When do better schools raise housing prices? Evidence from Paris public and private schools

https://doi.org/10.1016/j.jpubeco.2009.10.009Get rights and content

Abstract

In this paper, we investigate how housing prices react to the quality of education offered by neighboring public and private schools. The organization of secondary schooling in the city of Paris, which combines residence-based assignment to public schools with a well-developed and almost entirely publicly funded private school system, offers a valuable empirical context for analyzing how private schools affect the capitalization of public school performance in housing prices. Using comprehensive data on both schools and real estate transactions over the period 1997–2004, we develop a matching framework to carefully compare sales across school attendance boundaries. We find that a standard deviation increase in public school performance raises housing prices by 1.4 to 2.4%. Moreover, we show that the capitalization of public school performance in the price of real estate shrinks as the availability of private schools increases in the neighborhood. Our results confirm the predictions of general equilibrium models of school choice that private schools, by providing an advantageous outside option to parents, tend to mitigate the impact of public school performance on housing prices.

Introduction

An important issue raised by residence-based school assignment policies is the extent to which they affect both educational and residential segregation. Theoretical models (including Bénabou, 1993, Fernández and Rogerson, 1996, Epple and Romano, 2003, Rothstein, 2006) have shown that in the presence of peer effects in education, such policies yield income and residential sorting in equilibrium. One of these models' key parameters is the willingness of parents to pay for school performance. The higher the value of this parameter, the higher the level of sorting when admission to schools is tied to residence. However, this sorting process can be reversed once school choice is introduced (Epple and Romano, 2003, Rothstein, 2006). In particular, private schools, which operate under different admission rules, expand educational choices and might mitigate sorting effects. In this paper, we test the theoretical predictions of models of housing markets in which public and private schools coexist (Nechyba, 1999; 2000; 2003). We estimate the impact of school performance on housing prices and investigate how this effect varies with the availability of private schools in the neighborhood.

Several empirical papers have sought to test the empirical prediction that housing prices should be higher in areas where schools perform better. The estimation is complicated by the endogeneity of school performance in the housing price equation, since better schools tend to be located in wealthier neighborhoods and pupils drawn from privileged socio-economic backgrounds generally have higher academic achievement. Traditional hedonic regressions in which housing prices are regressed on school performance, controlling for housing and neighborhood characteristics, are therefore likely to be biased.1 In her 1999 paper, Black proposed a method to solve this problem, by comparing the prices of houses located near primary school attendance boundaries and sufficiently close to share the same neighborhood characteristics. Across these boundaries, mean differences in housing prices are assumed to reflect differences in school test scores only. Restricting the sample to sales located within 0.15 mile of a boundary, Black finds that a 5% increase in primary schools' test scores (approximately one standard deviation) raises housing prices by 2.1%. On UK data and using an alternative estimation strategy to correct for spatial fixed effects,2 Gibbons and Machin, 2003, Gibbons and Machin, 2006 find an effect of the same order of magnitude for primary schools. A study by Bayer et al. (2007) yields similar results using a discrete choice model instead of the usual hedonic approach. Furthermore, these authors provide evidence of significant heterogeneity in the marginal willingness of households to pay for school performance. A few papers have also exploited changes in school boundaries over time (Bogart and Cromwell, 2000, Reback, 2005) and show that housing prices react to these changes. In the longer run, however, Kane et al. (2006) stress the difficulty of separately identifying the effect of school performance from the effect of household sorting along school attendance boundaries.

In order to better understand the policy implications of such findings, several papers (including Downes and Zabel, 2002, Kane et al., 2003, Figlio and Lucas, 2004, Brasington and Haurin, 2006, Clapp et al., 2008, Gibbons et al., 2009) have tried to identify which school characteristics are valued by parents. Empirical findings suggest that parents are more sensitive to student peer quality than to the quantity of school resources. The evidence is more mixed as to whether information on school effectiveness (essentially in the form of valued added measures of school performance) affects school markets or not (Kane and Staiger, 2002, Hastings et al., 2007, Mizala and Urquiola, 2008).

If parents strongly value the peer quality of their children's classmates, then school admission rules can serve as an important policy tool to reduce educational inequalities. However, a proper evaluation of the potential effects of alternative school allocation schemes requires the inclusion of both public and private schools in the analysis. While most authors acknowledge the possibility that private schools are likely to affect residential and educational choices, almost all existing studies have focused exclusively on sorting into public schools. To the best of our knowledge, our paper is the first to investigate the effect of private schools on the housing price premium attached to public school performance.

With respect to this existing literature, our paper innovates in two main directions. First, we improve the estimation strategy originally proposed by Black (1999) to deal with the endogeneity of school performance in the housing price equation. We do so by developing a matching framework which consists in the careful comparison of sales across public middle school attendance boundaries. We also test the validity of this identification strategy by checking precisely if other neighborhood characteristics vary discontinuously across boundaries.

