STRATHCLYDE

DISCUSSION PAPERS IN ECONOMICS

AN INTERREGIONAL INPUT-OUTPUT ANALYSIS OF THE POLLUTION CONTENT OF TRADE FLOWS AND ENVIRONMENTAL TRADE BALANCES BETWEEN FIVE STATES IN THE US MID-WEST BY SOO JUNG HA, GEOFFREY HEWINGS AND KAREN TURNER

NO. 09-20

DEPARTMENT OF ECONOMICS UNIVERSITY OF STRATHCLYDE GLASGOW

1

An interregional input-output analysis of the pollution content of trade flows and environmental trade balances between five states in the US Mid-West
by

Ha, Soo Junga*, Hewings, Geoffreyb, and Turner, Karena

a

Fraser of Allander Institute, Department of Economics, University of Strathclyde, Scotland
b

Regional Economics Applications Laboratory, University of Illinois, US

Abstract In this paper we attempt an empirical application of the multi-region input-output (MRIO) method in order to enumerate the pollution content of interregional trade flows between five Mid-West regions/states in the US –Illinois, Indiana, Iowa, Michigan and Wisconsin – and the rest of the US. This allows us to analyse some very important issues in terms of the nature and significance of interregional environmental spillovers within the US Mid-West and the existence of pollution ‘trade balances’ between states. Our results raise questions in terms of the extent to which authorities at State level can control local emissions where they are limited in the way some emissions can be controlled, particularly with respect to changes in demand elsewhere in the Mid-West and US. This implies a need for policy co-ordination between national and state level authorities in the US to meet emissions reductions targets. The existence of an environmental trade balances between states also raises issues in terms of net losses/gains in terms of pollutants as a result of interregional trade within the US and whether, if certain activities can be carried out using less polluting technology in one region relative to others, it is better for the US as a whole if this type of relationship exists.

2

environmental trade balance. air pollution JEL Codes: C67. on which Soo Jung Ha is the Postdoctoral Research Fellow and Geoffrey Hewings is one of the main collaborators. Q53 and Q56 3 . Keywords: Multi-region input-output models.Acknowledgements This paper has been developed as part of Soo Jung Ha’s and Karen Turner’s research program under the latter’s ESRC Climate Change Leadership Fellowship (Grant reference RES-066-270029).

We follow Turner et al (2007). as the first step of attempting to develop a general equilibrium analytical framework for US Midwest framework. (2008) in two respects. and Wiedmann. carbon monoxide (CO).1. Second. In this paper. multiregion context. Our focus in the current ESRC Climate Change Leadership Fellowship Project is the application of such techniques at a sub-national regional level and secondly. increasingly (see Wiedmann et al. 2009. identifying 5 Midwest regions and the rest of the US (RUS). Introduction Input-output techniques are frequently used to account for emissions related to sectoral economic activity an. we introduce a greater level of sectoral disaggregation. We extend the analyses of McGregor et al. focussing on the two region case of Scotland and the rest of the UK (RUK). The process of working on developing interregional models for the US Midwest for this purpose is likely to be beneficial in terms of considering interregional interaction and providing opportunities for comparative analysis with the UK cases. for reviews) to and estimate the pollution content of trade flows and emissions ‘trade balances’ between regions/countries in a multi-sector. However. Therefore. We focus on the emission pollutant. the research constructs the structure of a Midwest Input-output Model using the 1992 dataset from MWREIM (Midwest Regional Econometric Input-output Model) for the Midwest economy (developed by Regional Economics Application Laboratory) and analyzes the interaction between the economy and CO emission inventory and illustrate CO emission attribution analysis to examine the interdependence between regions of the Midwest in terms of environmental spillover effects. to develop appropriate modelling frameworks to analyse the impacts of changes in policy and other disturbances on pollution trade balances. we introduce a greater degree of spatial disaggregation. First. 2007. who provide an empirical application for the UK. who propose the empirical application of the multi-region inputoutput (MRIO) method of accounting for pollution trade balances and McGregor et al (2008). we are currently in the process of updating this dataset to 2007 in 4 . analyzing the CO2 trade balance between these regions. we apply Turner et al’s method (2007) to the case of the Midwest states within the US. identifying 13 production sectors.

…. By multiplying emission intensity coefficient generated per unit of output X.…. The traditional input-output approach can be written in matrix form: (1) X = AX + Y (2) X = (I − A) Y −1 where A is the input-output matrix. A Midwest environmental input-output 2. Their exposition is given in terms of the 2-region case and applied it as the empirical example of Scotland and the rest of UK.. Here we extend the 2-region framework to 6-region case for Midwest in USA. N=13 commodities according to MWREIM structure. The accounting framework We apply the multi-region environmental input-output framework from Turner et al. and Wisconsin) and the rest of US (RUS). 5 . At this stage our main focus is explaining the potential information and analytical content of the environmental interregional input-output framework. Ohio. Michigan.1. (2007) to the case of the US Midwest. (3) P = E (I − A) Y = EX −1 As a extended form. 6 regions are composed of 5 MW states (Illinois. 2. Each region has i=1. Indiana. the amount of emission pollution can be written. and X is the output. Y is a vector of aggregated final demand.order to make the framework more empirically useful. N=13 production sectors producing j=1.

