Journal of Agricultural and Applied Economics, 28,2(December 1996):393-407 0 1996 Southern Agricultural Economics Association Evaluating the Impacts of Agricultural Exports on a Regional Economy Roman I. Bairak and David W. Hughes ABSTRACT Agricultural exports are important to many regional economies, as is the case for agricul- tural exports either produced in or shipped through Louisiana. A hybrid (revised and verified) IMPLAN model of the Louisiana economy is used to estimate the direct and indirect impact of agricultural exports. Original model estimates of foreign exports lacked holistic (overall) accuracy. However, other, more general uses of the model were unaffected by this lack of accuracy. While the contributions of agricultural exports to the state econ- omy were substantial, impacts were concentrated in unprocessed products. Increasing the export of processed agricultural products should enhance economic activity. Key Words: agricultural exports, holistic accuracy, IMPLAN, input-output models, pro- cessed exports. International trade is important to the econom- ic well-being of a nation and a region. In 1991, the United States was the world’s largest trad- ing nation, accounting for 14% of world im- ports and 12% of world exports. The European Community, Canada, and Japan are the major U.S. trading partners. However, exports to many developing countries, particularly in Asia and Latin America, have increased in re- cent years (see “The United States, ” in Trade Policy Review). Further increases in U.S. trade are project- ed due to signing of the General Agreement on Tariffs and Trade (GATT) and of the North American Free Trade Agreement (NAFTA). The former will eventually lower trade barri- ers on a worldwide basis, while the latter will The authors are former graduate research assistant and assistant professor, respectively, in the Department of Agricultural Economics and Agribusiness, Louisiana State University, Baton Rouge. The authors acknowledge the helpful comments of two anonymous reviewers. eliminate most trade barriers among the Unit- ed States, Canada, and Mexico. Exports of U.S. agricultural commodities increased dramatically in the 1970s, By 1990, agricultural exports accounted for about 15YO of total U.S. merchandise exports. The United States is the world’s leading exporter of feed grains, wheat, livestock products, soybean products, horticultural products, and rice (“The United States,” Trade Policy Review). Yet, U.S. agricultural exports are concentrated in low-value, often unprocessed products (Burfisher and Missiaen). Louisiana ports are major points of depar- ture for U.S. agricultural exports (Falgout). Louisiana ports shipped $16.5 billion worth of exports in 1992, making the state the sixth largest port of exit in the nation. Between 55% and 60% of all commodities shipped through Louisiana ports were agricultural products (Falgout). Louisiana also produces several agricultur- al commodities that depend heavily on foreign markets, such as cotton, rice, and soybeans.
Transcript
Journal of Agricultural and Applied Economics, 28,2(December 1996):393-4070 1996 Southern Agricultural Economics Association
Evaluating the Impacts of AgriculturalExports on a Regional Economy
Roman I. Bairak and David W. Hughes
ABSTRACT
Agricultural exports are important to many regional economies, as is the case for agricul-tural exports either produced in or shipped through Louisiana. A hybrid (revised andverified) IMPLAN model of the Louisiana economy is used to estimate the direct andindirect impact of agricultural exports. Original model estimates of foreign exports lackedholistic (overall) accuracy. However, other, more general uses of the model were unaffectedby this lack of accuracy. While the contributions of agricultural exports to the state econ-omy were substantial, impacts were concentrated in unprocessed products. Increasing theexport of processed agricultural products should enhance economic activity.
ing nation, accounting for 14% of world im-ports and 12% of world exports. The EuropeanCommunity, Canada, and Japan are the majorU.S. trading partners. However, exports tomany developing countries, particularly inAsia and Latin America, have increased in re-cent years (see “The United States, ” in Trade
Policy Review).
Further increases in U.S. trade are project-ed due to signing of the General Agreementon Tariffs and Trade (GATT) and of the NorthAmerican Free Trade Agreement (NAFTA).The former will eventually lower trade barri-ers on a worldwide basis, while the latter will
The authors are former graduate research assistant andassistant professor, respectively, in the Department ofAgricultural Economics and Agribusiness, LouisianaState University, Baton Rouge.
The authors acknowledge the helpful comments oftwo anonymous reviewers.
eliminate most trade barriers among the Unit-ed States, Canada, and Mexico.
Exports of U.S. agricultural commoditiesincreased dramatically in the 1970s, By 1990,agricultural exports accounted for about 15YOof total U.S. merchandise exports. The UnitedStates is the world’s leading exporter of feedgrains, wheat, livestock products, soybeanproducts, horticultural products, and rice(“The United States,” Trade Policy Review).
Yet, U.S. agricultural exports are concentratedin low-value, often unprocessed products(Burfisher and Missiaen).
Louisiana ports are major points of depar-ture for U.S. agricultural exports (Falgout).Louisiana ports shipped $16.5 billion worth ofexports in 1992, making the state the sixthlargest port of exit in the nation. Between 55%and 60% of all commodities shipped throughLouisiana ports were agricultural products(Falgout).
Louisiana also produces several agricultur-al commodities that depend heavily on foreignmarkets, such as cotton, rice, and soybeans.
394 Journal of Agricultural and Applied Economics, December 1996
Certain processed food products, such as Lou-isiana poultry products, are also shipped over-seas. Louisiana enjoys a strategic location forexport markets in Latin America. As a result,the state may receive a disproportional benefitfrom expected increases in agricultural ex-ports. This benefit will be enhanced if increas-es in Louisiana-produced agricultural exportsare concentrated in higher valued, processedcommodities.
This article summarizes findings regardingthe impact of agricultural exports on the Lou-isiana economy using a revised (a so-calledhybrid) input-output (I-O) model constructedwith IMPLAN (Alward et al.). Emphasizedhere are revised estimates of agricultural ex-ports in the model. Estimates of foreign ex-ports can influence accuracy of study resultswhen the variable is of direct concern. How-ever, estimates of foreign exports influence es-timates of regional supply available for re-gional consumption for IMPLAN models inall situations. Hence, a comparison is made ofthe effect of verified versus unverified foreignexports on general (holistic) model accuracyin evaluating the impact of foreign exports andin evaluating changes in final demand that areunrelated to foreign markets. Model resultsbased on revised estimates of agricultural ex-ports are used to examine the impact of pro-cessed versus unprocessed exports. Finally,model results are reviewed, summarized, andused to make policy recommendations.