Second, we investigate whether private school choice reduces the capitalization of public school performance in housing prices. We use comprehensive data on both schools and real estate transactions in the city of Paris over the period 1997–2004 to study how the local availability of private schools affects the willingness to pay for better public schools. The organization of middle schooling in the city of Paris, which combines residence-based assignment to public schools with a well-developed and almost entirely publicly funded private school system, offers a valuable empirical context for analyzing this issue.

Using different measures of school performance, we find that a standard deviation increase in school performance raises housing prices by 1.4 to 2.4%. The size of this effect is similar to existing US and UK estimates and can explain roughly 5% of observed differences in housing prices between adjacent school zones. We also find that the price premium attached to better performing public schools exhibits spatial heterogeneity and varies with the availability of private schools in the neighborhood. In line with the theoretical predictions of general equilibrium models of school choice, the presence of good private schools in certain neighborhoods tends to attenuate the capitalization of public school performance in housing prices, by providing an advantageous outside option to parents.

The remainder of this paper is as follows: Section 2 describes the estimation strategy used to measure the impact of public school performance on housing prices; Section 3 briefly presents the French educational system and school admission rules; Section 4 gives a description of the data; Section 5 presents the basic results regarding the capitalization of public school performance in housing prices while Section 6 evaluates how this capitalization varies with the local density of private schools.

Section snippets

Estimating the impact of public school performance on housing prices

Economists have traditionally relied on hedonic models to estimate the impact of school performance on housing prices. The standard hedonic housing price function describes the price of a particular sale as a function of the flat's observable characteristics, which include its intrinsic features (size, floor, etc.) as well as its neighborhood characteristics. The corresponding coefficients are interpreted as measuring the marginal purchaser's willingness to pay for each specific characteristic.

The French educational system

The French educational system is highly centralized and fairly homogenous until pupils reach the age of 14. Children spend 5 years in primary school (ages 6 to 10), 4 years in middle school or Collège (ages 11 to 14) and 3 years in high school or Lycée (ages 15 to 17). The school year starts in September and ends in July. The minimum school leaving age is 16 years.

Education in France is predominantly public,12

Data and summary statistics

To estimate the impact of school performance on housing sales in Paris, we collected data on school zones, school characteristics, individual property sales and local socio-demographic characteristics during school years 1997–2004.

The impact of public middle school performance on housing prices

In this section, we use the matching framework presented in Section 2 to estimate the extent to which the performance of Parisian public middle schools is capitalized in housing prices.

The mitigating effect of private schools

Previous studies have mainly focused on public schools, ignoring the other options provided by private schools in their analysis.30 Yet Epple and Romano (1998) show that it is crucial to take into account both public and private schools when studying sorting into

Conclusion

Using comprehensive data on middle schools and housing sales in Paris over the period 1997 to 2004, we find that the performance of public schools has a significant impact on housing prices by comparing price and school performance differentials across school attendance boundaries. A standard deviation increase in the average exam score at the school level raises housing prices by 1.4 to 2.4% depending on the chosen index of school performance. The size of this effect is similar to existing

Acknowledgements

The authors wish to thank Sandra Black, Esther Duflo, Robert Gary-Bobo, Steve Gibbons, Caroline Hoxby, Francis Kramarz, Steve Machin, Éric Maurin, Philip Oreopoulos, Thomas Piketty and Miguel Urquiola for their helpful comments and suggestions. We are also grateful to Florence Defresne, Françoise Œuvrard and Fabienne Rosenwald from the French Ministry of Education, Patrick Caille and Alain Mariani from the Rectorat de Paris for their precious help in collecting the data. This work was

References (34)

  • D. Brasington et al.

    Educational outcomes and house values: a test of the value added approach

    Journal of Regional Science

    (2006)
  • K.Y. Chay et al.

    The central role of noise in evaluating interventions that use test scores to rank schools

    American Economic Review

    (2007)
  • P. Cheshire et al.

    Capitalising the value of free schools: the impact of supply characteristics and uncertainty

    Economic Journal

    (2004)
  • S. de Monredon

    Les élèves du premier degré à la rentrée 2007 dans les écoles publiques et privées

  • D. Epple et al.

    Competition between private and public schools, vouchers, and peer-group effects

    American Economic Review

    (1998)
  • D. Epple et al.

    Neighborhood schools, choice, and the distribution of education benefits

  • R. Fernández et al.

    Income distribution, communities and the quality of public education

    Quarterly Journal of Economics

    (1996)
  • Cited by (193)

    • Residence-based priorities and school choice

      2023, Economics of Education Review
    View all citing articles on Scopus
    View full text