(I-A)-1 is the interregional Leontief rs rs inverse matrix. 2001). divided by total output in consuming sector j in region s. Due to a lack of appropriate data. CO.. In this study. i. the total emissions generated in region 1(Illinois). showing the physical amount of CO directly generated per unit of output. erx is a 1 × N vector of emission intensity coefficient for a single pollutant. xij . p1 are given by summing along the first row of each P matrix so that y y y y y y (5) p1 = p11 + p12 + p13 + p14 + p15 + p16 while the total emission in all regions of MW and RUS that are supported by region 1(Illinois) final consumption demand are given by summing down the first column of each P matrix so that y y y y y y (6) p1y = p11 + p 21 + p31 + p 41 + p51 + p 61 6 . (2008). produced by sector i in region r. the interregional Leontief inverse matrix is 78 × 78 (13 production sectors × 6 regions) j representing a conventional ‘Type І’ analysis.e. The A matrix is defined as aij = xij / X s . the amount of output produced by each sector j rs i in region r and used as input by sector j in region s. Our treatment of trade between MW and RUS follows the ‘consumption accounting principle’ (Munksgaard and Pederson. Hence. xi .(4) y ⎛ p11 ⎜ py ⎜ 21 y ⎜ p31 y p ⎜ 41 y ⎜ p51 ⎜ py ⎝ 61 y p 12 y p 22 y p 32 y p 42 y p 52 y p 62 y p 13 y p 23 y p 33 y p 43 y p 53 y p 63 y p 14 y p 24 y p 34 y p 44 y p 45 y p 46 y p 15 y p 25 y p 35 y p 45 y p 55 y p 56 y p 16 y p 26 y p 36 y p 46 y p 56 y p 66 ⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠ = ⎛ e1x ⎜0 ⎜0 ⎜0 ⎜ ⎜0 ⎜0 ⎝ 0 x e2 0 0 0 0 0 0 x e3 0 0 0 0 0 0 x e4 0 0 0 0 0 0 x e5 0 ⎞ ⎟ ⎟ 0 ⎟ 0 ⎟ 0 ⎟ x⎟ e6 ⎠ 0 0 ⎛ I−−AA11 ⎜ 21 ⎜ − A31 ⎜ − A41 −A ⎜ − A51 ⎝ 61 − A12 I − A22 − A32 − A42 − A52 − A13 − A23 I − A33 − A43 − A63 − A53 − A62 I − A44 − A54 − A64 − A14 − A24 − A34 − A45 I − A55 − A65 − A15 − A25 − A35 ⎞ ⎟ ⎟ − A46 ⎟ − A56 ⎟ I − A66 ⎠ − A16 − A26 − A36 −1 ⎛ y11 ⎜ ⎜ y 21 ⎜y ⎜ 31 ⎜y ⎜ 41 ⎜ y 51 ⎜y ⎝ 61 y12 y 22 y 32 y 42 y 52 y 62 y13 y 23 y 33 y 43 y 53 y 63 y14 y 24 y 34 y 44 y 54 y 64 y15 y 25 y 35 y 45 y 55 y 65 y16 ⎞ ⎟ y 26 ⎟ y 36 ⎟ ⎟ y 46 ⎟ ⎟ y 56 ⎟ y 66 ⎟ ⎠ y prs is a scalar of the amount of CO generated in production activities in region r to support regions s. we only carry out conventional ‘Type I’ 6 region input-output attribution analyses (McGregor et al 2008). X s . we do not extend our framework for direct emissions generation by household as final consumers or endogenise trade to close the system at the national level by following McGregor et al (2008) in their analysis using what they refer to as a trade endogenised linear attribution system (TELAS) in the analyses of McGregor et al. Therefore.