Input-Output Models and InternationalExports
International trade activity influences mostsectors in a regional economy. Sectors withdirect sales to foreign markets are linked tomany other local sectors. Because of these in-terindustry linkages and because of projectedgrowth in exports, agricultural exports are ex-pected to be an important determinant of fu-ture state economic growth. I-O analysis wasselected as the analytical tool for this studybecause of its ability to analyze interdepen-dencies among industries in an economy (Mil-ler and Blair).
I-O models have been used in a number of
studies examining the regional and nationalimpacts of foreign trade. Belous and Wyckoffused I-O analysis to look at the net effect ofincreases in imports and exports on the U.S.economy. Holding the composition of importsand exports constant, they concluded that amillion dollar increase in exports generatedmore jobs than were destroyed by a milliondollar increase in imports. Martin and Hollandexamined the sources of change in nationaloutput from 1972 to 1977. They concludedthat international trade had contributed to netincreases in U.S. total output. Balassa and No-land applied forecasts of U.S. imports and ex-ports for the year 2000 to I-O coefficients.They projected large employment losses in theapparel, footwear, and automobile industriesand large employment gains in the coal, com-puting equipment, and machinery industries.Sharp employed an I-O model to predict theeffects of increased exports to Mexico on theTexas economy. He predicted that increasedexports to Mexico could generate a total of113,000 state jobs. Hughes, Holland, andWandschneider used results from an IMPLANmodel of the Washington state economy to ar-gue that growth in exports could help fill an-ticipated losses in employment due to cuts inmilitary spending.
Holistic Accuracy and Hybrid I-O Models
An I-O model of the 1985 Louisiana economybuilt with the IMPLAN (IMpact PLANning)model building system is used in this study.IMPLAN is one of several so-called ready-made model building systems that provide ac-cess to databases and model constructionmethods in a computer software package.Such systems allow researchers to constructmodels by combining the national I-O tablewith secondary regional data. Hybrid I-Omodels are ready-made models that have beenadapted and verified for a particular set of usesby incorporating additional secondary and pri-mary data about the regional economy (Jen-sen; Brucker, Hastings, and Latham).
The Louisiana IMPLAN model had previ-ously been converted to a hybrid model byapplying secondary employment and income
Bairak and Hughes: Regional Impacts of Agricultural Exports 395
data, by introducing new regional absorptionvectors for eight major agricultural producingand processing sectors based on published andunpublished budgets for state firms, and by ad-justing levels of regional exports and importsfor 96 commodities based on expert opinion(Hughes). A major focus here is on incorpo-rating other primary and secondary data notinitially used in IMPLAN to more accuratelyestimate sales to foreign markets by Louisianaagricultural producers.
The concept of holistic versus partitive ac-curacy provides a guideline in constructing re-gional hybrid input-output models. Partitiveaccuracy can be defined as the closeness ofany given cell in an input-output table to theactual but unknown value for the representedeconomy. Holistic accuracy concerns the abil-ity of an input-output table to represent theessential elements of an economy (Jensen).Only the latter is a worthy goal in the con-struction of any regional input-output modelbecause the accuracy of relatively small valuesin an input-output table has little bearing onmodel results, as demonstrated by Jensen andWest.
The concepts of partitive and holistic ac-curacy are best understood in terms of modeluse—a fact that has been underemphasized inthe literature in our view. For example, as-sume that a major sector of a regional econ-omy is poorly represented in a regional eco-nomic model, but that the model is anadequate representation in other respects. Fora general study of the economy or for a studywhere the sector is directly or indirectly af-fected in a substantive way, the model in ques-tion would lack holistic accuracy (i.e., use ofthe model would cause researchers to drawmisleading inferences). However, if the sectoris not affected in a substantive way in the sce-narios evaluated for a given study, then themodel is accurate from a holistic perspectivefor that particular study (if not for other stud-ies). This perspective provides the justificationfor examining the impact of changes in exportestimates on the holistic accuracy of the IM-PLAN model.
While IMPLAN-based models are widelyused in regional impact analysis, little analysis
has been conducted concerning the accuracyof IMPLAN-based estimates of foreign trade.Therefore, a question arises concerning the ho-listic accuracy of IMPLAN in studying the re-gional effects of trade policies. Perhaps moreimportant-because estimates of foreign tradeare deduced from estimates of regional com-modity supply-estimates of the former caninfluence the holistic accuracy of IMPLANmodels for studies unconcerned with foreigntrade. Comparing model results under originalexport estimates to model results under newexport estimates should provide an indicationof how holistic accuracy is affected when thevariable is of direct concern and when esti-mates of foreign exports only indirectly influ-ence model accuracy.
Estimating New Levels of Louisiana
Agricultural Exports
The hybrid IMPLAN I-O models for Louisi-ana in 1985 provided initial estimates of in-ternational exports for state firms. Estimates ofexports for 20 agricultural industries in theIMPLAN I-O model were supplemented bysecondary and primary data. Using the Stan-dard Industrial Classification (SIC) sectorcodes, information was gathered on agricul-turally based products classified as productionagriculture (SIC 01 and 02), processed foods(SIC 20), wood and lumber (SIC 24), and pulpand paper products (SIC 26).
Assume that the level of agricultural ex-ports shipped through Louisiana is known.Also assume that all foreign agricultural ex-ports originating in Louisiana are shippedthrough its ports. Estimates of foreign exportsof Louisiana agricultural commodities are thenobtained by determining the proportion of theagricultural commodities shipped throughLouisiana that are produced in Louisiana.
