An Empirical Enquiry into the Impact of Urban Planning Policy on Urban Growth

Abstract

Recent research has stressed the need to evaluate the economic implications of urban planning policy. In this article, we present empirical evidence on the impact of this policy on the population growth of towns and cities. A simple theoretical model serves to highlight the mechanisms whereby this policy may affect urban growth. The model yields a reduced-form equation which we estimate for the towns and cities of Andalusia (Spain). The empirical model strongly supports the claim that urban planning policy considerably affects urban growth and thus, the distribution of population across space. Our results suggest that urban planning policy is contributing to the reduction of diseconomies of agglomeration in larger cities. On the negative side, we find that urban plans are subject to obsolescence, which slows down growth.

Acknowledgements

This study forms part of a wider research project entitled The Impact of Urban Legislation on Territorial Planning. This project was commissioned by the Economic and Social Council of Andalusia and carried out by a group of researchers from the Universidad de Granada. The authors would like to especially thank Manuel Martin Rodriguez for his guidance in this project and his comments on earlier drafts. Also, thanks are due for their comments to Gianni De Fraja, Juan Luis Eugenio Martin, Francisco Jose Gonzalez, Javier Rodero, Juan Toro and two anonymous referees. All remaining errors are ours.

Notes

Note that this empirical evidence refers only to the US.

Perfect competition requires, among other assumptions, perfect information on the part of all agents in the market. It is thus, a weak assumption, especially in the analysis of business location choices or real state markets, where the existence of imperfect and asymmetric information is pervasive. Asymmetries of information and their impact on market outcomes have been studied by Akerlof (1970) for the case of commodity markets and by Stiglitz and Weiss (1981) for financial markets. Asymmetries of information are relevant for our research question; it would be clearly interesting to consider them in future research.

A brief overview of the theoretical and empirical literature on externalities and urban growth can be found in Glaeser et al. (1992).

Among these, we include the path-breaking article by Tiebout (1956) and the huge literature it has generated (see Ross & Yinger, 1999, for a review), and the more recent ones by Black and Henderson (1999) Simon and Nardinelli (2002) and Glaeser and Shapiro (2003), which underline the importance of human capital, along with the work of Wheaton (1998), Bjorvatn (2000) and Chatterjee and Carlino (2001), which analyse the influence of infrastructure on urban growth processes.

Sakashita (1995) extends the model of Engle et al. (1992), by considering the residents' utility as an endogenous variable. His conclusions justify the establishing of growth controls only under certain circumstances and under certain compensatory conditions.

The studies by Brueckner and Lai (1996), Sasaki (1998) and Lai and Yang (2002) analyse questions concerning optimum control levels, which means they are less relevant to the objectives of this paper.

Pogodzinski and Sass (1990) provide an exhaustive well-structured overview of the literature on zoning. See also the work by Stull (1974), Rossi-Hansberg (2004) and Fischel (2004).

The work of Glaeser et al. (1995) was, to some point, a precursor to this type of analysis, by including the whole of the population in a metropolitan area, along with certain localization dummies, as determinant factors of the growth of cities.

For example, an insightful review of the literature that studies the impact of the initial level of education in the growth of a cross-section of countries can be found in Krueger and Lindahl (2001). Glaeser et al. (1995) follow the same approach in their study of population growth across US cities.

Note also that the constant term in the equation is a linear combination of the constants of equations (8)–(10) and the logarithm of the variation of utility between t and t+1. Therefore, it does not have a direct interpretation.

The first database contains information on the general planning and development forms used by Andalusian municipalities and the adaptations that have occurred in these urban planning regulations. The second provides information on the level of facilities in the Andalusian municipalities and, more importantly, the programming for residential and industrial land that, by 1995, the region's towns and cities had done. The third shows, for this same year, land uses in Andalusia on a municipal level, differentiating between urban land, industrial land, other built-up areas, agricultural land and protected areas. The fourth database includes exhaustive socioeconomic information for the region, detailed by municipalities.

In our database, 72 municipalities belong to a metropolitan area. Of these, 24 had a PGM in 1995, the remainder having NNSS.

The main explanation of this fact to be found in the literature is that urban planning policy allows to establish detailed zoning regulations. Thereby, it facilitates the residents' control over the city's population growth (in addition to the type of citizens arriving in a city). See, for example, Calabrese, Epple and Romano (2007).

Note the data on the amount of land programmed for residential use refer to the moment in which the plan is approved.

Land programmed for residential (or industrial) use is defined as that which the plan allows to be developed.

These are: Almería, Cádiz, Córdoba, Granada, Huelva, Jaén, Málaga and Sevilla.

This has been previously documented in the empirical literature for the metropolitan areas around central cities. See, for example, Sole and Viladecans (2004) or Voith (1998). Our data suggest that it is also true for more distant locations, as population growth is negatively related to distance to the central city.

We include this variable in logarithmic form to obtain an estimate of the elasticity of the rate of growth with respect to the initial level of population. RESET tests available from the authors upon request support our choice.

Following Godfrey and Orme (1994), we use the square, cube and fourth powers of the fitted value as test variables. These are added to the regressors in a new regression and the sample value of the F-test statistic compared against the critical value of the F(3,524−k−3) distribution. The null hypothesis is that all parameter estimates of the test variables are equal to zero.

The value of the test statistic is 2.429, while the critical value of the relevant null distribution F(7,521) for a 95% confidence level is 2.009.

See Footnote 14 and Fishel's (2001) “homevoter” hypothesis.

Recall that all municipalities within metropolitan areas had either NNSS or PGM in our period of study.

This is true when a one-tail test is carried out, as our starting hypothesis regarding this variable justifies.