MWREIM). there are 456 endogenous variables and 192 exogenous variables from 1969 to 2020 based on 1992 Input-Output table. 2. This allows us to determine the interregional intermediate demands matrix (78x78) by 6 regions and 13 production sectors. This model. 1997) integrates econometric and input-output components. based on the initial formulation of Conway (1990. investment and government) for 1992. Figure 1 shows the provisional structure of an interregional MW IO table with 6 regions. employment related variables for 5 States and Rest of US. (1996. Wisconsin and Rest of US (The Midwest Regional Econometric Input-output Model. an interregional A coefficient (78x78) and total final demand (78x3) for 5 MW states and RUS are extracted from MWREIM by the extraction method (Israilevich et al. wage rates. (1997) for more details). There are also Data Resources Inc (DRI) and WEFA (Wharton Economic Forecasting Associates) variables and MW variables as exogenous variables. 1991). output variables by 13 industry sectors and final demand. Endogenous variables are composed of employment. unemployment). first of all. 13 sectors and final demand (consumption. Ohio. In order to construct an interregional MW IO table. income. Michigan. employment and income) and several major economic aggregates (such as gross regional product. Construction of MW interregional IO table The Regional Economics Applications Laboratory (REAL) at the University of Illinois has constructed a number of impact and forecasting models for Midwest area that cover Illinois. Indiana. 1997). This model allows for the extraction and forecasting of input-output tables on an annual basis (See Israilevich et al. In MWREIM.2. enabling impact analysis to be conducted as well as annual forecasts for a 30-year horizon for 13 different SIC (Standard Industrial Classification) based industrial sectors (production. region 1 (Illinois)’s CO trade balance with other US regions would be calculated by the difference between eq (5) and eq (6).. 7 . income. and further developed by Israilevich et al.According to Munksgaard and Pederson (2001) method.

The first step is initially to estimate final demands using REIM for 5 states (Illinois. Figure 1. 8 . our assumption is that the proportion of each final demand of imported goods and services from other states has the same sectoral structure as their own final demand of local goods and services. and government final demand from other state’s industry sectors. And we calculate the fixed ratio of consumption. respectively. total final demand is necessary to split into the sub-matrices to identify final consumption of local state and imported goods and services from other states. firstly we need to get the rest of final demand consumption by total consumption minus IL own consumption and then apply each state r’s consumption ratio excluding IL consumption by industry i. For example. Indiana.However. Structure for actual MW region IO table IL Intermediate Final Demand sector i=1-13 IL i=1-13 IN i=1-13 MI i=1-13 OH i=1-13 WI i=1-13 RUS i=1-13 VA IL X IL IL FIL IN Intermediate Demand sector i=1-13 IN X IL MI OH WI RUS Total Final Gross Final Intermediate Final Intermediate Final Intermediate Final Intermediate Final Deman d z=1-3 IN FIL Demand z=1-3 Demand sector i=1-13 MI X IL Deman d z=1-3 MI FIL Demand sector i=1-13 OH X IL Deman d z=1-3 OH FIL Demand sector i=1-13 WI X IL Deman d z=1-3 WI FIL Demand sector i=1-13 RUS X IL Demand z=1-3 Demand Output z=1-3 RUS FIL US FIL IL X IN IL FIN IN X IN IN FIN MI X IN MI FIN OH X IN OH FIN WI X IN WI FIN RUS X IN RUS FIN US FIN IL X MI IL FMI IN X MI IN FMI MI X MI MI FMI OH X MI OH FMI WI X MI WI FMI RUS X MI RUS FMI US FMI IL X OH IL FOH IN X OH IN FOH MI X OH MI FOH OH X OH OH FOH WI X OH WI FOH RUS X OH RUS FOH US FOH IL X WI IL FWI IN X WI IN FWI MI X WI MI FWI OH X WI OH FWI WI X WI WI FWI RUS X WI RUS FWI US FWI IL X RUS IL FRUS IN X RUS IN FRUS MI X RUS MI FRUS OH X RUS OH FRUS WI X RUS WI FRUS RUS X RUS RUS FRUS US FRUS 1) White cells : existing data in MWREIM 2) Grey cells : estimated data Three steps are involved in estimating final demands by 13 sectors by 6 regions.S. property type income and government final demand by industry and by state in order to estimate consumption. Ohio and Wisconsin) and the rest of the U. in the case of other state’s final demand consumption from IL goods and services. Hence. Not only there is no available data to identify them but also each state REIM only explains their final consumptions from their own industry sectors. Michigan. investment. (RUS).

and other value added by industry and state from Regional Economic Information System (REIS) in Bureau of Economic Analysis (BEA) and applied the ratio to total primary input value that is able to get by subtracting total industry output and total estimated interregional intermediate demands. we calculated the ratio of compensation of employee. In the current application. therefore. 2008).r which is not provided from MWREIM. Second is that our main purpose of this exercise focuses on which state in Midwest region has a responsibility for emissions generated within Midwest regional economy although each region’s trade with the rest of world should be considered in more comprehensive analysis (McGregor et al. This exercise is one of the key objectives of our ongoing research. Emission Intensity (EMI) coefficients 9 .3. First is that MW imports from ROW and export to 5 MW states and RUS broken down by commodity are not readily available. we focus on interregional trade between MW states with the effective assumption of a closed US economy. 2.i TOT IL consumptionIL → r = total consumptionIL − consumptionIL × ( ) r ≠ IL i TOT IL investmentIL → r = total investmentIL − investmentIL × consumptionri ∑ consumptionri ( ) r ≠ IL investmentri ∑ investmentri governmentri × ∑ governmentri r ≠ IL goverment i IL → r = total government ( TOT IL − government IL IL ) The next step is to estimate each state’s primary input matrix VAi . taxes on production and imports less subsidies. Therefore. There are two reasons we have not included ROW trade between MW and/or RUS. It is necessary to point out that we have not considered either region’s import or export with the rest of world (ROW) in our MW IO construction.

