Acentury of Research on Rural Developmen

8/13/2019 Acentury of Research on Rural Developmen

1/33

A CENTURY OFRESEARCH ONRURALDEVELOPMENT

ANDREGIONALISSUES

ELENAG. IRWIN, ANDREWM. ISSERMAN, MAUREENKILKENNY, ANDMARKD. PARTRIDGE

Rural North America has undergone a major economic transformation over the past century due tolabor-saving technological progress, reductions in transport costs, and rising household incomes. The

results are greater rural economic diversity, selected rural population decline, increased ruralurbaninterdependence, emergent exurban areas, and amenity-led rural growth. We summarize key researchinsights and provide a selected review of the economics literature over the past 100 years with afocus on this economic transformation of rural places, its implications for rural communities, and keymodeling innovations and applications. The many important contributions by agricultural economistsare highlighted.

Key words: rural economies, ruralurban interdependence, amenity-led growth, rural developmentpolicy, exurban, land use modeling, input-output models, computable general equilibrium.

JEL codes: C67, C68, O18, R14, R23, R58.

A discussion of rural and regional issues ineconomics requires a broadly cast net tocapture the multiple areas of relevant inquiry:regional economics for its fundamentalfocus on subnational economic units; spatialeconomics for an understanding of the roleof distance; rural and urban economics fortheir respective foci; and land economics forits study of land use and values. Economic

geography encompasses all these concernsand thus provides a useful prologue. Simply,economic geography is about what happenswhere: Why do firms and people tend toconcentrate in a few locations rather thanspread evenly across the countryside? Whatexplains the variety of economic activities andland uses in a region? What are the economic

Elena Irwin is associate professor at the Department of Agricul-tural, Environmental and Development Economics, Ohio State

University: Andrew Isserman is a professor in the Departments ofAgricultural and Consumer Economics and Urban and RegionalPlanning, University of Illinois; Maureen Kilkenny is a profes-sor of Resource Economics at the University of Nevada; andMark D. Partidge is the Swank Chair in Rural-Urban Policy atthe Department of Agricultural, Environmental and Develop-ment Economics, Ohio State University. We are grateful to DavidBarkley, Steve Deller, Tom Johnson, and David McGranahan forproviding feedback on an earlier manuscript version of this article.The article benefited greatly from their excellent suggestions. Inaddition,we thank the following colleagues for their helpful reviewand comments on specific sections of the paper: RalphAlig, NancyBockstael, Alessandra Faggian, Thomas Harris, Stanley Johnson,Davis Lewis, Gerald Nelson, David Newburn,Rose Olfert,AndrewPlantinga,JackStablerand JunjieWu.Lastly,we thankDougWrennand Heather Stephens for their research assistance.

diversification prospects for low-populationrural places that are costly to get to and from?

As Samuelson (1983) noted in his articleThnen at Two Hundred, economic geogra-phy is not new. From the production locationtheory of von Thnen (1826/1966) to its fur-ther development by Weber (1929), Christaller(1933), Lsch (1940), Hoover (1948), andIsard(1956), a fundamental view has emerged of theimportance of trade-offs among scale, trans-port costs, and endowments in explaining theeconomics of establishment location. Writingat midcentury, Hoover summarized these keyinsights in terms of three foundation stones:

[A]n understanding of spatial andregional economic problems can bebuilt on three facts of life: (1) natural-resource advantages, (2) economiesof concentration, and (3) costs oftransport and communication. Inmore technical language, these foun-dation stones can be identified as(1) imperfect factor mobility, (2)imperfect divisibility, and (3) imper-fect mobility of goods and services.(Reprinted in Chapter 1 ofHooverand Giarratani 1984)

From the pioneering insights of Adam Smithon compensating wage differentials to the

Amer. J. Agr. Econ.92(2): 522553; doi: 10.1093/ajae/aaq008
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


2/33

E. G. Irwin et al. A Century of Research on Rural Development and Regional Issues 523

more recent work ofTiebout (1956),and thetheory of spatial equilibrium that emergedfrom the work of von Thnen (1826/1966),Alonso (1964),and other regional and urbaneconomists, the important role of local pub-lic goods and amenities is a further elabora-

tion of these foundation stones in explaininghousehold and establishment location choices.The contributions of economists to research

on rural and regional issues have been orga-nized around the study of these basic spatialeconomic processes and the implications forthe evolution of rural places and for policiesthat seek to promote the vitality of theseplaces (Castle 1991). Agricultural economistsin particular have made numerous contri-butions. In some cases, these contributionswere foundational in terms of methods and

empirical analyses. In other cases, the appliedwork of research and extension economists hasresulted in pushing the boundaries of highlystylized conceptions of space and improvingthe usefulness of the models.

This article provides a broad review of theeconomics literature on rural and regionaleconomic issues over the past 100 years andhighlights the important contributions by agri-cultural economists. While the focus is oncontributions to these topics within agricul-

tural economics, the discussion is necessarilybroader, as many of the important theoriesand research developments have come fromother areas, most notably regional science andurban economics. Our general approach is toidentify so-called lessons from discernible linesof scholarship that summarize key insights inthe evolution of the theoretical and appliedresearch. We provide support for these sum-mary lessons by reviewing the intellectual her-itage of each line of scholarship (which oftentakes us outside the boundaries of agricultural

economics) and then tracing the contributionswithin the agricultural economics literature.

Obvious constraints imposed by a journal-length article necessitate an abbreviated dis-cussion of topics relevant to rural and regionalissues. There is important work, some by agri-cultural economists, that regrettably has notbeen given its proper due here. In particularwe note the omission of explicit discussions ofrural poverty, endogenous growth, and humancapital research. Other topics, while not dis-cussed in the research review, are lifted up asimportant topics in our concluding discussionof future research directions. Lastly, with few

i i i h

Rural Economies

Rural people have made productive use ofthe vast natural resources as inputs for agri-culture, forestry, mining, and other primaryproduction. However, over the past century

rural North American economies have under-gone a major transformation fueled by labor-saving technological progress in agricultureand other natural resource industries (Simon1947), reductions in transport costs, and risingincomes. The results include improved ruralurban income parity, farm population decline,and significant rural economic diversification.

A comparison of the share of farm employ-ment in U.S. metropolitan and nonmetropoli-tan areas shows that in both cases, agriculturesshare was over twofold greater in 1969 than in

2007 (figure 1).The data suggest that even afterthe initial mechanization of farming had run itscourse, productivity growth still yielded fewerand fewer farm workers who were increas-ingly productive. This dramatic decline hasaffected even the most agriculturally intenseplaces. For example, in describing the transfor-mation of the local economy in Saskatchewan,Canada, one of the most agriculturally inten-sive economies in North America,Stabler andOlfert (2009) report that the farm share of

employment declined from 60% in 1931 to 8%in 2008.Despite far fewer farmers, the nonfarm U.S.

rural population has remained remarkably sta-ble over the century (figure 2). At the turnof the twentieth century, the rural farm pop-ulation share was 42%, the urban share was35%, and the nonfarm rural share was 23%.One hundred years later, the rural nonfarmshare remained nearly 20%, but the rural farmshare had declined dramatically to about 1%

Figure 1. Agricultural employment shares inmetropolitan and nonmetropolitan areas

Source: U.S. Bureau of Economic Analy-sis, Regional Economic Information System.
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


3/33

524 April 2010 Amer. J. Agr. Econ.

Figure 2. Percentage of U.S. urban,rural farmand rural nonfarm population, 18902000

Source: U.S. Census Bureau; Census 2000,Summary File 3: generated using AmericanFact Finder; ;U.S. Census Bureau, 1976; USDA-NASS,Trends in U.S. Agriculture. [http://www.usda.gov/agency/nass/pubs/trends/farmpopulation.htm].

