Housing Market Responses to Immigration Shocks; Evidence ... · PDF fileSona Kalantaryan Housing Market Responses to Immigration Shocks: Evidence from Italy. and wage growth. Finally,

Sona Kalantaryan Housing Market Responses to Immigration Shocks: Evidence from Italy.

Housing Market Responses to Immigration Shocks; Evidence from Italy

Ph.D. Student, University of Turin, Italy

(Preliminary draft)

Abstract

In this paper, I examine empirically the impact of immigration shocks on the dynamics of housing prices across Italian provinces during the period from 1996 until 2007. There is massive debate going on upon the impact of current intensive immigration flow on the well-being of native Italian population and Europeans in general. The ongoing research is mainly focused on the influence of immigration on Italian labor market outcome. However, if the ultimate subject of interest is the changes in real income and wealth, then the possible impact of immigration shocks on prices should be taken into account as well. Such an intensive inflow extra consumers with potentially different preferences in the housing consumption leads to an increase in demand for housing, as well as to the segregation among immigrants and natives in the housing market. Housing expenditures represents a big portion of family expenditures for renter and income for house owners. In addition, due to underdevelopment of Italian financial markets it is one of major direction for savings for many Italian households. Hence, the estimation of influence of immigration on housing prices can significantly improve the understanding of immigrants’ influence on real income and wealth of natives.

I used data on the self-reported housing values estimated using data from the Survey of Households Income Wealth in Italy to measure the changes in housing prices in Italian provinces. Using number of valid residence permits as a measure of immigration stock, I find that the increase in concentration of immigrants in Italian provinces has positive but declining effect on average housing prices in province. The instrumental variable estimation thought confirm the positive affect in all used specifications they do not the resulting coefficients at measure of foreign presence turn to statistically not significant ones. Keywords: Housing market, Immigration, House prices, Italy

1


1.1 Introduction and Motivation

There is massive flow of economic literature emerged during last two decades

dedicated to investigation of the influence of immigration shocks on the economic position of

host countries. Such a strong interest to the topic is motivated by recent sharp increase in the

intensity of labor force mobility. The scale and intensity of the ongoing research covers many

aspects of influence of immigration shocks on host economies; already now, it is possible to

draw conclusions about many of them. However, there are still active and open debates going

on upon the influence of immigrants on the well-being of native populations. Still, the

prevailing part of economic literature on immigration is dedicated to the labor market

outcomes. The vast majority of the ongoing research is focused on the influence of

immigration on the employment and wages of host country1. However, a comprehensive

picture can be obtained by considering immigrants not only as extra labor force with possibly

different labor force characteristics, but also as extra consumers with potentially different

preferences in the consumption process.

Estimation of changes in employment and wages alone does not give possibility to

evaluate fully the effect of immigrants on the real income and real wealth of population in the

destination country. Despite their insightfulness, these results tell only part of the story.

Actually if the ultimate subject of interest is the effect of immigration on the real income of

resident population, then the impact of immigration shocks on prices should be taken into

account as well. Indeed, changes in prices clearly have an effect on real wages and real

wealth. Moreover, changes in relative prices may have distributional effects in addition to

those arising from changes in wages.

It is worth to mention that despite the intensity of current economic research on

immigration impact on the labor market, there is no evidence that immigration shocks alter

wages much. For example, the meta-analysis carried out in Poots&Cochrane (2005) based on

eighteen papers from international literature suggests that the effect of immigration on local

wages is very mild; one per cent increase in local labor force leads to reduction in wages less

than 0,1 per cent. To explain the absence of strong reaction of wages to immigration shocks

response economic literature proposes three possible explanations. First, natives may choose

the areas being afraid of possible completion they may face because of immigrants’ inflow

(Filer(1992). Second, immigrants may choose the cities with positive shock in productivity 1 See, for example, Brucker and Jahn (2008), Clark and Drinkwater (2008).

2


and wage growth. Finally, the labor market is more elastic than it is considered (Lewis

(2004)).

