fedea

Fundación de Estudios de Economía Aplicada

Immigration and Housing Prices in Spain by Simón Sosvilla DOCUMENTO DE TRABAJO 2008-40

Serie Economía Regional CÁTEDRA FEDEA – Caja Madrid

October 2008

Paper prepared for the Fedea Report 2008

米 FEDEA, Unv. Complutense.

Los Documentos de Trabajo se distribuyen gratuitamente a las Universidades e Instituciones de Investigación que lo solicitan. No obstante están disponibles en texto completo a través de Internet: http://www.fedea.es. These Working Paper are distributed free of charge to University Department and other Research Centres. They are also available through Internet: http://www.fedea.es.

ISSN:1696-750

Jorge Juan, 46 28001 Madrid -España Tel.: +34 914 359 020 Fax: +34 915 779 575 infpub@fedea.es

Abstract:

In this paper we quantitatively explore the role of immigration in the evolution of housing prices in Spain and its autonomous regions for the period 1995-2007. We use estimations of an inverted model for housing demand with and without immigrant flows. The results we have obtained suggest that housing prices in Spain are overvaluated in relation to the value determined by its fundamental variables in a 14.75 percent if we exclude immigrant flows, and in a 7.37 percent if we take them into account. We can consequently conclude that the eventual correction in housing prices to which the Spanish economy is exposed would not be as intense as it has been suggested by some international organisations. For the different autonomous regions, the greater differences in the estimations of overvaluated housing prices are found in Murcia, Canary Islands, Valencian Community, Madrid, Catalonia, Andalusia and Balearic Islands, all of them regions particularly attractive for foreign immigrants as they offer the best employment opportunities. These regions also have a relatively high level of tourist activity and represent a favoured location by many European citizens for the acquisition of a second residence.

Keywords: Housing prices, Spain, Autonomous regions.

Codes JEL: C52, R21, R31

^∗ I gratefully acknowledge the comments and suggestions offered by Michele Boldrin, Antonio Cabrales, Florentino Felgueroso, John Muellbauer and Sara de la Rica, as well as the financial support of Cátedra Fedea-Caja Madrid de Economía Regional. Any opinions, findings or conclusions expressed in this paper are those of the author and do not necessarily reflect the views of Universidad Complutense, FEDEA or Caja Madrid.

1. Introduction

The average rate for the growth of house prices in Spain from 1995 to 2007 was 9.70 percent, while the average rate of inflation for the same period measured by the Consumer Price Index (CPI) was 3 percent.

The intense dynamism registered in house prices can have important macroeconomic consequences because, in the case of Spain, housing represents a fundamental component of household wealth, asserting a significant influence in their consumption and investment decisions.

At the same time, Spain has registered one the greatest rates of immigration of the world (moving from 1 percent in 1995 to 11 percent in 2008) which means that immigration to Spain is three or four times the average immigration rate of the USA and eight times that of France. This immigration flows might have contributed to assert an additional pressure in housing demand and, by extension, in housing prices in Spain. Immediately after their arrival, immigrants tend to rent a shared-dwelling¹, pushing up prices in the housing rental market and indirectly stimulating the overall demand for housing as native landlords, according to the process of housing reposition, take advantage of the rents obtained to finance the acquisition of new housing. As the labour situation of immigrants improves and the process of family reunification is completed, the immigrant population tends to acquire a property of their own and emulate the behaviour of the rest of the population. Indeed, according to the National Survey of Immigrants for 2007, a 40.3 percent of immigrants are on rented property, while the 38.1 percent are home owners.

Some international organisms have suggested that the housing price in Spain may be overvaluated almost by a 20 percent (International Monetary Fund, 2008). The implication of this overvaluation is that the Spanish economy may be exposed to a dramatic price correction. However, these estimations are based on an unexplained increase in housing prices based on a model which does not consider immigration as a determinant variable, resulting in a biased overestimation of the rise of housing prices.

¹ According to the Colectivo IOÉ (2005), the three main housing modalities for immigrants are linked to the length of time they have been resident in Spain and the level of household formation / reunification. At an initial stage, immigrants tend to live in a sub-rented accommodation, afterwards they rent a property, and finally they gain access to home-ownership.

