There might be distinct reasons for people to emigrate to another country. Whatever the reason, may it be for personal, political or economic reasons, whenever a large quantity of people immigrates into another country, it can affect the local population and economic situation of the destination country. One of those factors that may be affected by increased rates of immigration is housing prices and rents. Multiple studies exist which have already investigated the effect of immigration on housing prices using samples from different countries and varying degrees of aggregation. Saiz, Degen and Fischer and Gonzalez and Ortega look at the relationship in US metropolitan areas, swiss districts and spanish provinces, respectively, and find an unanimously positive effect of immigration on housing prices [1-3]. In contrast, however, Saiz and Wachter look at neighborhoods on metropolitan areas in the US and find a negative relationship between immigration and housing prices [4]. These contradicting findings may suggest that the level of aggregation of the data has an effect on the empirical relationship between these variables. Correspondingly, Sá investigates the effect of immigration on housing prices in the UK disaggregated by local authority. In line with the findings of Saiz and Wachter, she finds that increases in immigrant population leads to a decrease in housing prices [4]. To further probe into this relationship, Sá develops a series of models incorporating different factors that may contribute to this relationship. This paper aims to provide an overview of the findings presented by Sá [5]. Additionally, the results serve as input for a policy proposal which may alleviate the negative relationship between the rate of immigration and housing prices.

The remainder of the paper is structured as followed: The first part presents and overview of the study conducted by Sá including the model development, reasoning and main findings. The second part presents the policy proposal by discussing one of the underlying problems according to Sá, proposing improvements and discussing the advantages and disadvantages of its implementation [5].

In the past decades, the UK has seen an increase in immigration as well as an increase in housing prices. While the share of immigrants has increased by 5 percent points in the past 30 years, average house prices have increased from £ 60,000 to £ 180,000. While the assumption that the increase in house prices may be solely caused by the larger share of immigrants within the population is far-fetched, there might still be a causal relationship between those two observations. Past research regarding this relationship has produced quite contradicting findings which may be due to sample size and level of aggregation. To investigate this topic further, Sá looks at the causal relationship between increases in the share of immigrants and house prices in the UK at the local authority level [5]. This level of disaggregation is in line with the past research which suggests that a negative relationship exists at higher levels of disaggregation but this disappears or changes at higher levels of aggregation. Additionally, Sá then investigates further factors which may cause this relationship based on previous research.

The simple theoretical model to evaluate the relationship between a change in housing prices and a change in the stock of immigrants is based on previous research conducted by Saiz [1]. This model takes natives preferences for immigrants into account and allows high income natives to move to another city if their preference is low and the marginal utility of moving is high. Low income natives are assumed to be incapable of moving due to financial constraints. Moreover, the total demand and supply of housing is incorporated in the model as well as budget constraints and income.

The data for analysis is retrieved from the UK Labor Force Survey and Land Registry. After slight manipulation of the data to ensure a usable data set, the data covers the period from 2003-2010 over 170 local authorities and a working age population of 16-65 years.

To investigate the relationship between changes in the stock of immigrants and changes in housing prices, an OLS regression analysis is conducted. However, failure to include the fact that immigrants to not settle randomly but rather according to historic settlement pattern in areas of already high immigration, leads to incorrect findings. Accordingly, an instrumental variable is developed to account for this observation. The IV shows to be statistically significant in almost all models. Additionally, to account for the problem of spatial correlations being due to common fixed influences, the dependent variable is used in first differences which eliminate any time-invariant, area specific variables. Since the OLS results appear to be flawed due to the ignorance of historical settlement patterns of immigrants, only the IV results will be discussed in the remainder of this paper.

Sá finds a negative relationship between a change in stock of immigrants and house prices. The strength of this relationship differs with different levels of lag included in the model. Without a lag on socio-economic variables (unemployment rate, benefits rate, crime rate, average dwelling stock, index of housing quality), a 1% increase in the stock of immigrants leads to a 1.786 % decrease in house prices. Equivalent results are found if these variables are lagged by 1 year. If a lag of 3 years on socio-economic variables and changes in housing prices is included, a decrease of 2.879% in house prices is found. All results are statistically significant at the 1% significance level.

Previous research suggests that this effect may be due to immigrants’ level of education with immigrants with lower levels of education driving down housing prices. To investigate this hypothesis, Sá conducts a further analysis taking this into account.