The New Orleans U.S. Customs Districtexport data provided an estimate of agricul-tural products shipped through Louisiana. Thedistrict includes all Louisiana ports and portsin Mississippi, Tennessee, and Arkansas thatare situated on the Mississippi River and itstributaries. Of these ports, only the Louisianaports of New Orleans, Baton Rouge, Lake
396 Journal of Agricultural and Applied Economics, December 1996
Charles, and the port of South Louisiana (theMississippi River between Baton Rouge andNew Orleans) are capable of accommodatingocean-going vessels (U.S. Army Corps of En-gineers). Thus, all export data for the customsdistrict were compiled at one of these ports ofexit.
The value of all agricultural commoditiesshipped through the New Orleans CustomsDistrict for 1989 through 1992 was taken from“U.S. Exports and Imports of Merchandise”on CD-ROM (U.S. Department of Commerce,Bureau of the Census). To be consistent withthe 1985 IMPLAN model, these values weredeflated to 1985 dollars using the appropriateproducer price index.
The total estimated annual average valuefrom 1989–92 of agricultural exports shippedthrough Louisiana ports was $10.897 billionin 1985 dollars. Food grains, oil-bearingcrops, and feed grains were the three IMPLANindustries with the largest share of the valueof agricultural exports shipped through Loui-siana ports. These three industries includedwheat, rough rice, corn, sorghum, and soy-beans, which contributed more than 75% (or$8.212 billion) in total agricultural exports an-nually shipped through Louisiana over this pe-riod.
A telephone survey of major agriculturalexporters in Louisiana was conducted to ob-tain the percentage of agricultural exports go-ing through Louisiana ports that originatedthere. A stratified random sample, based onthe four-digit Standard Industrial Classifica-tion (SIC) code, was used to ensure coverageof all agricultural exports. A list of 100 tradingcompanies that exported agricultural productswas drawn from the Louisiana Agricultural
Export Directo~ (Louisiana Department ofAgriculture and Forestry). Firms were con-tacted based on industry prevalence in theLouisiana economy and involvement in exportmarkets. For example, 10 firms involved withthe export of milled rice—an important stateindustry that is heavily dependent on foreignmarkets—were surveyed. On the other hand,only three firms involved in the export of cot-tonseed oil—a less important state industry—were contacted.
Across the 20 major industry categories inthe Louisiana IMPLAN model shown in table1, the number of surveyed firms ranged fromthree to 10. For seven out of the 20 IMPLANindustry categories, seven firms were repre-sented (the mode for the distribution of con-tacted firms by industry category).
The distribution of respondents indicatedbroad coverage of Louisiana agricultural ex-porters. For agricultural products that theyhandled, 63 of the 100 firms were willing toestimate the percentage produced in Louisianaout of the amount shipped through Louisianaports. The number of respondents in each ofthe 20 IMPLAN industry categories rangedfrom one to seven, with four respondents rep-resented in six of the 20 categories (the modeof the distribution of responding firms). Atleast three firms responded to the survey ques-tion for 15 out of the 20 IMPLAN industrycategories. The percentage of surveyed firmsin each IMPLAN industry category willing toprovide an estimate of the percentage level ofexports ranged from 33.3 % to 100’-ZO.For 16out of 20 of the IMPLAN industry categories,at least 50% of the surveyed firms were will-ing to provide an estimate of the percentageof Louisiana-produced commodities versusgoods produced elsewhere. Further, responserates tended to be high for important state in-dustries with large levels of exports, such as a7070 (seven firms) response rate for rice mill-ing and an 85,890 (six firms) response rate forlumber products.
The estimated value of agricultural exportsoriginating in Louisiana, the percentages ob-tained from the survey, and the levels of ag-ricultural exports moving through Louisianaare provided in table 1. Within each major IM-PLAN industry category, responses by eachfirm were given equal weight in assigning thesurvey-based coefficients. 1 Louisiana portsprovided an export channel for $9.989 billionin agricultural commodities produced in otherstates. Louisiana itself exported $962.632 mil-
1An equal weight was used because surveyed firmswere, in general, unwilling to provide information con-cerning the actual dollar value of exports.
Bairak and Hughes: Regional Impacts of Agricultural Exports 397
Table 1. Trade and Survey Data Estimates of Agricultural Exports Originating in Louisianaas an Annual Average, 1989–92 (millions 1985 $)
lion in agricultural products grown or manu-factured in the state.
The paper products category reflected thelargest level of Louisiana agricultural exports,with a value of $268.781 million (table 1).Rice milling was also a major contributor withstate agricultural exports of $135.047 million.Lumber, sugar processing, and oil-bearingcrops were other industries with high levels ofagricultural exports. These five industries to-gether were responsible for 67.9% of agricul-tural exports produced in Louisiana.
Other data sources were used to evaluatethe assumption that all Louisiana-producedproducts were shipped through the state. Theonly studies addressing this issue consisted ofa set of publications concerning movement ofsoybeans, wheat, oats, sorghum, and corn.Studies of product movements were not avail-able for other unprocessed or for any pro-cessed agricultural products. Researchers es-timated that 3.7?Z0 of Louisiana soybean
exports and 590 of Louisiana wheat exportswent through ports outside Louisiana (Larson,Smith, and Baldwin; Reed and Hill). Accord-ingly, exports of Louisiana oil-bearing cropswere increased from $77,824 million to$80.772 million. Exports of Louisiana foodgrains were increased from $25.099 million to$26.420million. Other studies examining feedgrains (corn, oats, and sorghum) indicated noexport of these Louisiana crops through portsoutside Louisiana (Fruin, Halbach, and Hill;Baldwin et al.; Hill et al,).z
2Results from these other studies also indicated atotal increase of exports for the three sectors of only2.9%, implying that the assumption of all Louisianaagricultural exports moving through Louisiana portswas generally acceptable. Estimates of Louisiana for-eign exports for agricultural crops were also comparedto estimates derived from these sources for unpro-cessed agricultural crops and to U.S. Department ofCommerce survey-based estimates for processed agri-cultural products (food processing, paper, and timber
398 Journal of Agricultural and Applied Economics, December 1996
Transportation, wholesale trade, and port
margins were included in the estimates of
Louisiana agricultural exports because the es-
timates were at the port of exit. The margins
were allocated from each of the agricultural
industries to the proper trade or transportation
sector.