This paper attempts to quantitatively explore the extent to which the recent behaviour of housing prices in Spain and its autonomous regions has moved away from the evolution of its fundamental price as derived from its determinant variables. For this purpose, we are going to employ the estimations obtained from an inverted model of housing demand based on homogeneous available data for the period 1995-2007, which explicitly incorporates the demographic variables derived from the immigration flows.

Another way to analyse the evolution of housing prices is the approach based on assets pricing models which employ the existing similarities between housing and stock-market investments to establish that the housing prices should not be too different from the discounted flow of all future incomes (see, for example, Ayuso and Restoy (2007) and Martinez Pagés and Maza (2007)). Leaving aside the problems associated with the lack of data at regional level about the relevant variables for the empirical assessment of assets pricing, we have decided to employ a demand approach to be able to compare the obtained results with the estimations advanced by international organisms about the overvaluation of housing prices in Spain, because such estimations also adopt a demand approach.

The paper is organised as follows. Section 2 provides the theoretical framework for the empirical analysis. Section 3 describes the data employed in our analysis. Section 4 discusses the empirical results and offers alternative estimations of house price overvaluation in Spain. Section 5 summarises the main conclusions of our analysis.

2. Theoretical Framework

In the theoretical and empirical literature about housing, a number of different variables have been singled out as fundamental determinants in the evolution of housing prices, with factors associated to demand playing a key role (see, for example, European Central Bank, 2003). The predominance of the demand approach is partly the result of the relative rigidity of housing supply in the short-term as well as the lack of reliable and homogenous data about cost and production prices in the construction sector which can affect the behaviour of residential property investment.

Following Cameron et al. (2006), we start from the following demand equation for housing services which we presume to be proportional to the existing housing stock:

\[\ln (h s / p o b) = \alpha \ln (y / p o b) - \beta r _ {h}\tag{1}\]

where ln represents logarithm, hs the housing stock, pob the population, y the real income, and the real rental cost of housing. The coefficients α and are, respectively, the income and price elasticities of the demand for housing services.

Since the housing rental market in Spain is particularly small and not representative of the total housing stock², the real rental cost of housing cannot be directly observed. In equilibrium, however, it should equal the real user costs of housing, so we can write the following expression:

\[h p \cdot (r _ {a} + m + t _ {h} - \Delta \ln h p ^ {e}) \equiv h p \cdot u c _ {h}\tag{2}\]

where represents the real price of houses, tax adjusted real interest rate, m the expenditure on house maintenance and repair, the net rate of tax on housing, ∆lnph the expected rate of appreciation of real house prices and the real user cost of housing, expressed as a proportion of the real price of the house. The last variable provides an estimation of the costs of house tenure which would be incurred during a particular period of time and it is, therefore, a significant factor to take into account when analysing housing demand. Given the difficulties to assess the impact of some of these costs, the measures more commonly employed in practice to capture the user cost of housing combine the nominal interest rate of house mortgages and the expected gain or loss from the variation in residential property prices (∆lnph ) on an interest rate adjusted to adjusted to price expectations

² The last Housing Census for Spain (November 2001) indicates that the percentage of main residential units in the rental market was 11.4 percent. The data from the Household Budget Continuous Survey indicates that in 2005 the percentage of main residential units in the rental market was 9.3 percent.

Solving expression (2) for substituting it in (1) and then inverting it, we obtain the following expression for the inverted curve of housing services demand:

\[\ln h p = \frac {\alpha}{\beta} \ln (y / p o b) - \frac {1}{\beta} \ln (h s / p o b) - \ln u c _ {h}\tag{3}\]

As can be seen, expression (3) establishes that real house prices are positively related with real per capita income , while at the same time such prices depend negatively on the per capita housing stock and the real user cost of housing

The expression (3) can be estimated from the following econometric model:

\[\ln h p = \delta_ {0} + \delta_ {1} \ln (y / p o b) + \delta_ {2} \ln (h s / p o b) + \delta_ {3} r i _ {h} + \varepsilon\tag{4}\]

where lnu is substituted by and where ε represents a random disturbance term.