To evaluate the effect of immigrants’ level of education and house prices, Sá divides the local authorities into four groups based on the quartiles of regional distribution of the share of immigrants with education levels below GCSE. The analysis is conducted for each quartile separately. The results show a negative relationship only in those quartiles, in which the share of lower educated immigrants is high. A 1% increase in share of immigrants leads to a decrease in house prices between 1.3% and 1.9% in quartiles 3 and 4, respectively. This supports the hypothesis that the level of education of immigrants plays a determining role in driving down house prices. There are different reasons which may lead to this observation. Firstly, natives’ preferences for immigrants may lead natives to move out of the area and being displaced by immigrants with lower household income. In response, the income of the area may decrease leading to the inability to pay high rents. Similarly, if immigrants are substitutes in production, which is frequently the case in low-skilled manufacturing jobs, average wages in the area may decrease due to immigrants often earning lower wages than locals. Again, average income decreases driving down house prices and rents. Saiz and Wachter theorize that increases in immigrants may also lead to increases in crime and decreases in the quality of housing stock which may also drive down house prices. To investigate this reasoning, Sá conducts further analyses [4].

The displacement effect between immigration and native population has already been subject to previous research [6-8]. The model to investigate the displacement effect in this sample is derived from Card which was deemed accurate by Peri and Sparber who investigated the bias inherent in those models [9]. The analysis by Sá shows an almost one-for-one displacement of natives by immigrants, significant at the 1% significance level, evaluated by local authority, year and education group. This shows that for every immigrant moving into a certain area, one native moves away to another local authority.

A further factor which may drive down house prices is the decrease of local wage levels due to increases in the stock of immigrants. Based on a model developed by Dustmann, et al. Sá evaluates the effect of increases in the stock of immigrants on natives’ wages [10]. Data regarding the gross hourly wage is retrieved from the UK Labor force Survey. To eliminate the effect of outliers which may skew the data, the bottom and top percentile is capped from the data sample. Even though the results show a negative effect, which is stronger at the lower end of the local wage distribution, the results are statistcially insignificant suggesting that immigration does not affect natives’ wages.

Even though native wages may not be influenced by changes in the stock of immigrants, the displacement effect observed in the previous analysis may lead to changes in the local wage distribution. To further evaluate this hypothesis, Sá uses the model developed by Card to investigate this relationship [7]. National wages are divided into four quartiles and a regression is run on each quartile separately. The results show that findings are only statistically significant in Q1 (β=0.204, p≤0.01) and Q4 (β= -0.085, p≤0.05). This indicates that an increase in stock of immigrants’ equal to 1% leads to an increase of 2 percentage points of workers in the lowest wage quartile and a decrease of 0.8 percentage points of workers in the highest wage quartile.

Further considerations regarding the negative relationship between immigration and house prices include crime and the quality of the housing stock. Assuming that increases in immigrants may lead to increased rates of crime and a decrease of the quality of local goods and housing stock, an increase in immigration reduces the quality of a local authority and thus leads to a reduction of house prices. However, no such relationship could be established.

To examine the robustness of the analyses conducted above, Sá repeats the analysis using data from the Worker Registration Scheme which encompasses immigration data from the A8 accession countries since 2004. Since every worker had to register, the data offers the advantage of being more administrative and thus reducing the sampling error. The results confirm previous findings that increases in the stock of immigrants leads to a reduction in house prices. In this dataset, the effect is even stronger with a decrease of 3.6%. This may be due to immigrants from the A8 countries often being younger, male, single and childless.

A further consideration is the supply of housing in response to increases in the stock of immigrants. Based on the model by Gonzalez and Ortega and using a lag of 1 and 3 years, Sá finds a negative relationship with 48 and 89 fewer units being supplied in response to a 1% increase in the stock of immigrants [3].

Conducting the analysis on a more aggregated level, she finds no statistically significant relationship between changes in immigration and house prices supporting previous studies that there is no statistically significant effect when examining aggregated regions. This highlights the importance of evaluating the relationship at a more disaggregated level.