All point source SCCs and approximately 16% of area source SCCs have their associated SIC. This is because EPA has developed NEI from 1999. 92 constant millions $) where EMI is defined by the emissions per unit of activity (ton/92 constant million $) The general approach is to formulate the emission intensity from the 1999 National Emissions Inventory (NEI99). the emissions from NEI99 inventories based on SCC (National Emission Intensity based on Source Classification Code) are mapped into MW IO’s SIC code (13 sectors). solvent usage. We focus our attention here on CO EMI as a first attempt to do our environmental trade balance analysis since CO produces the highest proportion. The remaining 80% area source SCC are assigned to a particular SIC following the EGAS mapping (Economic Growth Analysis System) 1 developed by US EPA and some allocation through analysis of SCC and SIC coding.5μm). nitrogen oxide (NOx). we assume the emissions per real unit of activity in 1992 is constant with 1999 level and all emission changes result from only activity changes under the fixed EMI condition assuming no 1 http://www.. Note that the remaining 4% of the area sources related to household activities (e. space heating/cooling. The development of emission inventory is similar to traditional approaches. Under the Clean Air Act.The development of EMI for Midwest is derived from the REAL research funded by the US EPA STAR program (more detailed methodology can be found in Tao et al. and yard-waste burning etc) and on-road mobile sources are excluded in developing EMI. US EPA (Environmental Protection Agency) has set the criteria emission pollutants.5 with diameter less than 10 and 2.gov/ttn/chief/emch/projection/index. The main problem is that emission data are not available for 1992. volatile organic compound (VOC) and ammonia (NH3) as a air quality standard. First. in which emissions are determined by emission intensity and levels of emission activities: E (EMI) = P (Emission.g. 40% of total 7 emission pollutants in terms of physical volume (tons) in 1999. sulfur dioxide (SO2).html 10 . 2007). tons per year) / X (sectoral output .. Therefore.epa. particular organic compound (PM10 and PM2. carbon monoxide (CO).

57 1.57 0.66 MI 130.98 0.84 1.50 1.52 0. But the sectors with large EMI are not necessarily the ones with large total emissions since economic activity plays its role of producing total emissions.71 0.93 0.77 1.48 0.49 2.54 1.93 1. sector 6 (primary metal products) and sector 12 (other durable manufacturing) are in top three EMI (table 2).24 8.54 0. The resulting set of CO emission intensity coefficients for MW and RUS are shown in Table 1.68 0.43 0.44 0. Electronic and other Electric Equipment 10.56 0. Top three sectoral CO EMI IL sector 1 sector 6 sector 2 IN sector 1 sector 6 sector 12 MI sector 1 sector 2 sector 6 OH sector 1 sector 6 sector 2 WI sector 1 sector 13 sector 6 RUS sector 1 sector 6 sector 2 11 .78 0.98 0.51 0.62 1.88 2.51 3.52 1.61 RUS 43.54 0.76 7. Mining 3.45 1. Within the 13 sectors.48 0.86 1. Forestry and Fisheries 2. Construction 4.55 1. Food and Kindred Products 5.44 3.57 2. Agriculture. Transportation Equipment 11.63 0. CO emission intensity coefficients for MW and RUS Ton of CO per $1 million real output Sector 1.34 0.27 4. Table1.46 WI 71.75 2.23 0.emission technology changes in order to calculate environmental trade balance with 1992 MWIO table.45 0. Service. Primary Metals Industries 7.08 IN 57.83 0.44 0. TCU.72 1.55 0.57 0.84 0.18 0.47 0. forestry and fisheries) has the largest CO EMI for all MW states and RUS. Fabricated Metal Products 8. sector 1 (agriculture.26 OH 96.26 3.53 0. Other Non-durable Manufacturing Products 12.93 18.30 6.40 Table 2.80 1.64 0.54 0.56 1.87 1. Other Durable Manufacturing 13.86 1.54 1.55 0.76 2. and Government Enterprises IL 54. Industrial Machinery and Equipment 9.56 0.31 2.46 0. Chemicals and Allied Products 6.43 0.56 0.38 5.43 0.42 0.54 0. Sector 2 (mining).42 0.47 0.