Alternatively, we can expand equation (4) incorporating the immigration flows , which can modify housing demand, obtaining the following alternative model:

\[\ln h p = \theta_ {0} + \theta_ {1} \ln (y / p o b) + \theta_ {2} \ln (h s / p o b) + \theta_ {3} r i _ {h} + \theta_ {4} \ln d + \xi\tag{5}\]

where represents a random disturbance term.

3. Data

In this paper we have used the available data for Spain and its autonomous regions for the period 1995-2007. This is the longest time series we could elaborate with homogenous data.

The price series in our analysis are provided by the Housing Ministry (Ministerio de la Vivienda) and refer to housing market prices (new or second-hand) in euros by square meter. The data are deflected by the Consumer Price Index (CPI) (national and regional) and reescalated using 2000 as the base year. The hp obtained refers to the constant prices in euros of 2000.

Regarding real income (y), for the case of Spain we have employed data on both household disposable income and Gross Domestic Product (GDP), both figures expressed in million euros of 2000 and collected from the Spanish National Accounts (Base 2000) elaborated by the Spanish National Statistics Institute (INE). Since we do not have homogenous data for the household disposable income in the different autonomous regions, we have taken the real GDP as a proxy variable using the data from the Spanish Regional Accounts (Base 2000, INE). The fact that the results obtained with both variables for Spain as a whole are relatively similar give us certain reassurances when interpreting the regional results.

As for the population data (pob), we use the official figures of the annual revision of the Municipal Census (Padrón Municipal) elaborated by the INE for the period 1996-2007³, using for the year 1995 the de iure population obtained from the historical series of the INE.

For the variable housing stock (hs), we initially intended to use the data of total housing stock for Spain collected by the Bank of Spain. Since this data is not broken down at regional level, we have subsequently decided to employ as a proxy variable the real home equity stock in net terms (in millions euros) elaborated from the estimations of the BBVA Foundation (Mas Ivars et al., 2007) and expanded to the year 2007 with additional data from the data base BDMACRO (Díaz Ballesteros et al., 2008). The results obtained for Spain as a whole with both variables are relatively similar and support the employment of this alternative variable for the regional estimations. It has to be kept in mind that such stock does change annually but we presume the offer to remain fixed throughout the year. In this way, we attempt to capture the existing time gap between the start of the construction process with its required administrative permits and the actual offering of the finished property in the market, which otherwise may result in the attribution of certain rigidity to the offer in the short-term.

³ The data for 1997, not available from the INE, are calculated as the average value between the registered data for 1996 and 1998

The mortgage interest rate expectations , which represents our own approximate variable to real costs of housing usage, has been calculated subtracting from the nominal interest rate for household’s home acquisition offered by the central Bank of Spain the variation rate (national and regional) effectively observed in real house prices.

Finally, for the variable representing the immigrant flows (d), we have used data for the population aged between 20 and 39 years, by nationality, offered by the Spanish Labour Force Survey (EPA). We have chosen this age interval because it is the most commonly employed in the literature as the more relevant for household formation. It is worth mentioning that we have also included data for foreigners with residence permits collected from the statistics elaborated by the Immigration Observatory (Labour and Immigration Ministry), obtaining similar results.

4. Empirical Results

4.1. Model Estimations.

The results of our estimations for Spain and its autonomous regions are shown in Tables 1 and 2, for the specifications (4) and (5) (with and without immigrants respectively) of our inverted model of housing demand.

[Tables 1 and 2 here]

As can be seen, all the variables are statistically significant at the standard levels and show the correct signs. Moreover, the adjustment obtained from the model is particularly high, as clearly evidenced by the resulting high statistics. Finally, it is worth mentioning that validation statistics tests were performed (not included due to lack of space) and did not reveal any wrong specification in the estimated equations.