With increasing rates of immigration, understanding its effect on local, political and economic factors becomes increasingly important for policy makers. Sá investigates the effect of changes in the stock of immigrants on house prices and even delves further into this topic by examining possible reasons for her finding. Overall, she finds that increases in the stock of immigrants leads to a reduction in housing prices particularly when immigrants are lower educated, an almost one-for-one displacement of natives, a higher fraction of individuals in the lowest wage quartile and fewer housing units. However, native wages and the level of crime and quality of housing stock appear to be unaffected. Her study contributes to scientific research by not only examining these relationships at a more disaggregated level but also by including an instrumental variable which takes historic settlement pattern of immigrants into account and thus addressing an endogeneity problem inherent in this analysis.

Based on the research conducted by Sá, this part of the paper proposes a policy to alleviate the negative relationship between immigrants and house prices. The analysis shows that it is mainly the level of education of immigrants which leads to a decrease of house prices. Thus, the policy proposal focuses on immigrant education, specifically the provision of ESOL (English for Speakers of Other Languages) courses to improve immigrants’ employability and wages.

One of the major determinants of employability of immigrants in the UK is their English proficiency which raises employment rates by up to 22% [11]. Similarly, immigrants with high language skills are less likely to be employed in low-skilled jobs and are more likely to understand their employment contracts better thus can negotiate better terms and conditions. Additionally, low levels of English proficiency can have other negative effects as well including the inability of integrating into the local society. Even though the UK Government funds adult further education and ESOL courses for immigrants, the budget has been cut by almost 40% in recent years (Figure 1) [12]. This government decision is quite puzzling in face of increasing rates of immigration. The underfunding in recent years has led to a great discrepancy between demand and supply which manifests itself in long waiting lists and the underperformance of students’ due to infrequently offered courses. Over 80% of ESOL course providers report to have waiting lists with 66% of respondents attributing this circumstance to a lack of funding [12]. Moreover, there is currently no national strategy regarding ESOL provision in place which has led to great differences in course offerings and quality across local authorities. Since English proficiency has such a significant effect on employability and other socio-economic factors, such as integration and even state of health, its provision and strategy should be improved to respond to current requirements.

Figure 1: ESOL Budget and Number of Immigrants in recent years.

One of the first steps to improve the provision and quality of ESOL courses and thus raise the employability of immigrants is the creation of a centralized, national strategy. Moreover, budgets should be estimated and allocated to local authorities based on the expected number of immigrants. Thus, assuming a total of 600 learning hours within a 2-year course and average hourly costs of £5.73 per student, the average cost per student amounts to £3,314 [13]. Given the expected number of immigrants, the current ESOL budget is far below the required funding for adequate ESOL provision (Figure 2). To counteract this problem, national and local ESOL budgets should be estimated and approved based on the estimated demand.

Figure 2: Actual ESOL Budget versus Required Budget.

Even though an increase in ESOL funding puts a financial strain on the UK government, the costs of ESOL provision per student can theoretically be recovered through income taxes within the first two years of employment. This is based on an estimated salary of £18,000 - £25,000. Even in case of immigrants out-migrating again before entering the workforce, the income tax effect supplied by the remaining immigrants over their remaining time ensures a positive net present value of this investment. Moreover, higher levels of English proficiency have a positive effect on the immigrants’ ability to integrate into UK society. This may, in turn, alleviate the displacement effect found by Sá by reducing natives’ preferences for moving out of highimmigrant neighborhoods. Additionally, the eligibility for higher paid jobs influences household income positively. On the other hand, however, the centralization of a national ESOL strategy requires additional efforts and costs for its development, coordination and supervision. However, the advantages elaborated above are likely to counteract the disadvantages of increased costs and effort in the long run.

Sá shows in her analysis that changes in the stock of immigrants has a negative effect on house prices if immigrants are lowly educated. This also includes low levels of English proficiency which reduces employability and inhibits integration. To counteract this problem, the UK government offers ESOL courses to immigrants to improve their language skills. However, the funding for this kind of education has been cut dramatically in recent years leading to the under-provision of courses. Since English proficiency is such a major determinant of employability and thus household income, changes should be implemented to improve the current situation. Thus, a national strategy of ESOL provision should be developed and implement with a budget estimated according to the anticipated demand. Even though this leads to increased financial costs to the government, these costs can be recovered through income taxes in the following years. Even in case that immigrant decides to out-migrate in the near future, the income tax effects of the remaining immigrants can be expected to provide a positive net present value in the long-run. Moreover, better integration may alleviate the displacement effect of immigrants and natives. Thus, by tackling the problem of low education and English proficiency of immigrants, the negative relationship between immigration and housing prices may be weakened.