Spatial and Temporal Diffusion of Housing Prices in the Presence of a Linguistic Border: Evidence from Belgium

Abstract

This paper investigates the ripple effect hypothesis for Belgium and is especially interested in the existence of potential border effects arising from the language border that divides the country in two large linguistic regions, Flanders and Wallonia. We find that housing prices in districts located along the north-south axis are highly integrated, while those in more peripheral eastern and western districts converge almost exclusively with neighbouring districts in the same linguistic region. Where a similar model might only applied to a few multilingual countries, we provide some valuable insights on the mechanisms responsible for the observed ripple—and border effects.

Diffusion spatiale et temporelle du prix de l'immobilier en présence d'une frontière linguistique : le témoignage de la Belgique

La présente communication se penche sur l'hypothèse de l'effet d'entraînement pour la Belgique, en particulier sur l'existence d'effets de frontière potentiels découlant de la frontière linguistique divisant ce pays en deux grandes zones linguistiques : les Flandres et la Wallonie. Nous établissons que le prix de l'immobilier dans des districts situés le long de l'axe nord-sud sont extrêmement intégrés, tandis que ceux de districts orientaux et occidentaux plus périphériques convergent presque exclusivement avec des districts voisins dans la même zone linguistique. Lorsqu'un modèle similaire n'est peut-être applicable qu’à un nombre limité de pays multilingues, nous dévoilons des aspects importants sur les mécanismes déterminant les effets d'entraînement et de frontière relevés.

Difusión espacial y temporal de los precios de la vivienda en presencia de una frontera lingüística: evidencia de Bélgica

Este estudio investiga la hipótesis del efecto dominó en Bélgica y está especialmente centrado en la existencia de los posibles efectos fronterizos derivados de la frontera lingüística que divide al país en dos grandes regiones lingüísticas: Flandes y Valonia. Comprobamos que los precios de la vivienda en los distritos situados a lo largo del eje norte-sur están muy integrados, mientras que los de los distritos más periféricos del este y oeste convergen casi exclusivamente con los distritos colindantes en la misma región lingüística. Cuando solo se puede aplicar un modelo similar a algunos países plurilingües, proporcionamos información valiosa sobre los mecanismos responsables de los efectos de dominó y frontera observados.

标题 语言边界下住房价格的空间和时间扩散 : 比利时实证

摘要 本文探讨了关于比利时的涟漪效应假设, 着重将该国分为佛兰德和瓦龙两大语言区的语言边界产生的潜在边界效应。我们发现, 位于南北轴上的地区的住房价格十分统一, 而比较边缘的东部和西部地区的房价则几乎仅与属于同一语言区的邻近地区趋同。尽管类似的模型可能仅适用于几个多语言国家, 我们为所观察到的涟漪和边界效应的相关机制提供了一些具有价值的见解。

Keywords:

• Housing prices
• spatial dependence
• cross-sectional dependence
• linguistic border

Acknowledgements

The Flemish Policy Research Center on Housing analyses the (Flemish) housing market and housing policy. The authors would like to thank Geoffrey Meen, Frank Verboven, Frank Vastmans, Sven Damen, Annelore Van Hecke and the participants of the housing economics workshop of the ENHR conferences in 2012 in Lillehammer and in 2013 in Tarragona for valuable suggestions and comments on earlier drafts of this work.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1. Villa's, mansions and bungalows is a small subcategory with in general high transaction prices. Inclusion of these transactions might lead to unwanted outliers. Therefore these transactions are dropped out of our sample. (Residential) land is left out for obvious reasons.

2. Before the reform of 2012/2013, there were 27 judicial districts in Belgium. In the empirical analysis we use 20 district. To this end we have merged the districts of Tournai and Mons (Tournai), Verviers and Eupen (Verviers), Arlon, Neufchâteau and Marche-en-Famenne (Arlon), Huy and Liège (Liège), and Tongeren and Hasselt (Hasselt).

3. Data concerning the CPI is provided by Statistics Belgium and is unfortunately not available on a more disaggregated level.

4. Everything is related to everything else, but near things are more related than distant things.

5. Seemingly Unrelated Regressions.

6. The port of Antwerp provided employment for 145,836 full time equivalents (FTE) in 2010, of which 60,509 direct and 85,327 indirect. The 10 largest employers were: BASF, BNRC Group, Public sector, Antwerp Port Authority, General Motors Belgium, ExxonMobil Petroleum & Chemical, PSA Antwerp, M.S.C. Home terminal, Electrabel and Total Refinery Belgium. The port of Antwerp was responsible for approximately 19.2 billion euros of value added, which was approximately 5.5% of the total value added in Belgium in 2010 (Mathys, 2012).

7. In Appendix 2 we present the GIRFs for each district separately together with their bootstrapped (5000 replications) confidence intervals.

Additional information

Funding

The support of the Flemish Policy Research Center on Housing (Steunpunt Wonen) is gratefully acknowledged.