The wholesale trade margin for agricultural
exports was allocated to the IMPLAN industry
category of other wholesale trade. Such treat-ment of trade margins is standard in I-O mod-els. The IMPLAN table wholesale margins forhousehold consumption were used to estimatetrade margins for all agricultural exports (Al-ward et al.). The estimated total wholesalemargin was $44.769 million (4.7Y0 of the totalvalue of state agricultural exports).
The transportation margin was allocated tothe IMPLAN motor freight transportation andwarehousing industry, to the water transpor-tation industry, or to a combination of the twoindustries. This allocation was based on as-sumptions concerning how Louisiana agricul-tural products moved to ports. Products wereassumed to move to port by truck, by barge,or by a combination of the two. The mode oftransportation was based on information ob-tained from three Baton Rouge companies:Eckstein Marine Co. (a water transportationfirm), SAIA Motor Freight Co. (a truck trans-portation firm), and Union Pacific (a railroadcompany). Transportation charges obtainedfrom these firms were also used in calculatingtotal transportation costs.
Distances from point of production to ex-port port also had to be calculated to obtaintotal transportation costs. Unpublished stateemployment data (Louisiana Department ofLabor) were used to distribute exports of foodprocessing, paper, and wood products amongthe nine state agricultural production districts.The geographical center of each district wasthen used to estimate the distances betweenpoint of production and port of export. Unpro-
products) for 1987–89. In all cases, estimates of for-eign exports used in this study were closer to estimatesobtained from these other sources than were the orig-inal IMPLAN estimates. (For additional details, seeBairak.)
cessed agricultural products were treated in thesame manner. But estimates of farm produc-tion in the agricultural production districts(from Zapata and Frank) were used to calcu-late the distribution of crop exports among thedistricts.
Port service charges (National Ports andWaterways Institute) were deflated to 1985dollars. The estimated port service charge was$4.75 per metric ton in 1985 dollars. For eachof the 20 agricultural industries, the estimatedweight of total exports by industry was usedalong with the per ton charge to estimate atotal port charge. The margin for port activityfor all agricultural exports was allocated to thewater transportation sector based on the ap-proach used in other studies (e.g., Yochum andAgarwal).
The total transportation cost of exportingagricultural products produced in Louisianawas estimated to be $27.973 million in 1985dollars. Total port charges were estimated at$11.146 million for all Louisiana agriculturalproducts. Together, port and transportationcharges represented 4.1 % of the total value ofLouisiana agricultural exports. Of the $39.119million total charges, $25.297 million was al-located to the motor freight transportation andwarehousing sector, and $12.787 to the watertransportation industry.3
Agricultural exports for most Louisiana in-dustries were larger than the estimates of Lou-isiana agricultural exports in the original 1985IMPLAN hybrid I-O model. The total estimateof agricultural exports for Louisiana industrieswas $880.816 million (table 2), reflecting anincreased value of $286.866 million, or 48.390larger than the same total in the original model.
Estimates of exports for 13 out of the 20industries increased in the new version of thehybrid IMPLAN model (table 2). Industrieswith considerable increases in current versusoriginal estimates of foreign exports included
3The estimate of the total trade and transportationmargin (8.9%) was compared to national estimates ofmargins for processed agricultural products found inthe U.S. Departmentof Commercecensus data. As ex-pected, the estimates used here were less than the na-tional values because of lower transportation costcharges. (For further details, see Bairak.)
Bairak and Hughes: Regional Impacts of Agricultural Exports 399
Table 2. Original versus New Export Estimates by Louisiana Agricultural Industries in the1985 Hybrid IMPLAN I-O Model (millions 1985 $)
Estimated LouisianaAgricultural Exports
($ roil.)
Original New ChangeIMPLAN Industry by Sector Code Model Model ($ roil.)
feed grains, dairy products, and bread prod-ucts. Other industries, such as oil-bearingcrops, had small increases in export estimates.New export estimates were slightly smallerthan original estimates for food grains, cannedfruits and vegetables, and other processed fats,feeds. New export estimates were markedlyless than original estimates for cotton, pro-cessed fish and seafood, and rice milling.
Differences in original estimates of for-eign trade of Louisiana agricultural productsand those calculated in this study may be ex-plained by differences in the years covered(1985 versus 1989–92).4 But differences may
4A review of the 1991 Louisiana ready-made IM-PLAN model indicated that these estimates were alsoproblematic. For example, exports by sugar producers(not sugar mills or refineries) of unrefined sugar (ex-clusively sugarcane in Louisiana) were reported at $53million when it is well known that Louisiana sugarcaneis never exported prior to milling and refining.
also be explained by the way in which esti-mates of exports are calculated in the IM-PLAN modeling system. For a given indus-try, Louisiana’s proportion of nationalcommodity output was used in the originalIMPLAN estimates to calculate Louisiana’sshare of national exports in that commodity.While this approach is standard procedure forIMPLAN models, it may yield inaccurate re-sults because of differences in commoditymixes at the regional and national levels. Fur-ther, the method does not account for the lo-cational advantage (for a state such as Loui-siana) or disadvantage (for a given interiorstate) of a region in moving goods to port ofexport. The large difference (48.3 Yo) betweenthe calculations of foreign exports found inthis study and those contained in the originalIMPLAN export estimate implies that IM-PLAN users should be cautious in using un-verified estimates of exports in evaluating the
400 Journal of Agricultural and Applied Economics, December 1996
impacts of such markets on regional econo-mies.
