Valuing School Quality via a School Choice Reform

• DOI: http://doi.org/10.1111/sjoe.12133
• Published date: 01 January 2016
• Author: Kjell G. Salvanes,Stephen Machin
• Date: 01 January 2016

Scand. J. of Economics 118(1), 3–24, 2016

DOI: 10.1111/sjoe.12133

Valuing School Quality via a School

Choice Reform∗

Stephen Machin†

University College London, London, WC1H 0AX, UK

s.machin@ucl.ac.uk

Kjell G. Salvanes

Norwegian School of Economics, NO-5045 Bergen, Norway

kjell.salvanes@nhh.no

Abstract

Among policymakers, educators, and economists, there remains a strong debate on the extent

to which good schools matter. In this paper, we estimate school quality valuations based

upon house prices, by considering a policy reform regarding pupils’ choices to attend high

school. We exploit a change in school choice that occurred in Oslo county in 1997, where

school authorities altered policy from one based on catchment zones to an open enrollment

policy that allowed pupils to apply to any high school. Our estimates show that parents

substantially value better performing schools because the sensitivity of housing valuations

to school performance falls significantly (by over 50 percent) following the school choice

reform that made the move to open enrollment.

Keywords: House prices; school choice reform; school quality

JEL classification:C21; I20; R21

I. Introduction

The way in which housing markets are affected by differences in local

school quality is an important economic and social phenomenon, and has

become a highly topical and politically charged issue. All over the world,

parents are concerned that they obtain the best education for their chil-

dren. Empirical evidence shows that parents are willing to pay significant

amounts of money to buy houses located in the catchment areas of better

performing schools.1

†Also Research Director, Centre for Economic Performance, London School of Economics.

∗We would like to thank Sandy Black, Steve Gibbons, Caroline Hoxby, Sandra McNally,

Olmo Silva, two anonymous referees, and numerous seminar and conference participants for

some very helpful comments. We also thank Roger Bivand for his great help in producing

the figures.

1See the literature reviews of Black and Machin (2011), Gibbons and Machin (2008), and

Machin (2011).

CThe editors of The Scandinavian Journal of Economics 2015.

4Valuing school quality via a school choice reform

Many papers in the research literature on valuing school quality show

that there are sizable and significant housing capitalizations linked to school

quality (as typically measured by test scores). However, an ongoing issue

in this literature concerns the means by which the impact of school quality

on housing valuations is identified. This is pertinent because the main

empirical challenge is to ensure that results capture only the portion of

housing expenditures affected by variations in school quality, and not due to

other (observed and unobserved) housing characteristics or local amenities.

Early work tried to net out these correlated effects by conditioning on a

range of observables. More recent work has tried to deal more effectively

with the unobserved aspects. One strand models variations in house prices

across school catchment area boundaries in an attempt to net out unob-

served heterogeneity that could contaminate any house price–school quality

relation seen in the data.2Others have adopted a quasi-experimental ap-

proach based on the redrawing of catchment area boundaries.3Yet another

strand adopts a more structural approach, borrowing from the industrial

organization and consumer demand literatures, and sometimes combining

with the boundaries approach (see, for example, Bayer et al., 2007).

Despite these advances in methodology, there remain concerns and wor-

ries. For example, some critics argue that administrative boundaries are

endogenous to housing prices, possibly because of differential growth over

time or because they have differing institutional features (e.g., in the US,

local property taxes might differ on either side of the boundary). Simi-

larly, quasi-experimental variations based on redrawing boundaries are also

likely to be endogenous to school quality and/or house prices (i.e., there

are correlated reasons as to why the boundaries were redrawn).

In this paper, we adopt a different approach to causally identify the

relationship between house prices and school quality. We also base our

approach on a quasi-experiment, but one that is very different to almost

all of those considered to date.4We implement a research design based

on a school admissions reform. The reform increased pupil choice to high

schools by allowing mobility across catchment area boundaries that was

2There are by now quite a few boundary discontinuity papers. See, inter alia, the early papers

by Bogart and Cromwell (1997) and Black (1999), and the more recent work combining the

boundary approach with other empirical methods by Gibbons and Machin (2003), Gibbons

et al. (2013), and Bayer et al. (2007).

3In the context of a loss of neighborhood schools and re-districting in Cleveland, Bogart and

Cromwell (2000) adopt a quasi-experimental approach based upon the redrawing of catchment

area boundaries. Ries and Somerville (2010) also treat a redrawing of school catchment zones

in Vancouver in September 2001 (approved in January 2001) as an experiment that they claim

exogenously induces changes in school quality, which in turn can affect housing valuations.

4However, see Reback (2005) for closely related work relying on the diminished importance

of catchment area boundaries.

CThe editors of The Scandinavian Journal of Economics 2015.