05 4652.31 631421.68 559667.59 741673.08 32200.84%) 216624 (2.51 10514.08 1066129.00%) (100.42 5245.49 650854.56 1110027.46%) (14.00%) 12 .22 7457.29%) 3144 (2.89%) (16.58 6512.05 27094.15%) 6406 (4.49 1865.87%) 345389 (3.39%) (100.23 2677.06 8319.06 23541.97 960966.75%) 412899 (4.00%) RUS 9750601.73 3428.86 7973668.29 653966.61 (82.87%) Total 11646991.18 21304050.97 7480.3.95 292564. (7)) against the gross sectoral outputs from the MW interregional IO tables and shown in Table 3.76 620.20 22431.52 6615965.37 290553.55 16619.19%) 486957 (5.65%) 4789 (3.34%) 190636 (2.82 6052.09%) 7806899 (82.36%) 5907 (4.87 127693.75%) (21. we can estimate direct CO emissions generation by sector in each region with the MW environmental IO system under Muskgaard and Pedersen(2001)’s ‘production accounting principle’.52 853489.97 1133851.00%) (100.63 6436.44%) OH 523784.97 10905.71 9472.97 7726.58 3996.34 7288.86 60122.73 313380.92 (3.11 50925.02%) 2917 (2.26 228587.53%) 116248 (83.12 10474.41%) (25.67 135045.38 (100.44 25707171.60 719076.24%) WI 313736.51 7117.95 4522.77 5758.05 12173.57 371864. CO attribution analysis for MW and the rest of US Firstly. Direct CO pollution generated in MW and RUS in 1992 (tonnes) Sector 1 2 3 4 5 6 7 8 9 10 11 12 13 Total Total contribution of US Total contribution within MW Total output (92 millions $) (%) Total employment (thousands) (%) IL 429291.70 15883.93 6084.96 28403.89 1150.21 8250.37 3970.46 5381.63 19976.07 98834.14 8620.91 34391.44 223103.33 919658.60 21389.74 6697.09 30420.89 87376.01 278275.89 20898.60%) IN 183393.90 5858.26%) MI 446183.07 22044.00 107701.87 144915.31 52064.67 39580.42 (2.51 933051. P = EX y ⎛ p11 ⎜ y ⎜ p21 y (7) ⎜ p31 y ⎜ p41 y ⎜ p51 ⎜ py ⎝ 61 y p12 y p22 y p32 y p42 y p52 y p62 y p13 y p23 y p33 y p43 y p53 y p63 y p14 y p24 y p34 y p44 y p45 y p46 y p15 y p25 y p35 y p45 y p55 y p56 y p16 y p26 y p36 y p46 y p56 y p66 ⎞ ⎟ ⎟ ⎟= ⎟ ⎟ ⎟ ⎠ ⎛ e1x ⎜0 ⎜ ⎜0 ⎜0 ⎜0 ⎜0 ⎝ 0 x e2 0 0 0 0 0 0 x e3 0 0 0 0 0 0 x e4 0 0 0 0 0 0 x e5 0 ⎞ ⎟ ⎟ 0 ⎟ 0 ⎟ 0 ⎟ x⎟ e6 ⎠ 0 0 ⎛x ⎜ 11 ⎜x ⎜ 21 ⎜ x31 ⎜x ⎜ 41 ⎜x ⎜ 51 ⎜x ⎝ 61 x12 x22 x32 x42 x52 x62 x13 x23 x33 x43 x53 x63 x14 x24 x34 x44 x54 x64 x15 x25 x35 x45 x55 x65 x16 ⎞ ⎟ x26 ⎟ ⎟ x36 ⎟ x46 ⎟ ⎟ x56 ⎟ ⎟ x66 ⎟ ⎠ Table 3.65 156789.79 5515.25 3314.08 133896.83 19807.64 5268.10 9845.99 621282.87 963122.12 588005.57 (4.42 15226.95 205143.49 (3. The direct CO generation in each sector is calculated by multiplying the direct EMI (eq.69 858193.37 (2.63%) (21.

Table 5 shows how CO spillover or trade occurs between MW and RUS as accounted for in a conventional Type I attribution analysis using the accounting framework. 405.6% of CO pollution generated in RUS is explained as a result of MW final demand expenditure. policy and academic arenas. the largest CO pollution generator in MW. However. Michigan and Illinois are second and third contributor of CO pollution. forestry and fisheries) ranks first as a direct CO emitter. 13 . Top three sectoral CO emitters IL sector 1 sector 13 sector 6 IN sector 6 sector 1 sector 13 MI sector 1 sector 13 sector 10 OH sector 1 sector 13 sector 6 WI sector 1 sector 13 sector 11 RUS sector 1 sector 13 sector 11 Table 3 shows that total direct CO pollution generated from MW states is 17. CO pollution generated in MW by production supporting RUS final demands is bigger than the pollution generated in RUS by production supporting MW final demands. That is. The results suggest that (in our accounting year of 1992) 55.365 tonnes of CO pollution. 14. sector 1 (agriculture. This direct analysis explains total direct CO emissions through the purchase of goods and services in each state.9% of CO pollution generated in MW is to support. followed by sector 13 (TCU. and government enterprises) except Indiana where sector 6 (primary metal products) contributes the biggest CO emission (refer to table 4). In equations (4) to (6).75% of CO pollution. RUS final demand. only 9.Table 4. In terms of sectoral level emission.13% of total CO pollution in US. of CO pollution in MW is produced from Wisconsin. The smallest proportion. Within MW states. we are likely to be more interested in what share of pollution generation in each state is indirectly attributed to the final demands of other states or RUS. how economic activity in one region affects pollution generation in others. from the consumption accounting perspective that is gaining increasing attention in the public. On the other hand. services. Ohio directly generates 25. there is a positive CO trade balance for MW. directly or indirectly. Consequently. respectively.34%.