Effect of Foreign Trade Estimates on
Holistic Accuracy in Other Model Uses
Another relevant concern about model accu-racy arises because estimates of regional ex-ports to foreign markets can influence the ho-listic accuracy of IMPLAN-based models foruse in studies unrelated to trade analysis. Thesupply demand pool (SDP) coefficient is themaximum amount of regional supply that isavailable to meet regional demand. Or, it is theratio of regionally produced commodity sup-ply, net of foreign exports, to gross regionalcommodity demand. An SDP coefficient ofone means that regional supply at least equalsregional demand for the commodity in ques-tion. An SDP coefficient of less than one im-plies that the commodity will have to be im-ported even if the commodity is not adomestic export (Alward et al.).
The regional purchase coefficient (RPC)for a commodity is the ratio of local demandmet by local production to regional supply netof foreign exports. Hence, the ratio provides ameasure of how much local demand is satis-fied by local production. An RPC of 0.9 meansthat 109’oof the commodity consumed is im-ported into the area. Over time, if regionalfirms substitute imports for regional produc-tion, the RPC for the commodity would de-crease. As a result, the estimate of the regionalimpact of a given change in final demandwould decrease. RPCS for all nonservice com-modities in IMPLAN (1–445) are estimatedthrough an econometrically based procedure.RPC estimates for IMPLAN service commod-ities (446–528) are calculated based on ob-served 1977 state supply, exports, and im-ports. Because the SDP is the maximumamount of regional supply available to meetregional demand, it is an upper bound for theRPC values used in IMPLAN models (Alwardet al.).
A commodity’s SDP is calculated by firstsubtracting estimates of foreign exports fromgross commodity supply. Hence, foreign ex-ports always influence the coefficient. Foreign
exports influence the RPC for commoditieswhere the SDP coefficient equals the RPC(i.e., the independently estimated RPC is at itsSDP upper bound).
SDP and RPC values under the original ex-port estimates were compared to SDP andRPC values under the new export estimates.The comparison showed that under the newestimates of foreign exports, the SDP for 18commodities increased, while the SDP de-creased for 29 commodities. Similarly, theRPC increased for nine commodities and de-creased for 13 commodities. While most ofthese changes were small, a few commoditieshad large changes, such as the difference of0.4087 for condensed and evaporated milk(IMPLAN commodity 88).
To compare the potential effect of changesin RPCS on model estimates, the impact of a$10 million dollar change in final demand foreach of the 20 agricultural industries listed intable 2 was calculated for the state model withoriginal versus new estimates of foreign ex-ports. Changes in RPCS due to differences inthe estimates of foreign exports did not affectthe holistic accuracy of the model in this case.For example, estimates of the employment im-pacts under the two models only differed by0.3’ZO(7,487 versus 7,511). Substantially dif-ferent estimates of foreign exports of agricul-tural products had little impact on model re-sults. One can conclude that IMPLAN modelusers should not be too concerned with theeffect of estimates of foreign exports on ho-listic model accuracy where this variable is notof direct relevance.
Impact of Foreign Exports on theLouisiana Economy
A comparison of the level and composition ofthe impact of original and new estimates offoreign exports on the Louisiana economy alsoprovides insight into model holistic accuracy.If the impacts of foreign trade on the Louisi-ana economy under the two estimates are sim-ilar, then holistic accuracy is retained for theoriginal model for the purpose of examiningthe effects of such markets on the state econ-omy. Holistic accuracy in this case would im-
Bairak and Hughes: Regional Impacts of Agricultural Exports
ply that researchers using IMPLAN in exam-ining the regional effects of trade policiesshould not be overly concerned about the ac-curacy of the method used in IMPLAN to gen-erate regional foreign export estimates.
The impact analysis for Louisiana agricul-tural exports had two basic components. Onepart was the direct effect of current export lev-els in 20 agriculturally related industries. Thesecond part was the direct effect in the threetrade and transportation sectors of motorfreight transportation and warehousing,wholesale trade, and water transportation. Thisdirect effect occurred because Louisiana agri-cultural exports were moved to and shippedthrough state ports. The impact analysis sim-ulated the total effect of exports for agricul-tural products, including the three margin in-dustries.
Particular care should be taken in inter-preting model results because I-O models mayoverestimate actual levels of change in eco-nomic activity. Because of the assumptions offixed relative prices, fixed per unit of outputinput requirements, and unlimited supply offactors of production at constant costs, supplyresponse may be overestimated in comparisonto a model where relative prices and input re-quirements are allowed to change. Adjustmentin regional labor markets—and in this casesubstitution of domestic for lost foreign mar-kets-could be especially important in damp-ening the effect of the foreign export shock.Hence, estimates of changes in output, in-come, and jobs from the shock should be re-garded as an upper bound estimate of the ac-tual change under such a scenario.
For comparison purposes, the impact offoreign exports of agricultural products wasestimated in the original hybrid IMPLANmodel. Model results are discussed in terms oftotal industry output (TIO), total income, val-ue added, and employ ment.5 The original es-timates of agricultural exports ($593.950 mil-
5In the hybrid IMPLAN model, employment rep-resents the number of full- and part-time jobs for thesector in question. In the original 1985 version (butnot later versions) of original IMPLAN models, em-ployment is given in terms of full-time equivalent jobs(Alward et al.).
lion) resulted in a total$1.510 billion ($915.993induced effects), or 1.190
401
impact on TIO of
billion indirect and
of estimated TIO in
the Louisiana economy in 1985. The total ef-fect of agricultural exports on total incomewas $585.972 million, or 0.9% of total incomein the Louisiana economy in 1985. The totaleffect on value added was $667.532 million,or 1.1YOof state value added. An estimatedtotal of 25,818 jobs were generated in theLouisiana economy due to the export of Lou-isiana agricultural products.
The same procedure was repeated with thenewly estimated levels of Louisiana foreignagricultural exports. A comparison of impactanalysis with the two estimates of Louisianaagricultural exports implied that the originalexport estimates may have substantially un-derestimated the importance of such markets.The total effect of Louisiana agricultural ex-ports on TIO was $2.197 billion, which was a$686.867 million (45.5%) increase from theoriginal hybrid model TIO estimate (table 3).The export of agricultural products to foreignmarkets was estimated to be responsible for$854.886 million in total income, an increaseof $268.914 million over the original estimate.Louisiana exports were estimated to be re-sponsible for $979.411 million in total valueadded in the Louisiana economy, an increaseof $311.878 million over the original hybridmodel estimate.