Migrating people usually carry with themselves to the country of destination not only

their skills but also their traditions, customs and attitudes; this makes them different from

natives in many dimensions. Among many other fields, the cultural background affects the

behavior of immigrants as consumers. The resulting shift in composition of consumers affects

not only the scale but also the structure of consumption of goods and services in a country or

a region. Those changes in their turn alter the structure of the aggregate demand. The effect is

more vivid once the supply for particular good or service is relatively inelastic. In case of

inelastic supply, the shift in demand results in changes in prices. Housing market is usually

considered as one with inelastic supply in the short run. Hence, the shift of housing demand

due to the inflow of immigrants can alter the housing prices in the area. The resulting changes

cannot leave the real income and wealth of those previously living in the area unaltered. First,

housing represents a considerable share of households’ wealth. Second , the housing-related expenses

represent an important part of overall expenses for majority of households. House-price dynamics are

a key factor in the process of reallocation of household wealth (Davies and Shorrocks, 2000)

interacting with financial asset prices (Sutton, 2002) and conditioning labor mobility (Cannari, Sestito

and Nucci, 2000).

Taking into account the above mentioned arguments, in this paper I evaluate the

influence immigrants may have on the housing market in Italy. The choice of Italian housing

market as the subject for empirical estimation is motivated by several reasons. First, Italy was

traditionally considered as a country facing continues waves of emigration. The situation has

changed dramatically only recently. Immigration became one of the most distinct features of

Italian economic reality during last two decades. The country became a desirable destination

point for hundreds of thousands immigrants with European and non-European origin. The

number of legally registered immigrants increased from 648 to 2.414 thousands during period

from 1992 to 2007. However, the intensity of immigration flow was not homogeneous across

Italian provinces. Such a drastic change could not leave housing market uninfluenced.

Second, the peculiarities of Italian financial markets are such that houses, or real estate in

general, serves as an alternative way of wealth accumulation for many Italian families. Italian

households have very strong preferences towards housing wealth (Brandolini et al., 2004;

Faiella and Neri, 2004). and vivid orientation towards owner occupation (Paiella, 2001; Di Addario,

2002). For the considered period of time dwellings constituted approximately 80 percent of total real

3


assets of Italian households (Cannari et al., 2008). ( Though financial aspects of housing markets

are out of the scope of this paper, it is worth to mention that changes in real price of such an

important component of households` income and expenditures cannot leave real income or

wealth of natives unaltered. Third, according to Venturini&Del Boca (2003) Italian

population is immobile within country; hence, the inflow of immigrants coupled with

immobility of natives can intensify changes in local demand for housing units and housing

prices. Investigation of the link between international immigration and the housing prices in

Italy can serve as good opportunity first, to extend existing research on housing market

response in European region, and second, to enhance the previously done research on

influence that immigration has on real income and wealth of population in Italy.

To understand better the mechanism through which immigrants can influence prices

in general, let us see what the economic theory proposes. Economic theory suggests that

immigration affects prices through different and opposing mechanisms making the overall

effect ambiguous and difficult to predict. On the supply side, production costs may increase

or decrease depending on the way the changes in the overall composition of labor supply

affect relative wages. In a fully traded economy, it would not translate into changes in output

prices, but would rather result in changes in factor intensity or output mix. Still, part of the

output will typically be non-tradable. In this case, one can expect final prices to decrease for

those goods and services produced at lower cost, and reverse result for those goods and

services that immigration has made relatively more expensive to produce. Similarly, the

effect of immigration on the demand side is ambiguous as well. It depends heavily on the

changes that immigration shocks may cause in the composition of consumers, which in its

turn transforms into changes in demand for goods and services. However, these changes are

not necessarily homogenous across different goods and services.

The above-mentioned theoretical insight refers to the response of prices to

immigration shocks in general. What if, the subject of interest is a specific segment of

market. In this particular case, how can housing market respond to the inflow of immigrants?

The response of both demand and supply sides should be taken into consideration. Housing is

considered as a non-tradable good with relatively inelastic supply in the short term. Hence,

qualitative and quantitative changes in housing demand caused by immigration shocks may

translate into changes in housing prices and rents. However, the direction of these changes is

not easy to predict. Immigrants, as additional consumers, do not only create simple increase

4


in aggregate demand but also potentially change its composition. In fact, foreign population

may differ from natives in many aspects including tastes. In other words, immigrants can

have preferences different from the ones specific to native population. For example, due to

low income they may prefer to occupy relatively cheap housing units, or choose to live in

overcrowded flats. The price effect depends also on the reaction of natives to the inflow of

foreigners into area. On the one hand, if immigrants and natives are compliments in

production process, then the immigrants as additional consumer can increase the housing

demand, which may translate to increase in housing prices and rents. On the other hand, if

immigrant and natives are substitutes in the labor market, natives may prefer to leave the area

to avoid possible competition. In this case the outflow of natives may neutralize the effect of

positive immigration shock on housing market. As a result, prices decrease or remain

unchanged. Though housing market can be one of the major non-labor market channels

through which immigrants can influence the well-being of natives, the overall demand effects

are not clear a priori The uncertainty about the direction of the final effect leaves a room for

further empirical analysis.