12%) 228092.50 (2.00%) 741673.82 (9.03%) 1058925.48 (56.92 (100.20%) 593395.59%) 39928.75 (4.08 (27.36%) 21304050.18%) 631421. Table 6.1% of locally generated CO pollution supports external consumption – i.2%) 541660.99%) 42191.89 (0.9%) 2054013.0%) 2054013.89%) 21192.00%) 963122.6%) Table 6 summarises this information for each state in turn.93%) 509706.49 (28. reading along the Il row.38 (1.84 (58. it has the highest outward trade in CO (reading along with WI row.1%) 729491.82 (3.82 (91. The CO trade balance between MW states and RUS (tonnes) : Type I input-output CO pollution supported by final consumption demand: IL Pollution generated in : IL IN MI OH WI RUS 391390.82 (3.00 (1.38 (100.81%) 419672.30%) 410997.86 ( 9.58 (2. There is a negative CO trade balance for Illinois with all of other MW states. in the case of Illinois.43 ( 9. CO trade balance between MW and RUS Environmental trade balance MW CO pollution supported by RUS demand RUS CO pollution supported by MW demand MW CO trade surplus with RUS IL CO pollution supported by other MW and RUS’s demand IN CO pollution supported by other MW and RUS’s demand MI CO pollution supported by other MW and RUS’s demand OH CO pollution supported by other MW and RUS’s demand WI CO pollution supported by other MW and RUS’s demand RUS CO pollution supported by other MW and RUS’s demand 2459379.14 (0.84 (1.3%) 574460. On the other hand.7%) 694473.10 (7.79 (3.52%) 611383.92 (4.49 (100. only 58.31%) 779943.26 (56.61%) 6805.72 (41.58%) 542710.58 (0.61 (100.94 (6.42 (100.79 (3.89%) 21709416. 42% of CO emissions generated in Illinois are generated to support final consumption demand in Illlinois. where only 9% of CO emissions generated in Wisconsin are associated with production to supports final demand in Wisconsin).28 (3.96 (72.57 (100.63 (18.24 (0.72%) 404360.04 (4.06 (2.26 (3.00%) IN MI OH WI RUS Total regional emission of CO 35181.34%) 276263.45 (35.92%) 1133851.66%) 114373.45%) 25707171.77 (53.e.13 (84.1%) 532045.95%) 24126.e.52 (1.84 (90.12 (3.67%) 44783.66%) 68167.19 (64.60%) 474059.27 (5.73 (65.78 (0.85%) 15889.05 (50. In Table 7 we go onto to examine the CO trade balance between MW states.27 (2.11%) 11679.26%) 26633.92%) 268649. 14 .35%) 933051.02%) 711739.32%) 56960.08%) 6701.39%) 145307.87%) 5897.15%) 37769.29 (55.97 (71.59%) 15582.72%) 6420.37 (100. Table 5 shows that Wisconsin is the state with the highest share of its CO emissions (91%) produced to meet external (other MW and RUS) final demand – i.22 (2.00%) Looking at each Midwest state.04%) 209627.00%) Total 1337398.00%) 411553.00%) 20963.33 (0.6%) 405365.27%) 34536.86 ( 9.Table 5.10%) 33928.07 (2.68%) 19250036.

05 -11136. However. CO trade balance between MW states Environmental trade balance between MW states (IL→IN)-(IN→IL) (IL→MI)-(MI→IL) (IL→OH)-(OH→IL) (IL→WI)-(WI→IL) (IN→MI)-(MI→IN) (IN→OH)-(OH→IN) (IN→WI)-(WI→IN) (MI→OH)-(OH→MI) (MI→WI)-(WI→MI) (OH→WI)-(WI→OH) -20657.24% of total CO emission in WI 15 . note that CO generated in Wisconsin in production to support other MW states final demands is greater than the pollution generated in other MW states in production supporting Wisconsin final demands. However.80 -14158. the Wisconsin CO trade surplus with Illinois.08% of total CO emission in IN 0.79% of total CO emission in IN 2. the figure of 114373 tonnes from the Illinois entry of the Wisconsin row of Table 5 – ‘exports’ of CO to Illinois – minus the 6701.87 -24346. this transaction only accounts for 10% of the output in Wisconsin supported by Illinois.29 -9991.71 25975.73 -107672. Agriculture.53% of total CO emission in MI 1.This implies that the pollution generated in Illinois in production to support final consumption demands in other MW states is less than the pollution generated in other MW states in production to supporting Illinois final demands.83% of total CO emission in WI 2.05% of total CO emission generated in Wisconsin.05 tonnes (i. It is relatively high direct CO-intensity of production (71.86 tonnes of CO per $1million output from Table 1) that underlies this core element of the CO trade balance between Wisconsin and Illinois.672.e.58 Wisconsin entry in the Illinois row – ‘imports’ of CO from Illinois) is relatively big. For example. Meanwhile.54 -5259.16% of total CO emission in MI 0. note that 84% of the CO emissions in Wisconsin that are supported by Illinois final demands are generated in the Sector 1.51 607.58% of total CO emission in WI 2.92 -13989. accounting for 17. Forestry and Fishing.58 2.05% of total CO emission in WI 1. Table 7. 107.70% of total CO emission in MI 1.23% of total CO emission in OH 17.