The number of jobs generated in the Lou-isiana economy by foreign agricultural exportswas also larger than in the original hybridmodel of the 1985 Louisiana economy. Exportof Louisiana agricultural products generated35,241 jobs in the state economy (table 3), or9,423 (36,5%) more jobs than in the originalhybrid model estimates. The 35,241 jobs rep-resented 1.8% of the total work force of1,984,043 jobs in 1985. This percentage valuewas 0.5 % greater (1.8% versus 1.3%) than thesame estimate calculated with results from theoriginal hybrid model.
The total (direct, indirect, and induced) ef-fect of Louisiana agricultural exports on TIOprovided an indication of the nature of exportimpacts (figure 1). Of the state TIO impact,$970.438 million (44.2%) was characterized
402 Journal of Agricultural and Applied Economics, December 1996
Table 3. Total Effect of Updated Louisiana Agricultural Exports on Selected Industries (mil-lions 1985 $)
Total Output Total Income Value AddedIMPLAN Industry by Sector Code ($ roil.) ($ rnil.) ($ roil.) No. of Jobs
Note: Industries with output impacts under $17.7 million are not reported. However, totals include unreported industries.
as a direct effect, $595.756 million (27.170) asan indirect effect, and $630.625 million(28.7Yo) as an induced effect.G Paper productsreflected the largest TIO impact at $268.960million (table 3).
Of the five industries with the largest in-direct effects in TIO due to agricultural ex-ports, food grains had a $66.517 million in-direct effect and lumber a $47.407 millionindirect effect (figure 1). Both of these indirect
cThe direct shock differed slightly from the sur-vey-based estimates because IMPLAN sector 122,roasted coffee, was not included in the survey data.The original estimate of foreign exports ($5.991 mil-lion) for the sector was assumed to be accurate. Therest of the difference was due to the increases in foodgrain and oilseed product exports.
impacts were explained by the size of the in-dustries and their ties to further processing inLouisiana. For example, the food grains in-dustry required additional processing for itsproducts because rice accounted for most foodgrains. Rice was generally milled before beingshipped overseas. This caused large direct ex-ports in the rice milling industry, which wasreflected as an indirect effect for food grains.The petroleum refining and chemical productssectors also had significant indirect TIO im-pacts due to agricultural exports. Petroleum re-fining provided fuel to agricultural machineryand to the two export transportation sectors ofmotor freight transportation and warehousingand water transportation. The chemical prod-ucts industry is a major producer of fertilizers
Bairak and Hughes: Regional Impacts of Agricultural Exports 403
( InducedEffects ($ Ml.) 1469 Real Estate ($105.623)462 Retail Trade Not Restaumats ($98.741)503Health Sewices ($68.813)491 Eating andDrinkingPlaces ($37.536)456 Electric, Gas, SanitaryServices($36.557) I
Figure 1: Direct, indirect, and induced effects of agricultural exports on Louisiana TIO esti-mated with the Louisiana IMPLAN model with updated levels of agricultural exports
and agricultural chemicals such as pesticidesand herbicides.
Policy makers in many states have lookedto growth in agricultural processing as a wayto enhance general economic activity (Barke-ma, Drabenstott, and Stanley). In Louisiana,growth in agricultural processing has beenseen as a way to counteract losses in employ-ment and income in rural areas due to declinesin production agriculture and a more generaldownturn in economic activity due to markeddecreases in mining. Such a policy is based onthe assumption that increasing the contributionof agriculture to state economic activity wouldsubstantially reduce levels of idle resources,particularly in rural communities with fewother growth options.
Therefore, general categories of the Loui-siana economy were also used to analyze theeffects of agricultural exports. This approachenabled a comparison of the contributions tothe Louisiana economy of unprocessed andprocessed agricultural products. Such a dis-tinction is important, because local processingof locally produced raw agricultural products
increases jobs and income by adding anotherlayer of value to existing activity generated bythe commodity in question (Schluter and Ed-mondson). To accomplish this analysis, the en-tire Louisiana economy was separated into sixcategories: farm products, food processing,wood and paper processing, margin sectors,manufacturing, and services.
The contribution of agricultural exports toLouisiana value added was analyzed for eachcategory (figure 2). The $90.540 million infood processing value added was considerablyless than the $137.646 million in farm prod-ucts. Further, 22.9% of the food products im-pact was concentrated in rice milling. Ricemilling is an important part of the Louisianaeconomy, but it is not a high-value food prod-uct. Other processors with a greater potentialfor generating state economic activity, such aspoultry processors, had a smaller share of thefood processing value-added impact. Hence,assuming that markets are available, replacingexports of unprocessed agricultural productswith increasing sales of Louisiana food pro-cessors to foreign markets would increase the
404 Journal of Agricultural and Applied Economics, December 1996
Notes: Farm Products category includes all unprocessed agricultural commodities and Logging Camps (IMPLANsector 160). Food Processing category includes all food processing products. Wood and Paper Processing categoryincludes paper products and all lumber products except Logging Camps (sector 160). Margin Sectors category includesMotor Freight Transportation and Warehousing (sector 448), Water Transportation (sector 449), and Wholesale Trade(sector 460). Manufacturing category includes allnonagricultural manufacturing, i,e., allmanufacturing except indus-tries belonging to the Agricultural Processing and Wood and Paper Processing categories, and Logging Camps (sector160). Services category includes consumer, business, and government services except the margin sectors andagricul-tural services.
Figure 2. Total value added due to agricultural exports by major category of the Louisianaeconomy estimated with the Louisiana IMPLAN model with updated levels of agriculturalexports
contribution of state agriculture to overall eco-
nomic activity.