In the last two columns of Tables 1 and 2, we show the estimation of the income and price elasticities of the demand for housing services (αˆ and , respectively) which can be recovered from the estimated coefficients for and (hs/pob) As it can be observed, for Spain as a whole these elasticities reach the values 1.71 and 0.45 respectively, when we use the model without immigrants (Table 1), and 1.74 and 0.46 when we use the model with immigrants (Table 2), being this difference statistically significant. The values obtained for ˆ are relatively high when compared with the empirical literature, but they may be reflecting, to some extent, the easier access to mortgage lending which existed during the sampled period. Furthermore, the value estimated for is close to the 0.50 value frequently obtained in the empirical literature (Cameron et al., 2006). The average values for for the autonomous regions are 1.84 and 0.44, and 1.85 and 0.45, respectively, when we include or exclude the variable that tries to capture the impact of immigration flows. From Tables 1 and 2, it appears that Madrid is the region which presents a greater elasticity in relation to income (2.00 and 2.01 with or without immigrants), while Galicia has the lower value (1.66 and 1.68 with and without immigrants)

Regarding the effect of mortgage interest rate expectations, the results obtained suggest a semi-elasticity of housing demand in relation to this variable of a –0.02 for Spain as a whole and of an average value of –0.03 for the autonomous regions.

Finally, the estimated coefficient for the proxy variable of immigrant flows is 0.19 for Spain as a whole, with an average value of 0.11 for the autonomous regions.

4.2 Estimation of the overvaluation of housing prices.

Following the common practice of the empirical literature in this area of research, we can interpret the existing gap between the real house price and the estimated house price obtained from the econometric models developed in this paper as a clear indicator of the existence of an overvaluation in relation to the value determined by the fundamental variables. For this reason, Figure 1 shows the percent average deviations registered for the period 1995-2007 between the observed real prices and the adjusted real prices with and without immigrants for Spain and its autonomous regions as alternative indicators of the overvaluation of housing prices.

[Figure 1 here]

As can be seen in Figure 1, we estimate that in Spain the house price is overvaluated in relation to the price determined by the fundamental variables in a 14.75 percent when immigrant flows are not taken into account, and in a 7.37 percent when they are. By autonomous regions, La Rioja, Balearic Islands, Cantabria, Andalusia, Canary Islands, Valencian Community and Catalonia show the higher overvaluation in house prices as estimated with the models which do not incorporate proxies for immigrants; the overvaluation in all cases is above 11 percent. In contrast, Castilla-La Mancha, Asturias and Castilla y Leon are the regions with the lower overvaluation according to these models, in all cases below 9 percent. However, if we use the models which incorporate immigrant household formation, Extremadura, Navarra, Basque Country, Cantabria and La Rioja are the autonomous regions which have the higher overvaluation in house prices (above 9 percent), while Balearic Islands, Murcia and Madrid are the regions where the overvaluation is lower than 6.5 percent.

The differences between the overvaluation estimations based on the obtained estimations with and without immigrants are shown in Figure 2. As can be observed, these differences are above the 4 percentage points for Spain as a whole and for the following autonomous regions: Murcia, Canary Islands, Valencian Community, Madrid, Catalonia, Andalusia and Balearic Islands. These are precisely the regions with a high level of attraction for foreign immigrants, having the best employment opportunities, both in quantitative and qualitative terms, and a substantial representation of labour intensive sectors such as hotel and catering, construction, and intensive green-house agriculture. Likewise, they are predominantly tourist destinations which have experienced significant increases in the number of elderly residents and secondhome buyers, especially from other European countries.

[Figure 2 here]

In order to graphically illustrate our results in an alternative way, we devised the following contrafactual exercise: we proceeded to apply to the market house prices in 1995 the subsequent growth rates actually observed and estimated by the models with and without immigrants. In this way, we obtained the temporal evolution actually registered in the data and the temporal evolution which would have been experienced based on our estimations of the fundamental variables with and without immigrant flows. Figure 3 shows the results for Spain as a whole. As can be seen, the contrafactual fundamental values would have systematically remained below the prices effectively observed in the market. In fact, the average price per square meter during the 1995-2007 was 1212.34 euros, while the average estimated price without taking immigrants into account would have been 1070.31 euros, and 1133.45 euros with immigrants.

[Figure 3 here]

5. Concluding remarks.

The evolution of house prices in Spain and its possible overvaluation have attracted a good deal of attention in the last few years, primarily because housing represents a substantial share of household wealth. It is not surprising, therefore, that housing prices have a significant impact on the distribution of economic welfare and on the consumption and saving decisions taken by a household.

