The Influence of Local Fiscal Structure and Growth Control Choices on 'Big Box' Urban Sprawl in the American West

Report Part Title: Big Box Retail Decentralization in the Western United States
Report Title: The Influence of Local Fiscal Structure and Growth Control Choices on “Big Box” Urban Sprawl in the American West
Report Author(s): Robert W. Wassmer
Lincoln Institute of Land Policy (2002)
Stable URL: http://www.jstor.com/stable/resrep18267.6

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

Lincoln Institute of Land Policy is collaborating with JSTOR to digitize, preserve and extend access to this content.

Big Box Retail Decentralization in the Western United States

An empirical study of the degree of metropolitan retail decentralization must begin with a unit of analysis. For this study it is the 61 metropolitan areas in what the Census Bureau defined in 1990 as the continental western United States, less the seven metropolitan areas in Idaho, Montana, and Wyoming. The analysis is limited to metropolitan areas in the West for a few reasons. The first is that six of the eight states defined as Western enacted statewide ballot box or legislative restrictions on the local use of property taxes between 1977 and 1997. Through Proposition 13 (1978), Measures 5 and 47 (1990 and 1996), and Amendment 1 (1992); California, Oregon, and Colorado voters all used the citizen initiative to limit local property taxation. In Arizona, Nevada, and Utah, state legislatures took similar steps.${}^4$ These restrictions, which Sokolow (2000) classifies as harsher than in any other region in the United States, offers natural experiments by which to test the influence of changes in statewide municipal fiscal structure on metropolitan retail decentralization. Furthermore, most metropolitan areas in the western United States grew up in an era of rising populations, rising real incomes, and declining transportation costs. Metropolitan areas in Idaho, Montana, and Wyoming are not included because the metropolitan development patterns in these three states are very different in comparison to the western states retained in the sample used here.${}^5$

Fifty-four metropolitan areas are used in the analysis. A metropolitan area consists of the relevant component counties in the 1990 Census definition of either a Metropolitan Statistical Area (MSA) or a Primary Metropolitan Statistical Area (PMSA).${}^6$ Since the focus of this paper is retail activity in suburban locations, the suburban area within a metropolitan area is the constituent counties in a MSA or PMSA, less the “central places” included in the 1990 definition of urbanized areas in a metropolitan area.${}^7$ Central places are considered by the Census to be the dominant employment and residential centers in each urbanized area. For instance, the suburban area in the San Diego MSA would be San Diego County less the cities of Coronado, Escondido, and San Diego. Table 1 contains a list of the 54 metropolitan areas in the sample, the areas’ component counties, and the central places that are excluded from these counties to create the definition of a metropolitan area’s non-central places that is used here. This designation of suburbia is

${}^4$ Sokolow (2000) offers a comprehensive survey of property tax limitation in the western United States. See Chapman (1998) for a summary of the local public finance consequences of California’s 1978 passage of Proposition 13.

${}^5$ The largest central cities in each of these excluded states only had 1992 populations of 136,000, 84,000, and 52,000 respectively.

${}^6$ A PMSA consists of integrated counties that are divisible into smaller integrated units that consist of one or more counties. A MSA consists of counties that are not divisible into smaller integrated units.

${}^7$ This approach should be considered conservative in regard to defining excessive decentralization because many would consider the location of retail activity in a place not a central place in 1977 and 1987, but classified as such in 1990, as sprawl.

an attempt to account for the polycentric nature of most U.S. metropolitan areas through the use of existing data sources.

The three dependent variables in the regression analyses that follow are (1) total retail sales, (2) building material sales, and (3) new and used auto sales. Big box retail sales are represented by building material and new/used auto sales due to popular perceptions of what big box is ("Home Depots" and "Automalls") and the availability of comparable data for 1977, 1987, and 1997 for the 54 metropolitan areas under consideration. Table 2 offers a comparison of the ratio of non-central place building material (auto) sales to total metropolitan area building material (auto) sales for the 47 metropolitan areas in the sample for which data is available. The top eight data rows in this table report the averages for each state, where available, using metropolitan area as the unit of observation.

As shown at the top of Table 2, for building material and auto sales where multiple year averages are available, some western states (5 out of 12 cases) experienced an increase in the percentage of these forms of big box sales occurring in non-central locations relative to all metropolitan locations, while the other western states (7 out of 12 cases) experienced a decrease. Just based on these percentages it is hard to distinguish whether these forms of urban retail sprawl have increased or decreased over time. But as indicated above, it is perhaps best to not designate an increase in the percentage of retail activity occurring in non-central places as sprawl if population, income, and demographic shifts warrant such changes. In addition, there are many distinct metropolitan areas that experienced a decrease in the percentage of building material sales occurring in non-central places between 1987 and 1997 (Chico-Paradise, Sacramento, Denver, Bellingham, and Tacoma) and even more for auto sales (Chico-Paradise, Denver, Las Vegas, Medford-Ashland, Salem, and Spokane). Other than that there is a great deal of variation in the degree of non-central place big box retail activity occurring in western metropolitan areas and over time, it is hard to draw any specific conclusions from the information in Table 2. A regression analysis of the determinants of suburban retail activity is necessary to comprehend the causes of the observed variation.

8 Data gathering was complicated by the fact that the U.S. Census Bureau switched from using Standard Industrial Codes (SIC) in 1987 to designate different types of retail activity, to North American Industrial Classification System (NAICS) codes in 1997. For many types of retail this made switch made multi-year comparisons impossible. This was not a problem for new and used auto sales that is represented by SIC 551 (motor vehicle dealer—new and used cars) in 1977 and 1987, and NAICS 4411 (automobile dealers—new and used) in 1997. Building materials is represented by SIC 521,3 (building materials and supply stores) and SIC 525 (hardware stores) in 1977 and 1987, and NAICS 4441 (building materials and supply dealers) in 1997.

9 Table I in Wassmer (forthcoming) offers similar information for aggregate retail sales for all 54 of the western state metropolitan areas in the full sample.