(Kilkenny and Johnson 2007). These dramaticshifts in farm employment and population havebeen accompanied by an increasing reliance offarm households on off-farm incomes. Recentdata show that nearly 89% of farm householdsincome was from nonfarm sources (USDA2006), a shift that has led many to observethat agriculture is far more dependent on therural economy than the rural economy is onagriculture (USDA 2006).1

The shifts in rural employment and pop-

ulation reflect larger shifts in the nationaleconomy, including the expansion of employ-ment in services and a decline in the relativesize of the manufacturing sector. In fact, manyrural areas benefited particularly during the1970s and 1980s due to technological changeand the relocation of manufacturing into ruralareas. Other rural areas have benefited fromthe reductions in transportation and commu-nication costs that have increased the accessi-bility of rural places with high-valued natural

1 Rural definitions differ depending on the data source andresearch question at hand. For example, using data from the U.S.CensusBureau,asinFigure2,urban is definedas consistingof built-up areas with more than 2,500 people, andruralas everything else.Alternatively,the Officeof Management category ofnon-core iden-tifies counties without any urban areas of 10,000 people or moreandwithlimited orno commutingto such towns or largercitieselse-where.Micropolitanidentifies counties with urban areas of 10,000to 49,999 and surrounding counties linked to them through strongcommuting flows, leavingas metropolitan counties with urban areasof 50,000 or more and their surrounding areas linked through com-muting. Therefore,nonmetropolitan, which combines micropolitanand non-core counties as in Figure 1, includes a variety of land-scapesranging from small cities andtheircommuting sheds tosome

1,084 counties that have no built-up areas of even 2,500 people. In2000 the nonmetropolitan counties had 20 million residents livingin urban areas and 29 million living in settlements with fewer than2 500 people or in the country See Isserman (2005) for a detailed

amenities and made them desirable residentialdestinations for retirees, tourism-related busi-nesses, and service sector firms. Rural areasare also home to those who are attracted bya rural lifestyle and lower land and housingprices and are willing to commute to cities for

employment.Today, rural areas are exceedingly diversein terms of economic activity and employ-ment. Farming (6%), mining, forestry, andfishing (


4/33


remained remarkably constant (figure2) dueto manufacturing and service sector employ-ment growth. This isprima facieevidence thatthe farm sector is not the sole determinant ofthe rural economy.

While few midtwentieth century econo-

mists disputed the notion that agriculture andnatural resource industries were dominant,many were not inclined to assert that it wouldbe the engine of future growth.Hoover (1948),for example, referred to areas that dependedon agriculture, hunting,fishing, and other natu-ral resourcebased activities as primitive. Hepointed out that because foods share of thebudget declines with rising incomes (Engelslaw), industrialized areas would grow fasteras incomes rise, while agriculture-dependentareas would grow slower.

The emergence of auto transport by themidtwentieth century meant that commutingto urban employment was becoming a majormode by which rural people and communi-ties engaged with nonfarm urban economies.Schultz (1950, 1953) argued that urban agglom-eration effects extended far into the country-side. Rather than land quality, accessibility tourban markets was the key factor in the pros-perity of rural regions. Ruttan (1955) foundthat the median income of farm families was

related to the level of urban-industrial devel-opment in the area and that increased avail-ability of off-farm jobs to farm family memberswas more important than increased labor pro-ductivity in agriculture in raising farm familyincomes. Fox and Kumar (1965) described howurban commuting created functional economicareas that extended 50 miles out from citiesof 25,000 or more people in Iowa. As theystated, [We] are dealing withpeople-orientedrather thanresource-oriented regions (p. 62)[emphasis added]. They pointed out that even

the rural peripheries of functional economicareas were increasingly dependent on anchorurban areas.

Due to their historical focus on the farmsector, Fox (1962) noted that [a] few yearsago only a handful of agricultural economistshad much proficiency in interpreting interac-tions between agriculture and the non-farmeconomy (p. 3). Noting our tendency toequate farms with farmers and farmers withrural communities, Bishop (1967)asked in anaddress as president of the American FarmEconomic Association (AFEA): Surely, theproblems involved are as important to rural

l h i h

devoted more resources to the study of struc-tural changes in rural communities and to pub-lic policies relating to the location of economicactivities and of population? Years later,Pulver and Rogers (1986) studied nonfarmincome in farm-dependent counties and rec-

ommended that rural development policyfocus both on farm incomes, prices, and pro-duction and on vitality of the nonfarm ruraleconomy: The survival of rural Americanfarms and rural communities depends equallyon the expansion of nonfarm income andemployment opportunities in rural areas (pp.11867).

By the 1970s, the notion that rural wasnot agriculturalalone had become the domi-nant view among agricultural economists. AsEdwards (1981)explained:

Most of the rural economy in theUnited States is a nonfarm economy.Rural growth is not the equivalentto agricultural growth. In fact, techni-cal progress in agriculture can releaseredundant agriculture labor into thenearby nonfarm rural economy andhave a depressing rather than a stim-ulating effect. A growing commercialagriculture may havecloser economic

ties with the highly developed urbaneconomy, where it sells products andpurchases inputs,than with the spatialcontiguous rural nonfarm economy.(pp. 223224)

Today, agricultural and regional economistsemphasize the economic diversity of ruralplaces and do not associate long-term placeprosperity with static notions of sector prosper-ity(Barkley 1993;Isserman2001;Johnson 1997,

2007).This conclusion is supported by a varietyof empirical evidence of the wide and variedeconomic contributions of rural areas. Smith(1984)andPorterfield and Pulver (1991)usedsurveys of businesses to demonstrate that ser-vices and other nonmanufacturing businessesexport from rural areas. Therefore, in Smithswords, They can reasonably be includedwith manufacturing in employment genera-tion and economic development programs(p. 153).

Similarly arguing that the narrow focus ongoods-producing industries needs to be reex-amined,Stabler and Howe (1988)found that

h b d b h i
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


5/33


of the four western Canadian provinces dur-ing the 1970s (p. 312). Feser and Isserman(2009)provide the latest evidence of rural eco-nomic diversity by focusing on national valuechains. They identify 45 national value chains,and using a two-way county taxonomy that

considers population distribution and density(urban, mixed urban, mixed rural, and rural)as well as metropolitannonmetropolitan sta-tus (Isserman 2005), they find that rural non-metropolitan counties specialize in 28 of these45 chains. While many of those value chainsare agriculture- or primary sectorbased, suchas feed products, wood processing, and min-ing, rural specialties also include appliances,construction machinery, motor vehicles, andmachine tools. These shifts have affected eventhe most agriculturally intensive rural places.

Stabler and Olfert (2009) report that mod-ern rural Saskatchewan can be described asfollows:

Inputs used in agriculture are nowpurchased from all over the worldrather than in the nearest town.But even as the industrial linkagesbetween the agricultural industryand the rest of the rural econ-omy have eroded, a mutual depen-

dence between rural and urbanSaskatchewan has developed. Thetwo have become integrated as neverbefore through rural dwellers com-mutingtoafewlargecommunitiesforemployment, to shop and to accesspublic services. It is as if all of ruralSaskatchewan has become a geo-

graphically removed neighbourhoodof an urban center. (emphasis added)

These trends of increased accessibility, spa-tial contiguity with expanding urban areas,rising incomes, preferences for amenities, anddiversification of rural economies leads to ournext lesson on rural economies.

Lesson #2: Rural vs. urban is more than asimple dichotomy. There is a stronginterdependence that produces a continuum

from dense urban places to remote rural places.

Distinctions between rural and urban aptly

characterized the economic landscape for mostof civilization. Before the Industrial Revolu-tion, high transport costs and the relatively

Less than 10% of a regions population couldbe sustained in cities, and even the largestcities (outside of Asias more dense rice-basedcivilizations) were concentrations of no morethan 150,000 people (Bairoch 1988;McEvedy1992). Beginning in the nineteenth century, the

Industrial Revolution produced a number oftechnological innovations in production andtransportation that shifted production from thehomes to large factories. Labor-saving tech-nologies in agriculture and scale economiesin manufacturing fueled large-scale rural-to-urban migration and the emergence of manylarge and medium-sized industrial cities. Inno-vations in transit, including electric trolleysin the 1890s and automobiles in the 1910s,decreased the cost of intra-city transportionand led to city expansion into the countryside.

The extent of the city was nonetheless discrete,with the open countryside typically beginningright where the city ended.