This paper has undertaken a quantitatively explorative exercise to evaluate the role that immigration might have played in the housing prices in Spain and its autonomous regions for the period 1995-2007. We have employed for this purpose the estimations obtained from an inverted model of housing demand with and without immigrants.

The results suggest that in Spain the income elasticity of the demand for housing services is situated at around 1.74, while the price elasticity is estimated to be 0.45 and the user cost semi-elasticity is found to be –0,02. The average values of these elasticities estimated for the autonomous regions are 1.85, 0. 45, and –0.03 respectively.

Moreover, our estimations point out that the house price in Spain are overvaluated in relation to the value determined by the fundamental variables by 14.75 percent if we do not consider the immigrant flows, while this overvaluation is reduced to 7.37 percent if we take immigrant flows into account. Therefore, the correction in the house prices to which the Spanish economy is exposed would not be as intense as suggested by some international organisations. For the autonomous regions, the greatest differences between the estimations obtained from the models with and without immigrants are located in Murcia, Canary Islands, Valencian Community, Madrid, Catalonia, Andalusia and Balearic Islands, all of them regions with high levels of attraction for immigrants due to better employment opportunities and with a significant tourist industry. Many European citizens have also chosen these regions for the acquisition of a second residential property.

As in all empirical works, the results obtained in this paper should be treated with due caution, especially when they are based on a relatively short-term sample and make use of proxy variables for some fundamental determinants. Nevertheless, the results obtained from our estimations seem to confirm that the immigrant flows to Spain from 1995 to 2007 could have represented an important factor in determining the increase in house prices, and that its omission might lead to an overvaluation of house prices in relation to their long-term equilibrium value.

References:

1. Ayuso, J. y Restoy, F. (2006): "House prices and rents: An equilibrium asset pricing approach", Journal of Empirical Finance, Vol. 13, pp. 371-388,
2. Cameron, G., Muellbauer, J. y Murphy, A. (2006): "Was There a British House Price Bubble? Evidence from a Regional Panel" Discussion Paper 5619, Centre for Economic Policy Research.

Colectivo IOÉ (2005): “Inmigración y vivienda en España”, Documentos del Observatorio Permanente de la Inmigración No. 7, Madrid: Ministerio de Trabajo y Asuntos Sociales.

Díaz Ballesteros, A., García, E. y Sosvilla Rivero, S. (2008): Base de Datos Macroeconómicos de España (BDMACRO). Madrid: Cátedra FEDEA-Caja Madrid de Economía Regional.

European Central Bank (2003): Structural factors in the EU housing markets, Frankfurt am Main: European Central Bank.

1. Internacional Monetary Fund (2008): Housing and the Business Cycle, World Economic Outlook, April 2008.
2. Martínez Pagés, J. y Maza, L. A. (2007): “Análisis del. Precio de la Vivienda en. España”, Documento de Trabajo 0307, Servicio de Estudios, Banco de España.
3. Mas Ivars, M., Pérez García, F. y Uriel Jiménez, E. (dir.) (2007): El Stock y los Servicios del Capital en España y su Distribución Territorial (1964-2005). Nueva Metodología. Bilbao: Fundación BBVA.

Table 1: Estimation Results: Model without Immigrants

Source: Own elaboration from estimations of equation (4). Note: t statistics in brackets

Table 2: Estimation Results: Model with Immigrants

Source: Own elaboration from estimations of equation. Note: t statistics in brackets

Figure 1: Estimated House Price Overvaluation Source: Own elaboration from models estimations presented in Tables 1 and 2.

Figure 2: Differences in House Price Overvaluation with and without Immigrants Source: Own elaboration from models estimations presented in Tables 1 and 2.