The estimated annual average value from
1989 through 1992 of agricultural exports
moving through Louisiana ports was $10.952
billion, of which an estimated $9.981 billion
(91. 1%) originated elsewhere. Agricultural
products originating in other states but moving
through Louisiana to foreign markets may
have considerable influence on the state econ-
omy. A scenario was developed to estimate the
state impact of agricultural exports produced
outside Louisiana. To avoid overestimation,
the port sector of the Louisiana economy—
which was a component of the water trans-
portation sector in the Louisiana hybrid I-O
model—was assumed to be the only industry
directly affected by agricultural exports from
other parts of the United States.
The direct shock for agricultural exports
moving through Louisiana but produced else-
where was derived by first estimating the ton-
nage of agricultural products moving through
Louisiana ports to overseas markets, Using in-
formation provided by the U.S. Army Corpsof Engineers, it was estimated that 70,938,214metric tons of agricultural products wereshipped through Louisiana ports annually. Theestimated value of agricultural exports origi-nating in Louisiana was 2,346,691 metric tons.This value was subtracted from the total ton-nage of agricultural exports shipped throughLouisiana ports, resulting in a net value of68,591,523 metric tons. The 68,591,523 met-ric tons was then multiplied by the Louisianaport service charge ($4.75 per ton in 1985 dol-lars). The Louisiana water transportation sec-tor was estimated to receive a direct impact of$325.810 million because of the shipping ofnon-Louisiana agricultural exports throughLouisiana ports.
The total effect on the state economy ofnon-Louisiana agricultural exports shippedthrough Louisiana ports was $771.948 millionin TIO and 10,096 jobs (table 4). The impactof non-Louisiana exports on state total incomewas estimated to be $264.854 million. The in-dustry with the largest impact was the directly
Bairak and Hughes: Regional Impacts of Agricultural Exports. 405
Table 4. Effect of Agricultural Exports from Other States Shipped Through Louisiana Ports onSelected Louisiana Industries as Estimated with the Hybrid IMPLAN Model (millions 1985 $)
Total Total Valueoutput Income Added No. of
IMPLAN Industry by Sector Code ($ roil.) ($ roil.) ($ roil.) Jobs
Note: Industries with output impacts under $2 million are not reported. However, totals include unreported industries.
affected water transportation sector, with a to-tal TIO impact of $487.873 million and anemployment impact of 4,768 jobs. Other sec-tors with large impacts tended to be serviceindustries, such as real estate.
Results from the demand shocks for agri-cultural exports originating in Louisiana andthe demand shock for agricultural exportsshipped out of Louisiana but produced else-where were summed. The combination provid-ed information on the total impact of agricul-tural exports shipped through Louisiana,whatever their origin. Agricultural exportswere estimated to be responsible for $2.969billion (2. 1Yo) of state TIO. The contributionto state value added was $1.281 billion (1 .7’%oof total state value added). Agricultural ex-ports generated 45,338 jobs, or 2.2?10 of the2,015,084 average number of state jobs from1989–92. While 45,338 jobs may seem insig-nificant when compared to over two millionjobs, the Louisiana economy annually gener-
ated an average of 31,242 net jobs from 1988–92 (U.S. Department of Commerce, Bureau ofEconomic Analysis). Therefore, total Louisi-ana employment tied to agricultural exportswas approximately equal to the number of jobsgenerated by the economy over a year and ahalf.
Summary and Conclusions
A hybrid IMPLAN model of the Louisianaeconomy was used to estimate the state impactof agricultural exports. Results of the modelshould be of interest to policy makers and oth-ers concerned about the ability of foreign mar-kets to generate regional economic activity.
Model results should also be of interest toIMPLAN users. Provided here is a case studyconcerning the accuracy of the proceduresused in calculating foreign exports in IM-PLAN models. Study results suggested thatthese procedures may have underestimated
406 Journal of Agricultural and Applied Economics, December 1996
foreign exports of Louisiana agricultural prod-ucts. When the total impact of original versusrevised estimates of agricultural exports wasevaluated, large differences were observed inthe effect of such markets. Based on thesemodel results, IMPLAN users are urged toverify IMPLAN-based estimates of exportswith outside information for studies directlyconcerned with the impacts of exports.
It was also demonstrated that large differ-ences in the estimates of agricultural exportshad little effect on the holistic accuracy ofmodel results where general changes in finaldemand were evaluated. As a result, IMPLANusers should not be too concerned with theaccuracy of the estimates of foreign exports,if this variable is not of direct interest. In suchcases, researchers may want to verify esti-mates of foreign exports when export esti-mates of important regional commodities arelarge or if knowledge of the regional economysuggests that such estimates may be problem-atic.
Model results of interest to policy makersindicated that the effect of agricultural exportswas felt throughout the Louisiana economy.Based on the assumption that existing levelsof idle resources would be substantially re-duced, the widespread benefits of agriculturalexports to various components of the Louisi-ana economy provide a partial justification forstate efforts aimed at the expansion of agri-cultural export markets.
The expansion of Louisiana agriculturalexports appears to be feasible and desirable.However, the question remains concerning thetypes of agricultural products to emphasize inpromotion efforts. In terms of total contribu-tion to the state economy under current exportlevels, exports of raw farm products generatemore jobs and value added than the exports ofprocessed food products. But the export of ag-ricultural products in processed rather than un-processed form has greater potential for gen-erating additional economic activity. On a perunit basis, export of a given agricultural com-modity in a processed rather than unprocessedform has greater regional impacts because pro-cessing adds another layer of economic activ-ity to the impact of goods and services pro-
duced at the farm gate. Therefore, beyondincreasing the levels of state agricultural ex-ports, policy makers should evaluate a policyof emphasizing the exports of processed ag-ricultural products. A starting point of such apolicy should be an assessment of the com-petitiveness of processed Louisiana agricultur-al products in foreign markets.
References
Alward, G., E. Siverts, O. Olsen, J. Wagner, O.Serif, and S. LindaIl. Micro IMPL4N Users
Manual. Dept. of Agr. and Appl. Econ., Uni-versity of Minnesota, St. Paul, 1989.