79%) MI 32838.42%) 196174.55 (2.14 (3.19%) IN 28138.88 (5. and government consumption) in Illinois than other MW states do.92%) IN Investment 1926.55 (0.50%) 8949.21%) 31841.81 (1.68 (0.26 (1.23%) 3828.53%) HH 19886.29%) Govt 44360.29% of CO emission in Illinois are related with Ohio’s investment and government consumption.74 (2.00%) 9344.77 (0.25 (0. investment and government final demands in table 8 (e.19%) 5079.21%) 2511.89%) 289678.12%) 3550. Wisconsin and RUS generate relatively bigger amount of CO to supply their productions for all three final demands (household.52 (0.38%) 28311.24%) 113499.68 (0.75 (0.08 (0. There is a difference in the extent of interregional CO spillover between these final demand types.33%) 3105.73 (34.85 (2.98%) 4406.41%) OH 27461.35%) 80936.15%) Govt 2313.27%) 1401.26 (3.29 (1.10 (0.23%) HH 11641.29%) 3579.86 (4.08%) 4696.31%) 1699.05 (0.62 (2. the three Illinois-Illinois entries in the top left of the table sum to the single top left entry in Table 5).82 (0.31%) RUS 535613.22%) 1328.17 (0.68 (0.95 (3.93 (1.g.99%) 11961.39%) 208986.43 (0. Table 8.28%) 3311.31 (2.78%) 3153. which is the biggest proportion among other MW household consumption supported by Illinois. CO emission generated in Michigan is largely associated with all three final demands in Ohio and vice versa.27%) 40381.13%) 142397.62) (0.90 (0.96 (4. 0.15 (0.25%) 17725.57%) 62626.71 (1.30 (0.35 (0.41 (3.22%) 48247.46 (2. 2.38%) 16 .41%) 2228.68%) 20396.66 (19.21%) 39191.24 (0.49 (0.70%) 33944.36%) MI Investment 1765.28%) 2033. On the other hand.75%) 7301.57 (21.42 (0.51%) Investment 28031.46%) 4409.13% of CO emission in Illinois is generated as a result of Indiana household consumption.75 (2.19%) Govt 2174.22% and 0.39%) 7825.41 (0. The CO trade balance by three final demands between MW and RUS (continued over page) IL HH IL 318998.83 (0.15 (16. For investment and government consumption.11 (0.77 (0.26%) 2684.25%) 28037.06 (0.33 (3.42%) WI 102968.52 (0.20%) 32583. capital investment.52%) 39265. CO emissions in Indiana.70 (1.More detailed interregional CO trade is shown by household.00 (0.