These clear distinctions between rural andurban began to blur during the twentiethcentury with the continued rise in agricul-tural productivity, the greatly reduced costsof transporting farm products that put thewhole United States within any one farmsmarket area, and the reduced costs of com-munication and commuting. After World War

II the pattern changed dramatically: Urban-ization slowed, extensive road networks mademany once-remote places accessible, and pop-ulation and production decentralized, leadingto urban sprawl and exurban development.Most counties since the 1980s have containedboth rural and urban areas and are home tomany of theeconomicfunctionsthat previouslyhad occurred only in cities.

Formal modeling of the numbers and spatialextents of the smallest to largest cities in theurban hierarchy was advanced byChristaller

(1933)andLsch (1940).Their Central PlaceTheory (CPT) explains the sizes, distancesbetween, and sectoral compositions of cities,including towns in agricultural areas such asthe Great Plains (Olfert and Stabler 2002). Amajor advantage of CPT is its realism in mea-suring actual business and consumer servicesand their population thresholds (i.e., the popu-lation that is needed to support given servicessuch as a grocery store). The primary criti-cism of CPT is that it is static, requiring adhoc assumptions regarding changes in tech-nology and transportation costs to make itdynamic.

CPT d i
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


6/33


beginning in the 1950s. The declines in com-munication and transportation costs led torising population thresholds, which in turn ledto the consolidation of some private serviceindustries in larger urban centers (Stabler andOlfert 2002). This provided increased pressure

to consolidate government jurisdictions, whichpreviously reflected the needs of a nineteenth-century agrarian economy (Allen 1931; Foxand Kumar 1965).

CPT also motivated regionwide economicdevelopment policies for functional economicareas surrounding larger urban centers of25,000 people or more. Myrdal (1957) intro-duced the concepts ofspread effects, in whichgrowth diffuses to rural hinterlands throughdemand for agricultural products and rawmaterials, leading (if successful) to new cen-

ters of self-sustained economic expansion (p.31). But interdependence with the core couldalso be destructive. Myrdal called backwasheffects those in which urban growth attractspeople and capital away from the hinterlands.For further discussion of the positive and neg-ative dependence of rural areas on their urbancenters, see the reviews by Gaile (1980) andParr (1999a,1999b) and the critical prescrip-tions for regional prosperity byJacobs (1984).

The conceptual strength of the growth-

center model led to support for urban growth-center policy, as well as a debate about ruralparticipation in urban-led growth processes(Day 1968). Growth prospects being betterfor towns of sufficient size justifies targetinginvestment of scarce resources to towns withcritical mass. Efficient government fundingshould target places promising higher returns.Proponents of strategic investment in growthcenters argued that a worst first policy ofproviding funds to the smallest and poorestcommunities wouldbe ineffective because they

lacked the scale economies needed to sus-tain growth (Berry 1970;Hansen 1970). Theysupported targeting intermediate-sized citiesranging from 50,000 to 250,000 (Berry) or even1 million in population (Hansen). On the otherhand, agricultural economists argued that ruralregions shouldhave greater participation in theregional and national growth process and thattargeted growthcenters could be much smaller.Tweeten and Brinkman (1976)suggested thata city size of 10,000 people was large enough.The Appalachian Regional Commission desig-nated even market towns and service centersof 5,000 to 7,000 as centers, pointing out they

l f 250 000 (I

Despite the initial enthusiasm for regionalapproaches, research in CPT, regionalization,and growth centers waned in the 1970s formany reasons. First, in the absence of moderngeographic information systems (GIS) thathave greatly facilitated large data sets and

detailed spatial analysis, descriptive empiri-cal analyses were limited, which preventedresearchers ability to empirically test CPThypotheses. Second, migration into ruralareas led to the so-called rural renaissanceor population turnaround, in which non-metropolitan growth outpaced metropolitangrowth for the first time in the 1970s, especiallyto high-amenity rural areas (Bishop 1967;Beale 1975, 1977). With rural Americasprospects improved, the major institutionalchange needed to coordinate a growth-center

policy that engaged rural communities wasless urgent. Third, though rural towns lackedcapacity to unilaterally plan and conduct eco-nomic development programs, a regionwidescope was opposed by rural interests thatfeared the loss of political self-determination(Whisnant 1994). Fourth, as the AppalachianRegional Commission found, formidablepressure against growth poles comes fromofficials whose areas are not designated and,thus, are ineligible for funds (Isserman and

Rephann 1995).A final reason for the waning interest inCPT was a frustration arising from the lack ofeconomic progress in rural areas near growthcenters (Whisnant 1994). It appeared that therural population decline was mainly the con-sequence of urban growths counterproduc-tivebackwash effects. Proponents of growth-center policy responded that it takes morethan a few years of modest investments torejuvenate entire regions (Richardson 1976).Urban growth was also blamed for the

rural brain-drain,in whichthe highest-skilledwere especially likely to leave. Schultz (1950,1953),Sjaastad (1962), Huffman (1977, 1980),Tweeten and Brinkman (1976),Topel (1986),Barkley (1990),and many others have docu-mented thepulleffect of the relatively higherurban returns to human capital. Others arguedthat infrastructure policy that supported ruralcommuting to growth centers simply encour-aged more out-migration by increasing ruralresidents familiarity with urban locations(Jansma et al. 1981).

Following a relative dearth of research onruralurban linkages for a couple of decades,h i i d d i i h
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


7/33


adjacent rural growth. Tolley (1981) arguedthat a significant part of the increase inrural population is undoubtedly explainedby people living in a rural area and com-muting to a job in an urban area, a situ-ation explained by the seeking of residen-

tial amenities (p. 254). Regionwide develop-ment approaches were needed to deal withthe environmental problems associated withsprawl and the increasing costs of infrastruc-ture and public service provision (Barkley,Henry, and Bao 1996; Drabenstott, Novack,and Weiler 2004). Demand for regionwideapproaches arose once again in remote ruralregions that lacked natural amenities, sinceimproved links to urban economies repre-sented one of their few avenues for growth(Stabler and Olfert 2002). Complementing the

policy demands were innovations in GIS thathelped researchers provide empirical measuresand spatial econometrics.

Spread and backwash represented onetheme in the revived assessment of ruralurban interdependencies and the effects ofurban growth centers (Hughes and Holland1994; Rephann and Isserman 1994; Barkley,Henry, and Bao 1996; Henry, Barkley, andBao 1997). The study by Henry, Barkley, andBao (1997) was particularly innovative in its

use of GIS data and spatial econometrics.Kahn, Orazem, and Otto (2001) investigatedwhether job growth in one county producesa competition with neighboring counties byattracting new migrants from these nearbycounties. Like spread effects, they find that

job growth is more complementary becauseit also leads to population growth in neigh-boring counties through increased commuting.These studies concluded that rural areas offer-ing higher quality of life enjoyed urban spreadeffects such as in-migration and higher rural-

to-urban commuting. Rural areas offering lowamenities experienced urban growth backwasheffects and out-migration. However, they alsoindirectly found that the backwash effects wereweakeningin concert with the declining rateof rural farmer out-migration.

Partridge et al. (2008b) considered hun-dreds of metropolitan areas separated by ruralspace to investigate population growth in non-metropolitan and small metropolitan areas.They created detailed measures of access to thefive nearest higher-ordered tiers in the urbanhierarchy because urban proximity effectsare multitiered. They also investigated the

di i d d ld i h h

in which information technology and lowtrans-portation costs (Glaeser and Kohlhase 2004)are claimed to have eliminated the tyrannyof distance. They concluded, however, thatnot only is distance not dead, but its effectsare becoming stronger, most likely due to

the expanding share of the service sector, ashopping-goods industry, given that humantransport remains costly.

Other studies have focused on how com-muting from nonmetropolitan to metropoli-tan regions affects nonmetropolitan localeconomies. Renkow (2003)found that about60% of the adjustment to local nonmetropoli-tan job growth was accommodated throughchanges in commuting flows and another 30%was through migration. A key implication isthat job growth is not fully accommodated

through increases in local labor force par-ticipation. In weighing costs and benefits ofdevelopment policy, local officials need to rec-ognize that the economic benefits may not goto underemployed locals.

