-run (up to 30-year) geometric average rate of growth forte standard ‘measure of production, gross domestic product adjusted for inflation (real GDP); a standard measure ofthe general price level, consumer prices; and three commonly used definitions of money (MO, MI, and M2), We also look for correlations over two specific subsamples of counties, ‘One of the subsamples consists of 21 OECD countries: the other consists of 14 Latin American countries.’ The countries within each of the two subsamples are more ho- ‘mogeneous than those in the full sample in terms of avail- able technology, education, and level of development of f- nancial (and other) institutions. We consider the findings from these subsamples as a crude test of robustness of our full sample facts. ‘The data we use come from the CD-ROM version of the Intemational Monetary Fund's International Financial Statistics (HFS). The time period we consider is from 1960 to 1990, For each country with 10 or more years of data (110 counties), we calculate the geometric rate of growth for consumer prices (line 64 ofthe JFS tables); three defi- nitions of money—MO, cumency plus bank reserves (line 14); M1, money easily used in transactions (line 34); and M2, money easily used in or converted into use for trans- actions (the sum of lines 34 and 35}—and real GDP. The growth rite of real GDP is calculated by subiracting the growth rate of consumer prices from that of nominal GDP (ine 99%). ‘The Facts [Now we will restate each of the three results of our study, describe each in detail, and discuss how our results com- pare with those of previous studies. Money Growth and Inflation In the long run, there is a high (almost unity) carelation between the rate of growth of the money supply and the rate oF inflation. This holds across three defintions ‘of meney and across the full sample of counties and two subsamples. ‘The evidence on the long-run relationship between the rate ‘of money growth and the rate of inflation From ou sam- ple of 110 countries is presented in Table 1. It shows a high correlation between money growth and inflation for ‘both narrow and broad definitions of money. For the fall ‘sample, for each of the three definitions of money we con- sider, the correlation coefficient between the rate of change of the money definition and the rate of change of consum- et prices is 0.925 or higher. ‘The evidence from our sample also suggests tht the ‘growth rates of MI and M2, which are broader definitions ‘of money, are slightly more highly correlated with inflation rates than isthe grovith rate of MO, which isa narrow defi- nition, The correlation coefficients for the broader defni- tions are both approximately 0.95, whereas that fr the nar- rower definition is slighly lower, 0.925. ‘The evidence from the subsamples of OECD and Latin ‘American countries, also shown in Table 1, confirms the robustness of the high conelation between money growth and inflation, For these subsamples, the correlation coet= ficients between money growth and inflation are always higher than 0.89, and the relation is always weaker for MO than for the broader monctary aggregates, ‘The high correlation between money growth and infla- tion suggests thatthe relationship between these two vari- ables is very close to linear. The natural question is, What {is the slope of the relationship? Here the slope is very close to unity, ass illustrated in Chart 1, where we plot average rates of change of the M2 definition of the money supply and average rates of change of consumer prices forthe full 110-country sample. Each point in the chart represents the ‘observations on money growth and inflation for a particu- Jar country. In the chart we have also drawn a 45-degree line through the grand means of the observations. Inspec- tion shows that the individual country observations lie on ‘or very close t0 such a ine ‘The finding that money growth and inflation have a line «ear relationship with a slope very close to unity brings to mind the quantity equation. The quantity equation is Q) MxVaPxy where Mis the money supply, Vis the velocity of money (roughly, how many times each dollar in the money supply is spent each year), P is the price level, and Y is real out- pat. Written in terms of growth rates, the quantity equation becomes Q) mtvepty ‘where the lowercase letters in (2) refer to the growth rates of the variables represented by the uppercase letters in (1). ‘The growth rate version of the quantity equation implies that there should be a linear relationship between money ‘growth and inflation with a slope coefficient of unity when ¥ and y are treated like constants ‘The evidence in Chart 1 seems to support the quantity equation, atleast as long-run constraint on the effects of ‘monetary policy. That the 45-degree line through the grand means does not go through the origin of the graph sug- gests that central bank cannot generate a particular long- run rate of inflation by choosing an equal long-run growth rate for the money supply. The long-run inflation rate is influenced by the growth rates of ral output and velocity as well as by the growth rate of money. However, a cen- tral bank can be confident that over the long run a higher growth rate of the money supply will result in @ propor- tionally higher inflation rate (Our finding is consistent with what other studies have found. A sampling of them we summarize in Table 2. This tuble shows that the existence of a hich comelation be- {tween money growth and iniation has been found in many studies, but these studies have focused almost exclusively ‘on broad definitions of money. Lucas (1980), for example, applies filters that progressively emphasize the long-run relationship in U.S, data between MI and the consumer price index, He finds thatthe relationship becomes more regular, with a coefficient closer to one, the more the filter stresses the low frequencies (the long-run relationships). ‘Lucas (1980, p. 1005) claims thatthe low-frequency rela- tionship he finds represents “one way in which the quanti- ty-theoretic relationships can be uncovered via atheoretical rmetivods from time-series which are subject to a variety of ‘other forces.” Other evidence for the long-run relationship between money growth an inflation has come from studies using

    You might also like