In this paper, I investigate the effects of the growth in the immigrant population on the

housing prices across Italian provinces over the period from 1996 until 2007.

The rest of the paper is organized in the following way. Section 2 presents and

analyzes the related literature Section 3 presents the identification strategy, and the

econometric issues. Section 4 describes the data. Section 5 reports and discusses the results.

Section 6 concludes the paper.

5


2 Related Literature 2.1 Papers on prices

The literature considering the effect of immigration on prices exists but it is scarce.

There is some limited empirical evidence which is mostly single country analysis focused on

the influence of unskilled immigrants on prices of different goods and services. Besides the

scarcity of research in this particular direction, the prevailing part of it is focused on countries

considered as traditional destinations for immigrants` flows, such as the USA, Canada, New

Zealand.

To my best knowledge, there are three main recent articles, which consider the

influence of immigration shocks on the dynamics of prices. The problem was first elucidated

in Cortes (2008) and then Frattini (2008), who investigate the effect of immigration shocks

on prices in the UK and in the USA respectively. According to Cortes (2008) and

Frattini(2008) immigration shocks have significant however quantitatively limited effects on

prices, and that those effects are different for services and tradable goods. The question about

the influence of low-skilled immigration waves to the price dynamics in the USA was first

taken up in Cortes(2008). The empirical results state that the increase of immigrant to native

ratio by 1 per cent leads up to 0.2 per cent decrease in prices of services. These results are

confirmed also in Frattini (2008) where changes in price dynamic as a result of immigration

shocks in the UK from 1996 till 2006 are considered. The empirical study states that

immigration flow had dual effect on prices in the UK during the considered period. On the

one hand, immigration contributed to the reduction of price growth of services in sectors

where the concentration of low-wage workers is high. The effect is stronger when prices of

such services as restaurants, bars, and take-away food are considered. The inflow of relatively

cheap labor force lead to reduction in production costs of these services during the considered

period. Plus, the inflow of immigrants could increase competition in the sectors providing

these services; very often immigrants run bars or small restraints. In other words, the

observed negative effect on the prices of such services is probably achieved through labor

supply channel. On the other hand, the opposite effect is documented once the prices of low-

value grocery goods are considered. The inflow of immigrants could cause increase in

demand of such goods, which later could be translated into changes in their prices. Hence, in

this case prices were probably influenced through demand channel.

6


Some opposite effect is documented in the empirical work by Lach (2007), who

shows reduction of grocery prices as a result of immigration shock. It is necessary to mention

that this paper examines the behavior of prices following the unexpected arrival of a large

number of immigrants from the former Soviet Union to Israel during 1990. Once the size of

native population, city and time effects are controlled the estimations show that a 1

percentage point increase in the ratio of immigrants to natives in a city decreases prices by

0.5 percentage point on average. However, the documented negative effect can be explained

by the fact that Former Soviet Union immigrants had higher price elasticity and lower search

costs than the native population. Actually, most of them were not active in the labor market.

The market economy reality where they suddenly appeared could motivate the grocery

shopkeeper to attract new potential customers by temporary price decrease.

As it is mentioned above, the changes in prices due to immigration shocks cannot be

considered as simple change in scale. The response of prices is not homogenous across

various goods and services. Such a change in the distribution of prices may cause

distributional consequences of real income in the country of destination. For example,

according to the estimation presented in Cortes(2008) the low-skilled immigration wave

during the period from 1980 to 2000 had the following effects. It increased the purchasing

power of high-skilled workers living in the 30 largest cities of the USA by in average of 0.32

per cent and decreased the purchasing power of native high school dropouts by a maximum

of 1 per cent and by 4.2 per cent of Hispanic low-skilled natives. Frattini (2008) reports

similar evidence. The author states that the types of items that experience the highest price

reductions (food and drinks out of home, dry cleaning, hairdressing) are those that tend to be

relatively less consumed by low-income households. At the same time, the share of

expenditure for food and drinks which are consumed at home and for which positive price

effects is found, tends to be inversely proportional to the household's income. Coupled with

the distributional effects on wages for the same period of time highlighted by Dustmann et

al.(2008), it can be stated that there indeed was income distributional effect in the UK due to

recent immigration shocks.