Hedonic studies have examined how ruralwages and housing costs reflect remoteness.Wu and Gopinath (2008) found that remote-ness, defined as nonadjacency to a metropoli-tan area, is the primary factor in the spa-tial variation in wages, accounting for 76%

of the expected differences in average wagesbetween the top and bottom 20% of counties.They found that it was much more importantthan other key factors, including amenities andhuman capital. Partridgeet al. (2009)used finerdelineations of urban tiers to measure the spa-tial decay in both wages and housing costsacross U.S. counties. Relative to a metropoli-tan area with at least 1.5 million population,they found that remote rural wages were up to43% lower and housing values 58% lower, allelse constant.

Coupled with these empirically motivatedstudies of ruralurban interdependence, the-oretical innovations associated with the NewEconomic Geography (NEG) beginning withKrugman (1991) generated great enthusiasmfor regional and urban economics. These mod-els formalized the trade-offs between enjoy-ing internal and agglomeration economies ofscale and avoiding transportation costs withrespect to spatially immobile resources. Prod-uct variety, either among intermediate inputsfor firms or among final goods to consumers,allows for monopolistic competition, average-cost pricing, and higher profits where there

fi hi h i l i h
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


8/33


spatial concentration of production and urbancenters.

The analytical tractability of the canonicalNEG model and its ability to mimic urban-ization reinvigorated broader interest amongeconomists in spatial and regional economics.

But canonical NEG models cannot explainthe simultaneous existence of small and largepopulation centers. Agricultural economistsinterested in rural development made sig-nificant improvements in order to explainit. Kilkenny (1998) presented the first NEGmodel to rationalize the simultaneous exis-tence of small and large concentrations ofpopulation and diverse economic activity inone competitive spatial economy. She did soby formalizing features ignored by the canon-ical model but critical for understanding rural

economies, such as the decision to invest ina locality as determined by latent or poten-tial demand for the localitys product and thenon-inframarginal effectof the entrance of thatnew business on local wages and local marketprices. In addition,Kilkenny (1998) introducedKuhnTucker equations into general equilib-rium to formalize the discrete choice of whereto live and work to maximize real (as opposedto nominal) income. Her model rationalizeskey features of urbanrural interdependence,

including the spatial decay in urbanruralwages and effects of transportation costs on thegeographic concentration or dispersion of eco-nomic activity and population. It also suggeststhe importance of geo-labeled rural products,rural place specificity, and rural quality of lifefor modern rural growth.

Agricultural economists have played animportant role in testing the implications ofdeclining transport costs and lack of agglomer-ation economies in rural areas. Their findingsdemonstrate the competitive disadvantages

faced by rural areasfor example, the impor-tance of positive feedback arising from estab-lishment co-location (Shonkwiler and Harris1996;Barkley and Henry 1998). Other workexamines the importance of other externali-ties that generate regional growth, e.g., thatarise from labor pooling, learning by doing,knowledge spillovers, and innovation (Barkley,Henry, and Kim 1999).

A primary reason that few agriculturaleconomists interested in rural developmenthave adapted NEG models is that the canon-ical NEG model lacks predictive capacity forNorth American settlement patterns. A stan-d d NEG d l ld l h f il d

population (figure 2) even though the farmworkforce has dramatically declined. Like-wise, given the economy of the 1950s andthe ongoing declines in transportation costs,an economist using a standard NEG modelwould have predicted greater concentration in

the Northeast and Great Lakes manufacturingbelts. Instead, there was massive migration tothe hinterlands and other remote regionsof the 1950s, including the Sunbelt, the PacificNorthwest, and the Rocky Mountains. MostNEG models lack predictive capacity becausethey could only simplistically account forinstitutional and amenity effects that had first-order influences on firm and household loca-tion. Such shortcomings led North Ameri-can regional economists to prefer a Tieboutsorting or a Roback (1982) spatial equilib-

rium approach, which better account for jointfirm and household behavior (Deller et al.2001; Glaeser 2007). This leads to the nextlesson.

Lesson #3: Since World War II, migration flowshave been explained increasingly by amenitiesand quality-of-life differences across regions.Consumption of natural amenities has becomeone of the primary determinants of rural

growth.

The attraction of natural amenities and theirinfluence on the economic development ofrural areas is a long-standing trend in theUnited States, one that dates back to the 1920sand that by the 1950s had begun to transformthe spatial pattern of economic development atlocal and regional scales. Two forces drove thisemerging pattern. First, amenities are a nor-mal or superior good. Hence, as householdsacquire sufficient incomes to sustain an ade-quate lifestyle, they increasingly demand a

higherqualityoflifeandmovetolocationswithniceclimates and natural amenities.The secondforce driving this migration was the advent ofnew technologies, most particularly advancesin air conditioning, which made living in theSunbelt more attractive.

While this transformative trend did not havewidespread impacts on rural areas much beforethe 1970s, early land economists noted theincidence of amenity-led rural developmentas early as the 1930s. By this time, much ofthe cutover lands, harvested for their hard-wood timber in the nineteenth century, hadreverted from private to public ownership

d d j h ll bli
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


9/33


economists explored the suitability of theselands for potential recreational development(Wehrwein and Spilman 1933) and demon-strated the economic benefits of establishedrecreational lands in attracting tourists andsupporting new employment that filled the

gap in the tax base created by the declines inthe forest industry (e.g.,Wehrwein and John-son 1943).

Ullman (1954) appears to be the first todescribe amenity-led migration in the UnitedStates and provide a discussion ofhypothe-sized factors, including an increasing numberof footloose households and firms, improvedroads, and universal car ownership. Clawson(1962), then at Resources for the Future, notedthe combined influences of increasing incomesand leisure time and hypothesizedthat these

would imply a greatly increased demand forrecreation uses of rural land. By the 1970s,amenity migration was the key factor forreversing the decades-long rural-to-urban netmigration (Beale 1975;Dillman 1979;Gravesand Clawson 1981). Empirical investigationof population migration provided evidenceof the important role of regional amenities,most notably climate, in regional migration(Cebula and Vedder 1973; Liu 1975; Graves1976) and led to an important extension of

the spatial equilibrium model of urban eco-nomics (Alonso 1964) to regional amenitiesand migration. Graves and coauthors reasonedthat if amenities and rising incomes inducedhouseholds to move to amenity-rich places,then market rents and wages would adjustuntil utility was constant over space, result-ing in higher housing costs and lower wagesin desirable places (Graves and Linneman1979; Graves 1980). This theory wasalso artic-ulated by Rosen (1979) and further devel-oped by Haurin (1980) and Roback (1982),

leading to what is now known as the spa-tial equilibrium Roback model. An impressiveempirical literature followed that sought tovalue amenities by estimating compensatingwage and housing price differentials acrossmetropolitan regions, a literature to whichagricultural and environmental economistsnotably contributed (Bloomquist, Berger, andHoehn 1988).

The spatial equilibrium model provided analternative hypothesis to observed wage dif-ferences that were hypothesized to be theresult of disequilibrium employment differ-entials (Greenwood 1975) and new evidencei h j b l d b h h d

Stein 1964;Muth 1971). In his seminal article,Graves (1980)demonstrated that there is norightorwrongsign associated with incomeand unemployment variables in a model ofpopulation migration, since in a spatial equilib-rium worldthey simply compensate for better

or worse amenities. In addition,he showed thatinclusion of climate in a model of gross migra-tion of selected U.S. metropolitan areas addssubstantialexplanatory power to the model. Itis difficultto overestimate the importance ofhis finding, which has provided the foundationfor subsequent work on amenities and regionalgrowth.

The compensating differentials researchfocused oninterregional migration, and it wasnot until sometime later that work began onthe relationship between natural amenities and

rural population growth (Cromartie and Nord1996;Beale and Johnson 1998; McGranahan1999; Rudzitis 1999). Duffy-Deno (1998)appears tobe the first to introduce amenitiesinto the simultaneous population-employmentregional adjustment model first proposed byCarlino and Mills (1987).A limitation, how-ever,of the Carlino-Mills model is that it beginswith a partial adjustment model for employ-ment and population growth without any struc-tural interpretation of labor demand and labor

supply (Partridge and Rickman 2003). Thus,the resulting coefficients are difficult to inter-pret in anycausal form in terms of, say, jobsversus people debates.