Bairak, R. “The Impact of International Export ofAgricultural Products on the Louisiana Econo-my. ” Unpublished master’s thesis, LouisianaState University, August 1995.
Balassa, B., and M. Noland. “Changes in U.S.Trade Specialization: Projections to the Year2000,” J. Policy Modeling 12,3(1990):587-603.
Baldwin, E., T. Smith, D. Larson, and H. Traylor.“Oats Movements in the United States: Inter-regional Flow Patterns and Transportation Re-quirements in 1985,” Bull. No. 794, Agr. Exp.Sta., University of Illinois, Urbana–Champaign,1990.
Barkema, A.D., M. Drabenstott, and J. Stanley.“Processing Food in Farm States: An EconomicDevelopment Strategy for the 1990s. ” FederalReserve Bank of Kansas City, Economic Review
75(July/August 1990):5–23.Belous, R. S., and A.W. Wyckoff. “Trade Has Job
Winners, Too.” Across the Board 24,9( 1987):53-55.
Brucker, S.M., S.E. Hastings, and W.R. Latham.“Regional Input-Output Analysis: A Compari-son of Five ‘Ready-Made’ Model Systems. ”Rev. Regional Stud. 17(1987): 1–16.
Burfisher, M., and M. Missiaen. Developing Coun-
tries’ High- Value Agricultural Trade: Implica-
tions for U.S. Exports. Agr. Info. Bull. No. 615,USDA/ERS, Washington DC, September 1990.
Eckstein Marine Co. Verbal communication regard-ing water transportation and charges. BatonRouge, 1995.
Falgout, C. “Louisiana’s Trade with Mexico UpSharply.” The [Baton Rouge] Morning Advo-
cate, 7 May 1993.Fruin, J., D. Halbach, and L. Hill. “Corn Move-
ments in the United States: Interregional FlowPatterns and Transportation Requirements in
Bairak and Hughes: Regional Impacts of Agricultural Exports 407
Hill, L., M. Patterson, J. Vercimak, S. Fuller, andD. Anderson. “Sorghum Movements in theUnited States: Interregional Flow Patterns andTransportation Requirements in 1985.” Bull.No. 791, Agr. Exp. Sta., University of Illinois,Urbana-Champaign, 1990.
Hughes, D. “Measuring the Effect of LouisianaAgriculture on the State Economy ThroughMultiplier and Impact Analysis. ” Res. Bull. No.849, Agr. Exp. Sta., Louisiana State University,Baton Rouge, 1995.
Hughes, D., D. Holland, and I? Wandschneider.“The Impact of Changes in Military Expendi-tures on the Washington State Economy.” Rev.Regional Stud. 21,2(1991 ):3 11–27.
Jensen, R.C. “The Concept of Accuracy in Region-al Input-Output Models. ” Znternat. RegionalSci. Rev. 5(1980): 139–54.
Jensen, R. C., and G.R. West. “The Effect of Rel-ative Coefficient Size on Input-Output Multipli-ers. ” Environment and Planning A 12(1980):659-70.
Larson, D., T. Smith, and E. Baldwin. “SoybeanMovements in the United States: InterregionalFlow Patterns and Transportation Requirementsin 1985. ” Bull. No. 792, Agr. Exp. Sta., Uni-versity of Illinois, Urbana–Champaign, 1990.
Louisiana Department of Agriculture and Forestry.Louisiana Agricultural Export Directory. Officeof Marketing, Baton Rouge, 1994.
Louisiana Department of Labor. Unpublished Stateof Louisiana Employment Data, 1993. Researchand Statistical Unit, Baton Rouge, 1994.
Martin, R. I?, and D. Holland. “Sources of OutputChange in the U.S. Economy. ” Growth andChange 23,4( 1992):446–68.
Miller, R.E., and RD. Blair. Input-Output Analysis:Foundations and Extensions. Englewood CliffsNJ: Prentice-Hall, Inc., 1985.
National Ports and Waterways Institute. “Lake
Charles Harbor and Terminal District: A ProjectFeasibility Evaluation.” Louisiana State Uni-versity, November 1992.
Reed, M., and L. Hill. “Wheat Movements in theUnited States: Interregional Flow Patterns andTransportation Requirements in 1985.” Bull.No. 795, Agr. Exp. Sta., University of Illinois,Urbana-Champaign, 1990.
SAIA Motor Freight Company. Verbal communi-cation regarding truck transportation and charg-es. Baton Rouge, 1995.
Schh.rter, G., and W. Edmondson. USDA’s Agricul-
tural Trade Multipliers: A Primer. Agr. Info.Bull. No. 697, USDA/ERS, Washington DC,April 1994.
Sharp, J,R. “A Closer Look: Economic Impact ofFree Trade Between Texas and Mexico. ” Spe-cial Financial Rep., Texas Comptroller of PublicAccounts, Austin, November 1991.
Union Pacific Company. Verbal communication re-garding railroad transportation and charges.Baton Rouge, 1995.
“The United States. ” In Trade Policy Review. Ge-
neva: GATT Publication Services, April 1992.U.S. Army Corps of Engineers. Waterborne Com-
merce of the United States. Washington DC,1989.
U.S. Department of Commerce, Bureau of the Cen-sus. “U.S. Exports and Imports of Merchan-dise” (CD-ROM). USDC/BC, Washington DC,1994.
U.S. Department of Commerce, Bureau of Econom-ic Analysis. “Regional Economic InformationSystem” (CD-ROM). USDCE3EA, WashingtonDC, 1994.
Yochum, G., and V. Agarwal. “Economic Impactof a Port on a Regional Economy. ” Growth and
Change 18,3(1987):74–87.Zapata, H., and D. Frank. “Agricultural Statistics
and Prices for Louisiana, 1986–1992. ” AEAInfo. Ser. No. 116, Agr. Center, Louisiana StateUniversity, Baton Rouge, September 1993.