2.2 Papers on housing prices and rents

The influence that immigration shocks have on housing prices and rents can be

considered as a particular case of immigrants’ influence on prices in general. However,

7


relying only on the existing economic literature it is quite complicated to find common

judgment upon the influence of immigration shocks on the housing market in the countries of

destination.

There are number of reasons that make this task non-trivial. First, the existing

economic literature on responses of housing markets to immigration shocks is scarce.

Research on the determinants of the price of low quality housing was focused mostly on the

effects of zoning and land use regulation (Malpezzi&Green (1996)) or profitability of

constructing low quality housing (Ohls (1975)) Second, most of scientific works are case

studies, so it is not easy to draw common conclusion. Finally, the existing scientific works

examine mostly the housing markets of countries considered as traditional destinations for

immigrants, such as the USA, Canada, New Zealand and Australia. The question becomes

even more complex once the task is the application of the existing results to the countries for

which inflow of immigrants is just recent experience; to Italy in this particular case.

The first attempts to document the response of American housing market to

immigration shocks were made in the 80s. For example, Muller&Espenshade (1985),

Burnley, Murphy&Fagan(1997) as well as Ley&Tuchener (1999) find strong between inflow

of immigrants and housing prices. However, these early studies have rather descriptive

character. The first attempt to measure empirically the influence of immigrants on the US

housing market was made by Susin(2001) and Saiz(2003), which consider the changes in

rental prices in Miami after the Mariel boatlift, when the exogenous immigration shock added

extra 9 per cent to Miami’s renter population during 1980. Saiz (2003) found that the

unexpected immigration shock led to an increase in rents in Miami from 8 to 11 per cent

more than in the comparison groups between 1979 and 1981. By 1983, the rent differential

was still significantly positive. The change in rents was mainly for dwellings occupied by

low-income Hispanic residents, probably because of the tendency of immigrants to settle

initially in the districts populated by Hispanic residents. The paper states that the rental price

for units of higher quality was not affected by the immigration shock. For the same period,

the relative housing prices moved in the opposite direction. Despite the relative increase of in

rents in Miami, the immigration flow did not alter the rent to income ratio or so called “rent

burden”2 in Miami (Gleulich, Quigley and Raphael,2005). The effect is robust with respect

to rental units less probable to be occupied by immigrants. A similar conclusion is drown in

2 Rent burden is rent to income ratio. (To give more detailed explanation to seem more serioud )

8


Card(2007). Estimating the influence of immigrants on the US cities the author finds that the

magnitude of effect on average wages he estimates is very similar to the one found by

Saiz(2006) for housing market. Hence, the so called “rent burden” remains roughly constants

“Mariel boatlift” case described in Saiz (2003) and Susin (2001) is very special (in particular

point of time in particular city) and could hardly be generalized. However, this study together

with the previously mentioned ones suggests that the labor market is not the only area where

the consequences of immigration shocks should be looked for.

An attempt to get more general picture upon the impact of immigration shocks on

housing rents in American cities can be found in papers that are more recent. For example,

Saiz(2006) demonstrates that there is positive association between rents growth and

immigration inflows for all metropolitan areas; 1 per cent immigrants` inflow to a city

population leads to 1 per cent increase in average rents and housing values. These results

confirm the initial expectations of the author about the magnitude of effect on housing

market. Indeed, the order of magnitude of housing market responses to the immigration

shocks is much bigger than the ones observed in the labor market, which can at least partially

explain the fact that wages do not respond strongly to immigration shocks.

The further extension of the research topic was made through consideration of labor

and housing market simultaneously. Ottaviano&Peri(2007) use a general equilibrium

approach to evaluate the effects of immigration flows on skill-segmented labor and housing

market in the USA. According to the model developed by the authors, the inflow of

immigrants is associated with long run higher average wages and higher average rents. The

rental price of units occupied by high-educated residents is much more sensitive to

immigration shocks compared with those occupied by low-educated ones. As regards to

wages, the model predicts the largest positive effect for the most educated ones and small

negative effect on least educated ones. Once the results generated by the model are compared

with the real data, the following predictions were confirmed First, due to the complimentarity

between natives and immigrants, the overall production effect is positive on natives. Second,

immigrants that are more educated increase the competition for the housing in the best areas

and make the prices to increase from 0.6 up to 2.3 per cent. This finding contradicts to the

results of Saiz (2003) who estimates the strongest effect to low quality dwellings occupied by

Hispanic immigrants. Third, even the natives in the lowest skill group have higher house

ownership rate hence each group of education receive positive transfer from immigrants.