Deller et al. (2001) estimated that thesimultaneous relationship among population,employment, and income changes as a func-tion of a wide array of natural amenities andother control variables. Their paper, publishedin theAJAEand one of the most cited paperson amenities and rural growth, demonstratesthe positive impact on rural county growth

of a wide range of natural amenities, manyof them local and malleable and thereforepotential targets for rural development policy.This and other work (e.g., Carruthers and Vias2005; McGranahan andWojan 2007; McGrana-han 2008)provides empirical evidence of theattraction effects of natural amenities andpotential benefits to rural areas, suggestingthat environmental policy and preservationof natural areas may have important ruraldevelopment implications (McGranahan 2008;Wu and Gopinath 2008). Additional workunderscores the differential effects of ameni-ties on different population cohorts (Fer-

l h d ibb
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


10/33


spurring amenity-led growth (Jensen andDeller 2007).

Recent work has taken into account theexplicitly spatial nature of amenities andrural growth. For example, Partridge et al.(2008b) use geographically weighted regres-

sion (GWR), an empirical method thataccounts for local spatial heterogeneity, toexamine differences in the association betweenamenities and growth. They find substantialdifferences across U.S. counties, e.g., moun-tains are more favorably related to popu-lation growth in the western United States,and lakes have a higher positive associationwith growth in the arid southern Great Plains.Other work has accounted for other aspectsof space, including spatial spillovers (Fergu-son et al. 2007), distance effects (Schmidt

and Courant 2006), and spatial error depen-dence (Rupasingha and Goetz 2004; Kim,Mar-couiller, Deller 2005;Ferguson et al. 2007). Ingeneral, these and other studies have foundamenities to be strong and statistically signif-icant contributors to rural economic growth,although this effect may be specific to U.S.regional development (Ferguson et al. 2007).

Like most economic processes, amenity-ledeconomic development creates trade-offs andmay not translate into other measures of rural

well-being (e.g.,Isserman, Feser, and Warren2009). Amenity-led growth is associated withfaster population and job growth. Yet,to main-tain equal utility across locations in spatialequilibrium,greater site-specific amenities andquality of life imply offsetting lower realper capita income (and vice versa). In equi-librium with mobile populations, one does notobserve both a high quality of life and ahigh real per capita income. This reality hasset off a debate as to whether amenity-ledgrowth worsens income inequality because it

is associated with only, say, low-wage peoplecleaning hotel rooms (Marcoullier, Kim, andDeller 2004). Another complicating featureis that amenity-led growth is usually comple-mented by man-made facilities such as skiresorts, boat marinas, and other recreationalfacilities (Deller et al. 2001;Kim, Marcouiller,Deller 2005). Thus, in addition to a high-quality natural environment, private or publicinvestments are usually necessary to spur thisgrowth.

The discussion thus far has highlightedthe contributions that agricultural economistshave made to the research on rural economies,

i l l i d i d l i i

historical role as natural resource reserves tothe diverse economic places they are today.Economists have also played an importantrole in understanding the implications of thesechanges and, in particular, the implications forrural policies. Next we highlight the contribu-

tions of agricultural economists to the ruraldevelopment policy needs that have emergedwith this rural transformation.

Lesson #4: Sector-based policies are neitherefficient nor effective rural development

policies.

Prior to World War II, most farm householdswere much poorer than urban nonfarm house-holds. Schultz (1953, p. 286) reported thatthe typical industrial workers real earnings

were more than double the typical farmersearnings in 1940. After World War II, ruralpoverty was identified as primarily a prob-lem of low farm incomes (Bryant, Bawden,and Saupe 1981). When off-farm jobs in ruralindustries (those having a major advantagein locating near agricultural or forestry rawmaterial or local agricultural markets) wereproposed as a solution to the low-income prob-lem in agriculture,Robock (1952) concludedthat despite the emotional attachment of ruralindustry for people in agriculture, other fields

of industrial, trade, and service employmentmade more sense. Low income was a mainfactor behind the large migration from farmsto urban areas (Tweeten and Brinkman 1976).Recent evidence from the USDA (2007) showshow different these realities are today. As of2004, median farm household incomes were21% higher than their nonfarm-householdcounterpart, and 95% of farm householdshad greater assets than the median U.S.household.

Evidence that these remarkable gains infarm incomes have lifted the prosperity ofcounties most dependent on farm incomeis, at best, mixed. Data from the EconomicResearch Service (ERS) of the USDA andfrom the Bureau of Economic Analysis (BEA)show that between 1969 and 2007, farming-dependent2 counties lost about 10% of theirpopulation,with about 6% lost over the shorter19902007 period. This compares with 14%

2 The USDA Economic Research Service defines farm-dependentcountiesas havingeither15 percent or more of averageannual labor and proprietors earnings derived from farming dur
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


11/33


growth on average for all nonmetropolitancounties over the 19692007 period and a 7%gain over 19902007. When viewed througha Tiebout lensthat people vote with theirfeet in search of the highest levels of utilitythese farm-dependent local economies have

not fared well. However,when judged based onother measures of rural prosperitynamely,lower rates of poverty, unemployment, schooldropout, and housing problems (Isserman,Feser, and Warren 2009)the farm-dependentrural counties are doing well: 26% did betterthan the nation on all four measures in 2000,and 62% did better on at least three.

The traditional belief that the best way tosupport rural prosperity is to lift farm incomesexplains why U.S. federal rural developmentpolicy falls under the purview of the USDA.

Farm support policies have not typically soughtto promote local job growth, however, and infact may detract from rural growth given theneed for productive farms to shed labor toremain competitive. On the other hand, thereis no agreement on the effectiveness of alter-native rural development policies or even themore basic welfare question of what the objec-tive of rural development policy should be.Nonetheless, a consensus among rural devel-opment economists has emerged on policies

tailoredto place versus those based on sectorsor on people (Fluharty 2002; RUPRI 2002).Economists have long supported people-

based policies of facilitating education, work-force training, and workforce supports such aschildcare and transportation (when needed)and of providing relocation assistance fromstagnant locations to more economicallyvibrant ones (e.g., Glaeser 1997; Tweetenand Brinkman 1976). Many economists haveadvocated an even stronger notion of spa-tially neutral people-based policies, as most

famously stated by Harvard economist EdwardGlaeser when he argued that it was bet-ter to provide each resident of the city ofNew Orleans $200,000 than to have fed-eral funding to rebuild it after HurricaneKatrina (Pettus 2006). Fully spatially neutralpeople-based solutions are much less effectivewhen constraints limit labor-force movementstoward economically stronger locales and leadto a spatial mismatch between the available

jobs and the people who need them.Partridgeand Rickman (2008) provide evidence thatmobility is limited in more remote rural areas,questioning the effectiveness of fully spatially

l li i H d M i i h (1989)

neutral tax or incentive policy can have spa-tially varying effects on economic activity andbusiness location because of spatial hetero-geneity in endowments, spatial heterogeneityin the types of industries locating across theseheterogeneous places, and the multidimen-

sional differences in distances to the variousdifferent input source locations and markets.Likewise, spatially neutral people-based poli-cies can have spatially heterogeneous effects.People who have chosen to live in places withdifferent endowments and employment oppor-tunities are likely to have different preferencesabout market goods and policy alternatives.Thus, the same people-based policy can leadto many different outcomes in different places(Kilkenny and Huffman 2003; Blank 2005).

The implication of fixed costs that there is

a minimum efficient scale (or critical mass/population size) below which per unit costsrise also suggests that there is no such thingas one size fits all.Tweeten and Brinkman(1976), Shonkwiler and Harris (1996), andWensley and Stabler (2002), for example, pro-vided empirical evidence of the different mar-ket sizes or critical mass, in terms of both pop-ulation and business interdependencies, belowwhich different types of establishments are notsustainable. The implication is that attempt-

ing to provide every person everywhere withthe same level of services may be prohibitivelyexpensive in low-density rural places.

Policies based on place are those for whichthe location or spatial category of the bene-ficiary is a key criterion for eligibility. Manyeconomists are skeptical of policies that arenot spatially neutral, and instead argue thatplace-based policies should be incorporated asa last resort (seeWorld Bank 2009).Kilkennyand Kraybill (2003) discuss six pitfalls andshortcomings of place-based policies:They may

(1) generate nothing but rents for the (poten-tially absentee) owners of property in targetedplaces; (2) attract or retain (trap) poor peo-ple in poor areas; (3) distort business as wellas human migration decisions; (4) enable thepostponement of necessary adjustments; (5)create dependencies; and (6) be abused byplace-based politicians.