54%) 37839.66 (1. comparatively.35%) 459875.79 (1.61 (100.32 (0.77 (4.54 (0.16%) 2641.93 (0.13%) 7862.32 (0.32%) 45410.20%) MI 58912.69 (6.06 (0. our main finding from the empirical analysis here is that 55.63 (11.50 (4.40 (0.6% of CO emission generated in RUS supports consumption in Midwest.47%) 2795.42%) 102867.48 (0.17%) 67526.80 (2.77 (0.30%) 51151.72 (8.34 (0.48 (38.53%) 100060.28 (0.69%) 3693771.82%) 67501. 9.33%) Govt 73549.62%) WI 16385. Thus.88%) 87941.2% of the total CO generated in Midwest. The construction of Midwest interregional input-output table is provisional at this moment due to availability of appropriate data.66 (4.24%) Govt 2702.58 (5.83%) 3129.86%) 91764.72 (9.39 (40.42 (100.34%) 933051.39%) 42693.00%) 21304050.87 (0.09 (26.53 (0.29%) 4669.13%) Govt 857.99%) 1349213. We find out that Wisconsin is a biggest net loser in terms of CO emission pollutants within the Midwest region (due to the importance of Wisconsin’s very CO-intensive Agriculture. For more accurate analysis.20 (0.Table 8.48%) HH 4392.71 (10.12%) HH 361588.00%) 1133851.64%) 57129.37 (100.00%) 631421.92 (100.44%) WI Investment 1451.76%) 93015.10%) 43056.88 (66.56%) 306212.49 (100.12 (0.29 (0.22%) 6162.) OH HH IL 16401.00%) 741673. Another key finding is that Michigan and Ohio are very closely related to each other in terms of CO emission trade by production supporting each other’s consumption.85 (0.00%) Total 4.09%) 983.01%) 24872.57 (0.50%) 14207051.38%) 3964.65 (9.16 (3.70 (0.62 (0.13%) 778. we are setting up a more updated and detailed system of MW IO table with improved 17 .85 (7.73 (4.69%) RUS Investment 38920.46 (6.04%) 2656. the results of our environmental trade balance analysis should be regarded as an example of how we can examine environmental spillovers between Midwest states using the environmental IO framework.67%) Investment 2087.47%) 51447.12%) OH 301820.76%) IN 23704.85 (48. The CO trade balance by three final demands between MW and RUS (cont.63%) 6125.54 (0.36%) 1154. Summary and further research We use an interregional input-output framework for the Midwest regional environmental attribution and trade balance analysis in this study.75%) 264204.57 (100.18 (0.17 (5.47 (1.11 (42.25%) 27419.45 (0.60%) RUS 355686.27 (0.62%) 407889.65 (0. There is a CO trade surplus between Midwest and RUS. Illinois is a net gainer.22 (6.01%) 1921.73 (17.9% of CO emission generated in Midwest supports consumption in RUS while only 9.00%) 963122.07%) 6404. However.95 (0.12%) 1511.13 (35.08%) 850.41%) 4443.92 (1. Forestry and Fishing sector in interregional trade) and.

18 . We will also develop this through construction of an interregional SAM system for the Midwest and application of this as the core database of an interregional CGE modeling framework.interregional trade flow data (for 2007).

Agriculture. Government Expenditure 19 . Mining 3. Investment (Residenatial + Nonresidential and Equipment) 3. Nondurables. Industrial Machinery and Equipment 9. Other Durable Manufacturing 13. Other Non-durable Manufacturing Products 12. Consumption (Autos and Parts. and Service) 2. Chemicals and Allied Products 6. Construction 4. Transportation Equipment 11. Service. Fabricated Metal Products 8. Mnemonics of industry sectors and final demands Industry sector i=1~13 1. Forestry and Fisheries 2. and Government Enterprises Final demand z=1~3 1. TCU.Appendix 1. Other Durables. Primary Metals Industries 7. Food and Kindred Products 5. Electronic and other Electric Equipment 10.

and Hewings. Regional Input-Output Modelling: New Developments and Interpretations. Donaghy. G. T... Papers in Regional Science 75. G. Dewhurst. Israilevich. 20 . Barrett.R. 2009.A. Tao. K.J. 2007. Hewings. In: J. 662-672. Hewings. G.D. 141-166.J. A socio-economic method for estimating future air pollutant emissions-Chicago case study. J. The choice of input-output table embedded in regional econometric input-output models. Williams.. P. G.K. 1-12. and Turner. Forecasting Structural Change with a Regional Econometric Input-Output Model.G. Israilevich.ecolecon. Lenzen. P. R.D.2009. Examining the global environmental impact of regional consumption activities-Part1:A technical note on combining input-output and ecological footprint analysis. M.08. Energy Policy 29. K. and R.L.D. Lenzen. 1997.. Sonis. Atmospheric Environment. T. K.. McGregor..R. J.R.. and Barrett.R.. The CO2 ‘trade balance’ between Scotland and the rest of the UK: Performing a multi-region environmental input-output analysis with limited data. 103-119. J. C. Z.S. and Mahidharha.. J. and Schindler.J. 2001. K. Ecological Economics 66. A.R. Jensen (eds. Ecological Economics.References Conway.R.T.1016/j.. Munksgarrd.. Journal of Regional Science 37. 2007.. Conway. 1991..). Examining the Global Environmental Impact of Regional Consumption Activities – Part 2: Review of input-output models for the assessment of environmental impacts embodied in trade. 37-44 Wiedmann. Turner. M.. doi:10. M. The Washington Projection and Simulation Model: A Regional Interindustry Econometric Model.. 2007.. 565-590. and Pedersen. P. 1990. Ecological Economics. J. 1996.G. Ecological Economics 62... An Empirical Comparison of Regional Multipliers. Swales. Wiedmann.G. and Turner.D. Wiedmann. 61(1): 15-26. 327-334.. CO2 accounts for open economies: producers or consumer responsibility?.. Aldershot. 178-195. R. A review of recent multi-region input-output models used for consumption-based emission and resource accounting. Avebury. 2008. Schindler.. International Regional Science Review 13.R. H. Hewings.026...S. K.

Sign up to vote on this title
UsefulNot useful