Table: House Price Evolution in Spain Source: Own Elaboration from Table 1 and 2 estimations

1. 2008-40: “Immigration and Housing Prices in Spain”, Simón Sosvilla.
2. 2008-39: “Modeling the Immigration Shock”, Ana Montes y Michele Boldrin.
3. 2008-38: “Immigration and the Demand for Health in Spain”, Sergi Jiménez, Natalia Jorgensen y José María Labeaga.
4. 2008-37: “Immigration and Students' Achievement in Spain”, Natalia Zinovyeva, Florentino Felgueroso y Pablo Vázquez.
5. 2008-36: “Immigration and Social Security in Spain”, Clara Isabel González, J. Ignacio Conde-Ruiz y Michele Boldrin.
6. 2008-35: “Complements or Substitutes? Immigrant and Native Task Specialization in Spain”, Catalina Amuedo-Dorantes y Sara de la Rica.
7. 2008-34: “Immigration and Crime in Spain, 1999-2006”, Cesar Alonso, Nuno Garoupa, Marcelo Perera y Pablo Vázquez.
8. 2008-33: “A Social Network Approach to Spanish Immigration: An Analysis of Immigration into Spain 1998- 2006”, Rickard Sandell.
9. 2008-32: “The Consequences on Job Satisfaction of Job-Worker Educational and Skill Mismatches in the Spanish Labour Market: a Panel Analysis”, Lourdes Badillo Amador, Ángel López Nicolás y Luis E. Vila.
10. 2008-31: “Students’assessment of higher education in Spain”, César Alonso-Borrego”, Antonio Romero-Medina.
11. 2008-30: “Body Image and Food Disorders: Evidence from a Sample of European Women”, Joan Costa-Font y Mireia Jofre-Bonet.
12. 2008-29: “Aggregation and Dissemination of Information in Experimental Asset Markets in the Presence of a Manipulator”, HelenaVeiga y Marc Vorsatz.
13. 2008-28: “The Measurement of Consensus: An Axiomatic Analysis”, Jorge Alcalde-Unzu and Marc Vorsatz.
14. 2008-27: “Macroeconomic Consequences of International Commodity Price Shocks”, Claudia S. Gómez-López y Luis A. Puch.
15. 2008-26: “The Effect of Short-Selling on the Aggregation of Information in an Experimental Asset Market”, Helena Veiga y Marc Vorsatz.
16. 2008-25: “Adult height and childhood disease”, Carlos Bozzoli, Angus Deaton y Climent Quintana-Domeque.
17. 2008-24: “On Gender Gaps and Self-Fulfilling Expectations: Theory, Policies and Some Empirical Evidence” Sara de la Rica, Juan J. Dolado y Cecilia García-Peñalosa.
18. 2008-23: “Fuel Consumption, Economic Determinants and Policy Implications for Road Transport in Spain”, Rosa M. González-Marrero, Rosa M. Lorenzo-Alegría y Gustavo A. Marrero.
19. 2008-22: “Trade-off Between Formal and Informal care in Spain”, Sergi Jiménez-Martín y Cristina Vilaplana Prieto.
20. 2008-21: “The Rise in Obesity Across the Atlantic: An Economic Perspective”, Giorgio Brunello, Pierre-Carl Michaud y Anna Sanz-de-Galdeano.
21. 2008-20: “Multimarket Contact in Pharmaceutical Markets”, Javier Coronado, Sergi Jiménez-Martín y Pedro L. Marín.
22. 2008-19: “Financial Analysts impact on Stock Volatility. A Study on the Pharmaceutical Sector”, Clara I. Gonzalez y Ricardo Gimeno.
23. 2007-18: “The Determinants of Pricing in Pharmaceuticals: are U.S. prices really so high?”, Antonio Cabrales y Sergi Jiménez-Martín.
24. 2008-17: “Does Immigration Raise Natives’ Income? National and Regional Evidence from Spain”, Catalina Amuedo-Dorantes y Sara de la Rica.
25. 2008-16: “Assessing the Argument for Specialized Courts: Evidence from Family Courts in Spain”, Nuno Garoupa, Natalia Jorgensen y Pablo Vázquez.
26. 2008-15: “Do Men and Women-Economists Choose the Same Research Fields?: Evidence from Top-50 Departments”, Juan J. Dolado, Florentino Felgueroso y Miguel Almunia
27. 2008-14: “Demographic Change, Pension Reform and Redistribution in Spain”, Alfonso R. Sánchez Martín y Virginia Sánchez Marcos.
28. 2008-13: “Exploring the Pathways of Inequality in Health, Access and Financing in Decentralised Spain”, Joan Costa-Font y Joan Gil.