Consider places with chronic unemploymentproblems. Research shows that in the UnitedStates, labor typically adjusts to local unempl-oyment shocks through migration (Topel 1986;Bartik 1993). All else equal, people move fromplaces with poor employment opportunities to

l i h b Th l h i
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


12/33


negative spatial externalities, this better out-come costs society nothing. However, policiesthat mobilize people out of low-income ruralareas unintentionally add to the size disadvan-tage implied by scale economies by pushingrural communities further below their mini-

mum efficient scale (Kilkenny and Johnson2007).Gabe and Kraybill (2002)provide evidence

that a place-based rural development policythat waivestaxes or provides subsidiesto enticebusinesses to employ more people in ruraltowns may do nothing more than increase thenetprofitsof absentee owners.There is no guar-antee that the businesses will redirect the newrevenues to hire more of the local unemployed.Indeed, place-based policies may simply intro-duce perverse incentives. If larger tax breaks

or subsidies are targeted to places with higherunemployment, the less that businesses do toreduce local unemployment, the larger aretheir gains from policy transfers. Furthermore,businesses that do not depend on local pur-chasing power for their sales, the oft-chasedexport industries, have the weakest incentivesto reduce local unemployment or to bid localwages up.

In contrast, Kilkenny and Johnson (2007)argue for policies that are tailored to fit

local conditions, specifically place-tailored asopposed to place-based. Place heterogene-ity justifies place-tailored policies. For example,where the school-age population is less dis-persed, less need be spent on either transportor infrastructure in order to provide all citi-zens with the same outcomes. Locally sensitivepolicies should also be more effective or bettermatched to local preferences than one-size-fits-all. However, as Kilkenny and Johnson noted,place-tailored policies are welfare improvingonlyifthecostoflocaltailoringisnotexcessive.

Rural Development Modeling

Rural development is distinct from ruralgrowth. Growth usually means more ofeverything: more population, more resourceemployment,and more aggregateincome with-out a significant change in industry mix, tech-nology, productivity, or income per capita.One definition of development, on the otherhand,consists primarily in employing existingresourcesinadifferentway,indoingnewthings

i h h ( h 1 11 1 61 6 )

varied nonfarm rural industry and land uses,new rural occupations, and higher income percapita. The relevance of multiple sectors andmultiple factors, the interplay of demand andsupply, and the need to understand householdand producer responses to market signals and

policies are obvious in this setting, which leadsus to the next lesson.

Lesson #5: Rural development is a generalequilibrium problem that requires generalequilibrium tools. These and other quantitativetools provide a necessary foundation forcommunity economic analysis.

Input-output models Wassily Leontief wonthe 1973 Nobel Prize in Economics for devel-oping the input-output method. Said the Royal

Swedish Academy of Sciences, This impor-tant innovation has given to economic sciencesan empirically useful method to highlight thegeneral interdependence in the production sys-tem of a society. In particular, the methodprovides tools for a systematic analysis of thecomplicated inter-industry transactions in aneconomy. Even by that time, few anticipatedthat the input-output method would become sowidely used within agriculture economics.

The first appearance of an input-output

model in the refereed economics literatureappears to be Leontiefs (1936) analysis oftechnical progress in the United States. Leon-tief documented the changes in the use ofeach industrys output as intermediate goodsby other industries. Soon after this paper,the usefulness of input-output modeling as ageneral equilibrium planning tool was demon-strated. Knowledge of the countrys input-output relationships enabled the federal gov-ernment to coordinate production among allsectors to deliver the ordnance needed for

the war effort in World War II simultaneously(Leontief 1951). Despite the obvious differ-ences between planning a centrally adminis-tered, centrally financed, centrally coordinatedexpansion of production to meet new finaldemands for military goods and predictingthe effects of decentralized changes in eco-nomic opportunities, Leontief subsequentlytouted input-output models as a forecastingtool: The immediate purpose of interregionalinput-output analysis is the determination ofthe differential impact of any given change inthe final bill of demand (Leontief 1952, p. 9).

Agricultural economists have long madei ib i i d
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


13/33


(1951), Heady and Schnittker (1957) presentedthe simple arithmetic of the method, results foragriculture differentiated by region using 1949data,and an assessment of the methods limita-tions.Ram (1958)created a five-sector systemfor eight North Carolina counties with labor

and land equations to demonstrate how input-output could be used in a small homogeneousagricultural area. Davis and Goldberg (1957)traced the forward and backward linkages offarming, food processing, and fiber processingand coined a novel word to describe the linkedindustries: agribusiness. USDA continues touse this basic approach when identifying, moni-toring, and analyzing the food and fiber system(e.g.,Henry and Schluter 1985;Schluter, Lee,and Edmonson 1986;Lee and Schluter 1993;Edmonson 2004, 2008).

Maki, Barnard, and Suttor (1964) wereperhaps the first to embed an input-outputmodel within an econometric model. Littleand Doeksen (1968) devised a procedure tomeasure leakages from the local economy.Heady and Sonka (1974) used input-outputand math programming to address implic-itly the question Is agricultural policy goodrural development policy?They developed animpressive eleven-region interregional mathprogramming model to estimate the agricul-

tural sectors input demands, employment,andfarm household income under four farm sizescenarios in eleven regions. They then appliedeleven tailor-made regional input-output mod-els to simulate the local economy-wide effectsof the differences in farm input demand andfarm household demand for consumer goodsand services that they estimated using themath programming models. They found thathaving smaller farms, rather than the largerfarms that were encouraged by agriculturalpolicies, would support greater income genera-

tion in rural communities. Apparently agricul-tural economists have not yet econometricallytested this intriguing simulation result.

Agricultural economists have also usedinput-output to predict the changes in sup-plies of various occupational types of laborneeded to accommodate a new local enterprise(e.g., Doeksen 1972;Kraybill and Dorfmann1992;Lego, Gebremedhin, and Cushing 2000)and, conversely, to predict which occupationaltypes may become unemployed due to contrac-tions (e.g.,Waters, Holland,and Weber 1994) oralternative natural resource uses (e.g., Seunget al. 2000).

h ibili di h i

modeling attracted the giants of the field,Leontief (1970)and Isard (1969), but the inno-vative work of agricultural economists Hiteand Laurent (1972; Laurent and Hite 1972)stands out because their approach was prac-tical and operational. They added a matrix of

resource use or emissions per dollar of grossindustry output to the input-output model,thereby enabling calculations of changes inpollution as a result of changes in output,albeit with the typical input-output modelingassumption that pollution is proportional tooutput.

The initial period of modeling innovationwas followed by much research focused on costcutting. Building survey-based input-outputaccounts for an area can be very expensive, soresearchers sought ways to adapt the national

input-output accounts and create what becameknown as nonsurvey or partial survey meth-ods (seeRound 1983for a critical review). Thegoal,one influential model builder explained, isholistic accuracy, getting the general pictureright (Jensen 1980). Agricultural economistswere particularly creative in building modelsfor small areas. Robison and Miller (1991) usedideas from CPT in modeling a timber economyinIdahowithsixcommunitiesandabout20,000people. Robison (1997) divided two counties

with fewer than 12,000 residents into sevencommunity-centered regions.By 1986 the technology for quick construc-

tion of an input-output model for any countyor groups of counties had proceeded so far thatBrucker, Hastings, and Latham (1990) wereable to conduct an experiment. Five modelersagreed to run the same seven simulations usingtheir systems and compare the results. Two par-ticipants were federal government agencies:theBEA,inventorofRIMSII(RegionalInput-Output Modeling System), and the USDA

Forest Service, inventor of IMPLAN (ImpactAnalysis for Planning).

Since that time, IMPLAN has swept thetable. USDA officially adopted it to esti-mate the potential job creation effects of theAmerican Recovery and Reinvestment Actof 2009 (Kort 2009). According toScott andJohnson (1998,p. 51), IMPLAN became pop-ular among applied economists at land grantuniversities after a small investment by theUSDA Extension Service and the four USDARegional Rural Development Centers led totraining programs starting in 1986. IMPLANcounty data are key components in Miller andR bi b i d l d
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


14/33


AJAE articles, from bioenergy (de la TorreUgarte, English, and Jensen 2007) to tourism(Cline and Seidl 2009). Scott and Johnson(1998) claim, The availability of IMPLANhas influenced decisions involving millions ofdollars of private and public funds (p. 51).

It all began when the National ForestManagement Act of 1976 required the For-est Service to prepare five-year manage-ment plans based on analyses of alternativeland management options and their predictedsocioeconomic effects on local communities.The Forest Service made a major investmentto develop the IMPLAN modeling software(USDA-NRCS 2009). In 1988 MinnesotasDepartment ofAgricultural Economics startedproviding software and support to nonForestService users, and in 1993 these activities were

spun off as a private company. The princi-pals continue to innovate, recently releasinga multiregional system for analyzing effectsof activities in one county on other counties(Lindall, Olson, and Alward 2006).

The use of the input-output method has nowbecome routine, which is a danger (Henry andJohnson 1993). One can just push the buttons,but that usually is a mistake. The IMPLANsoftware allows users to override the defaultassumptions, which is important because agri-

cultural economists often are interested in newindustries that are not among the 509 sec-tors of the model, like dry-mill ethanol, andbecause technologies vary greatly across thecountry. Knowledge of the regional economyand local industry is very important to beable to adjust the IMPLAN model and specifythe scenario reasonably well, asLazarus et al.(2002)demonstrate when analyzing the Min-nesota pork industry. Low and Isserman (2009)show the many complex decisions that mustbe made outside the model without the help

of the model when estimating ethanols localeconomic impacts.

Another temptation lurks.Leones, Schluter,and Goldman (1994, p. 1123) warn, Poli-tics often motivates undertaking studies onthe size or importance of agriculture in astate or region. State and federal depart-ments of agriculture and state-supported col-leges of agriculture use this information tolobby either on budgetary or policy issues bytouting the importance of agriculture to theeconomy. . .. Agricultural economists have aresponsibility to make sure that such stud-ies follow acceptable procedures in economic

i il l (1 1) i

The value-added simulated using an input-output model measures a net benefit onlyto the extent that unemployed resources orlabor are brought into production or are usedmore productively.Watson et al. (2007) cau-tion against describing input-output results as

economic benefits and call for a consistentand explicit semantic, distinguishing amongeconomic activity, economic contribution, andeconomic impact.

Community Economic Analysis

In 1966 AFEA had no specialization in ruraldevelopment, but by 1982, 3.5% of the mem-bership in the Agricultural and Applied Eco-nomics Association (AAEA), 112 individuals,claimed community resource economics as

their specialization (Swanson 1984). The bookMicropolitan Development, by Tweeten andBrinkman (1976), claimed apparently accu-rately to offer a comprehensive look at ruraldevelopment totally lacking in any previouspublication (p. vii). Its approach, grounded intheory and methods, is activist in intent: Ouroverriding concern is to improve thewell-beingof micropolitan people, wherever they even-tually reside (p. 6). Community EconomicsbyShaffer (1989)discusses more quantitative

methods, including the use of concepts fromCPT-like demand thresholds and market areas,as well as input-output, but mastery of themethods is not an end in itself. The methods areto be used with people in a community, becausecommunity economic analysis is a compre-hensive effortto changetheeconomic situationwithin the community (p. 8).

A prominent step on this movement to bringinput-output and related methods into the ser-vice of rural communities has been the Com-munity Policy Analysis Network (see Scott

andJohnson1998). Community PolicyAnalysisModeling (Johnson, Otto, and Deller 2006) isthe product of a multistate effort to go beyondthe capability of the input-output model toconsider labor markets and fiscal impactsusing cross-sectional and panel data economet-ric models. The fiscal impact modeling tech-niques developed by agricultural economists(e.g.,Johnson and Keeling 1985;Halstead andJohnson 1986; Swallow and Johnson 1987;Johnson 1988) provide state and federal gov-ernments with estimates of the magnitudes anddistributions of the benefits and costs, includingthe deadweight losses, of proposed changes infi l d i d i l i d h
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


15/33


Waters, Holland, and Weber 1997;Seung andKraybill 2001). Input-output is also used topredict the employment and income payoffsof particular investments, projects, and devel-opment programs (e.g.,Berck, Robinson, andGoldman 1990; Berck and Hoffmann 2002;

Deller and Shields 1998).Agricultural economists also developed amore realistic treatment of the rural labor forceand how it changes, on the basis of their econo-metric analyses of rural labor markets (e.g.,Renkow and Hoover 2000; Mills and Hazarika2001;Khan, Orazem, and Otto 2001;Renkow2003; Rupasingha and Goetz 2004; Barkley,Henry, and Li 2004). Recent econometric anal-ysis of exogenousexport sector employment,rural income, and population growth over timehas provided a summary lesson: Rural labor

supplies are not perfectly elastic (Kilkenny andPartridge, 2009). Rural labor supply appears tobe inelastic with respect to increases in labordemand and elastic with respect todecreases.

Targeting regional economic development(Goetz, Deller, and Harris 2009), with 30authors, mostly agricultural economists, isthe most recent collaborative effort andwas preceded by papers that developed theinput-outputbased methods to inform and

evaluate local industrial targeting initiatives(e.g.,Leatherman, Howard, and Kastens 2002;Deller and Shields 1998). Another Harvardresearch project, that of Porter (1998), hadlaunched a wildly popular quest to fosterregional competitiveness, including in ruralregions (Porter et al. 2004). Vintage methodsand concepts are in play, such as location quo-tients, shift-share analysis, input-output,importsubstitution, location analysis, industrial com-plexes, and others discussed authoritativelyhalf a century ago in Methods of Regional

Analysis(Isard et al. 1960). The key analyticalstep, identifying existing or potential spatialclusters of interrelated industries, commonlyinvolves the location quotient, introduced inthe United States by theNational ResourcesPlanning Board (1943) and debuting in therefereed literature in Review of Economicsand StatisticsbyHildebrand and Mace (1950).Although high-speed calculations with largedatabases and inter-industry linkages revealedby national input-output tables are involved,so is concern with the community processwithin which the analysis is used. By combin-ing the process of targeting analysis with the

f h l i G D ll d

TRED exercise can be as much an educationaltool as it is the foundation of a technicalreport (p. 2).

Computable General Equilibrium Models

The abstraction of input-output models fromprice endogeneity and the inability to accountfor factor supply inelasticity are inherent lim-itations of the I-O method. For this reason,when the technology to compute general equi-librium was discovered in the early 1970s,agricultural economists pursued it. The devel-opment of computable general equilibrium(CGE) modeling techniques by agriculturaleconomists has been a major contribution tothe entire economics profession. Computablegeneral equilibrium models are the only ones

that endogenously determine primary factorsupplies as well as all prices and incomes in aneconomy.

Many founding fathers of CGE wereinternational development economists con-cerned about agriculture and developmentpolicy (Adelman and Robinson 1987;Dervis,de Melo, and Robinson 1982; de Janvry andSadoulet 1987). Other key progenitors wereeconomists interested in tax, trade, or indus-trial policy (Shoven and Whalley 1973;Hertel

and Tsigas 1988; Mutti, Morgan, and Partridge1989), many of whom are also agriculturaleconomists. The study by Shoven and Whal-

ley (1973), the first CGE article to appearin a refereed journal, was followed by aninitial trickle. Models were laboriously writtenin FORTRAN code. The number of refereedCGE journal articles took off after Herb Scarfdeveloped a practical algorithm for solvingfixed-point problems (see Kehoe, Srinivasan,andWhalley2005),theadventofmicrocomput-ers, and release of the General Algebraic Mod-

eling System (GAMS) software by Brooke,Kendrick, and Meeraus (1988). In the decadessince the early 1970s, almost one thousandarticles using applied or CGE methods havebeen published in refereed social science jour-nals.

CGE models are important for rural pol-icy analysis, and rural development economistshave made important contributions to CGEmodeling of regional economic issues. InAdelman and Robinsons (1987)early modelof Korea, for example, the rural economyconsisted of the agricultural sector, and ruralhouseholds were those employed in agricul-

h li i l
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


16/33


other markets and the many other factorsthat account for why rural is more than justagriculture.

The first multisector two-region (rural andurban) Walrasian CGE model to explic-itly model interdependent rural and urban

economies and the associated market segmen-tation was published in AJAE by Kilkenny(1993). The model endogenizes commuterincome, absentee landownership and transre-gional capital income flows, ruralurban goodstrade, rural government and intergovernmen-tal transfers, and the offsetting rural and urbancurrent/capital account imbalances. With thatmodel, Kilkenny showed that coupled farmsubsidies were not as effective as decoupled(nonfarm) income transfers at promoting ruralprosperity.

CGE methods have also been developed byagricultural economists to predict the effectsof, for example, agricultural liberalization onintersectoral labor reallocations (Kilkenny andRobinson 1990), macroeconomic imbalanceson a state economy (Kraybill, Orden, andJohn son 1992), cuts in defense spending ona state economy (Hoffmann, Robinson, andSubramanian 1996), and property tax limita-tions (Waters, Holland, and Weber 1997), aswell as to describe the role of agriculture in

the rural economy (Midmore et al. 1997), localeconomic trade-offs between recreational andagricultural water uses (Seung, Harris, andEnglin 2000), and the local economies of smalltowns (Schwarm and Cutler 2003). Given theirexpertise with CGE, agricultural economistswere also able to adapt, use, and properlyevaluate the tools of NEG for modeling ruraldevelopment (Hite 1997; Kilkenny 1998).

Rural Land Use Modeling

Land andits location arelong-standing areas ofinquiry within economics. Fundamental to landeconomics are the works by Ricardo (1817),whose concept of rent as equal to the differ-ences in product caused by the difference insoil fertility was fundamental to the economictheory of value, andvon Thnen (1826/1966),whose work on the role of transport costs indetermining the value of location in space pro-vided the foundation for location and landeconomics (Isard 1956).While traditionally thepurview of resource economics, the economicsof land use has emerged as an important rural

d i l i i h h i b

of exurbia as a recognized category alongthe ruralurban continuum. Current land useeconomics research is found in both environ-mentalandresourceeconomicsjournals as wellas regional and urban economics journals.

The beginning of land economics in the

United States is marked by the establishmentof the Division of Land Economics within theUSDA in 1919, the founding of the Institute ofLand and Public Utility Economics by RichardT. Ely at the University of Wisconsin, and thelaunch of the journal Land Economics (knownthen as the Journal of Land and Public Util-ity Economics) in 1925. The original scopeof land economics was broad and includedthe economics and management of all naturalresources under the general heading of land,as well as land policies and reforms, private

and public land ownership, land values, tax-ation, and other institutional concerns. Landutilization was and continues to be an impor-tant topic as well. Indeed,Salter (1942)calledthe study of problems associated with majorchanges in the pattern of economic use of landa primary organizing topic in rural land useeconomics.

Much of the early work of economists onland utilization was descriptive, e.g., analy-ses of regional land use patterns that linked,

for example, farm management considerationswith land suitability and other regional fac-tors that influenced farm-level decision mak-ing (Salter 1948,p. 27). Following World WarII, agricultural economists modeling land asa factor of production transformed land eco-nomics research by applying the new methodof linear programming (LP) to quantify opti-mal land use decisions. The focus was nat-urally on farmer decision making, and thusthis work was concerned primarily with allo-cations of land to different agricultural com-

modities. At the forefront of this new researchwere two agricultural economists, Earl Headyand Glen Johnson. According toCastle et al.(1981), their work is one of the most signif-icant developments in agricultural economics(p. 414). For example, Heady and Egbert(1964) created a land-constrained cost mini-mization programming model of agriculturalcommodity production to allocate farm landoptimally in 122 producing regions of theUnited States. The regions were delineated bysoil quality and other physical land features,and thus their approach defined the spatialdistribution of heterogeneous landscape fea-

f l d i bl f i l H
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


17/33


underlying normative assumptions and aggre-gation bias problems with the large regionaland national LP models (Taylor and Howitt1993) were problematic. From a rural landuse planning perspective, these models ulti-mately failed to provide reasonableprojections

of future land use allocations, and thus policy-oriented researchers were forced to developalternative econometric models of land usechange (Alig, personal communication).

Following the marked increase in housingdemand and population suburbanization afterWorld War II, concern over urban sprawland its impacts on both cities and rural regionsbegan to attract the attention of an increasingnumber of researchers (Bishop 1967;Clawson1962; Harvey and Clark 1965; Kaiser 1968).Theplanning needs for future predictions of rural

land use transitions prompted new reduced-form approaches to the empirical modelingof land use and land use change. Agricul-tural and resource economists were key inno-vators. We summarize this contribution inthe first of two lessons on rural land usemodeling.

Lesson #6: Reduced-form empirical models ofland use and land use change have provenuseful in examining land use outcomes and in

identifying the effects of policies on rural landconversion.

A large literature on econometric modelingof land use change has emerged since itsbeginnings in the early 1980s. In addition to theplanning needs of federal agencies that soughtlarge-scale regional projections of future landneeds, these models were spurred by growingconcerns over the urbanization of rural landand analyses by researchers at USDA of grossland use transitions that revealed much more

dynamicrurallanduseprocessesthanwhatwasrevealed by net statistics.

The key turning point came in the devel-opment of reduced-form econometric modelsthat provided a full accounting of all landacross alternative land use categories. Twoagricultural and natural resource economists(Alig and White) working in tandem withan urban/environmental economist (Healy)were the first to develop land use and landuse transition models that accounted for landacross multiple land uses (e.g., cropland, pas-ture, forest, urban) using state-level data onland use and socioeconomic variables fromh S C ( hi d l i 1980

highly aggregated, sparse data beginnings, agri-cultural economists have forged the way inmaking continual improvements to economet-ric land use modeling through the use of betterdata, theory, and methods. Following on theinnovations of the early and mid-1980s, agricul-

tural economists continued to improve uponthe basic land use share model, e.g., to accountfor heterogeneous land quality (Stavins andJaffe 1990; Wu and Segerson 1995), uncer-tainty (Parks 1995; Schatzki 2003), forestrydynamics (Plantinga 1996) and land use transi-tions (Lubowski, Plantinga, and Stavins 2008).These models have been usefully appliedto study the effects of a variety of fed-eral policies on regional land use alloca-tions, including federal flood control projects(Stavins and Jaffe 1990), wetlands restoration

programs (Parks and Kramer 1995), climatechange strategies (Lubowski, Plantinga, andStavins 2006; Plantinga, Mauldin, and Miller1999), federal farm policies (Lichtenberg 1989;Plantinga 1996) and deforestation in Brazil(Pfaff 1999).

The GIS data revolution of the 1990sspawned new sources of spatially disaggre-gated data and further transformed empiricalland use modeling. It became possible to matchthe spatial scale of the data with the scale of the

economic behavioral process. This integrationof microspatial data (e.g., at the scale of landparcels or subdivisions) with discrete choiceeconometric models provided new opportu-nities for examining the associations of spa-tially heterogeneous features with fine-scaleland use patterns and improved economet-ric parameter estimation. Chomitz and Gray(1996) and Nelson and Hellerstein (1997) wereamong the first to estimate spatially disaggre-gated land use models using remotely senseddata from Belize and Mexico respectively. In

both cases, a theoretical model of deforesta-tion was used to motivate a reduced-formempirical model in which access to roads is akey variable. A seminal contribution is foundin Bockstael (1996), the text of her Waughlecture at the American Agricultural Eco-nomics Association in 1995. This paper laysout a basic framework for a discrete choicemodel in which land is fully allocated to theuse that maximizes its net present discountedvalue of returns. It then applies the model tourban land development in a central Marylandregion.

The spatial disaggregation that is the key-f h d l h b
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


18/33


and facilitated the use of newer methods thatenable the empirical identification of policyeffects. Endogeneity and selection bias prob-lems frequently arise in the context of eval-uating land use regulations. This is due tothe endogenous nature of most local land

use policies, e.g., growth controls are mostoften imposed in areas that alr

Documents

Acentury of Research on Rural Developmen