9


Somewhat different results are documented for New Zealand. Stillman&Mare (2008)

empirically estimates the response of housing market to immigration shocks in New Zealand

during period from 1986 till 2006. New Zealand along with the USA, Canada and Australia is

another country traditionally considered as one of the main destinations for immigrants.

Hence, one can expect results similar to ones stated in the papers considering American

housing market responses (for example, Saiz(2006), Ottaviano&Peri(2006)). The estimation

results demonstrate that 1 per cent increase of population in the area is associated with

increase in local housing prices from 0.2 to 0.5 per cent. However, the authors find no

evidence of positive relationship between inflow of foreign-born immigrants to an area and

local housing prices. The only strong positive relationship found is the one between inflows

of New Zealanders previously living abroad into an area and local housing prices, which

however is not robust over time.

With the exception of the work by Gonzales&Ortega (2009), to my best knowledge

there has been no economic research on the link between international immigration and the

price of urban housing in European countries. Gonzales&Ortega (2009) presents empirically

estimated effect of immigration on housing prices and residential construction activity in

Spain over the period from 1998 to 2008. During the mentioned period, Spain was

experiencing both a spectacular immigration flow3 and an impressive housing market boom.

The authors find sizeable causal effect of immigration shocks on both quantities and prices in

the housing market. The inflow of immigrants increased housing prices by about 52 per cent

and caused growth in construction of new housing units equal to 37 per cent; it accounts for

one third of the housing boom in terms of both new construction and prices.

As it can be seen from the reviewed literature the prevailing part of papers are focused

on countries that used to be traditional destination for immigration flows4. In other words,

immigration has always been serving as one of the sources for extension of their labor force.

However, there is no common agreement upon either the existence, magnitude or the

direction of the effect of immigration shocks on housing market.

3 The average Spanish province had an immigrant inflow equal to 17 per cent of its initial working-age population 4 See for example, Saiz(2003), Saiz(2006), Ottaviano& Peri(2007), Card (2007) etc.

10


2.3 Displacement of natives by immigrants

The magnitude of the effect immigrants have on the local housing prices depends

heavily on the reaction of natives on the presence of immigrants in the area. Absence of

housing price effects does not imply absence of immigrants’ impact on housing opportunities

of formerly residing population. If there is sizable displacement effects for natives, the cross-

province estimation will not capture the full impact on housing prices. In other words, if

immigrants displace natives, local housing price effects of immigration will be attenuated. If

the displacement is one to one, in case of perfect substitution, then the price effect should

vanish completely. In case of modest displacement, the estimates will provide information

upon lower bounds of the full magnitude of the effects.

The effect of immigration shocks on the migration decision of natives depends on

several factors. There are some questions to answer and the answers are not trivial. For

example, are immigrants attracted to a region by the same characteristics as the natives?

Alternatively, what kind of preferences do natives have? On the one hand, immigrants’

decision about settlement pattern may differ from the native one. Often the key factor for

immigrants is the initial settlement of immigrants from the same country of origin. They are

more probable to be supported by their compatriots (Pedersen&Pytlikova&Smith, 2008). On

the other hand, the utility of natives from immigration shocks depends heavily on their

complimentarity in the labor market. Particularly, if immigrants and natives are compliments

in the production process then the areas where immigrants settle will become more attractive

for natives as well. Exactly the opposite is expected, if natives and immigrants are substitutes

in the production process. In this case, natives can consider the areas densely populated by

immigrants as less attractive. Natives, being scared by possible completion created by

immigrants in the labor market, may try to avoid those areas.

The research on the USA regarding the possibility and extent of displacement of

natives conclude that there is no “one-for-one” offsetting outflow of natives caused by inflow

if immigrants5. To my best knowledge, there is no economic paper considering the

displacement of natives by immigrants in Italy. However, some preliminary picture can be

5 For comprehensive research on possible displacement and size in the USA see, for example, Card and DiNardo (2000), Card (2001, 2005), Federman, Harrington, and Krynski (2006), Ottaviano &Peri (2007), Borjas (2006)

11


drawn based on from the situation in the labor market. If there is competition between natives

and immigrants in the Italian labor market, it can signal about possibility of the displacement

of natives from areas densely populated by immigrants. People usually form their opinion

upon economic situation in the country relying heavily on information they find through

mass media. Recently held social surveys6 show that Italians consider the received large

masses of uneducated workers as negative phenomenon; they depressed wages, worsened

skill-intensity of the economy, hurt native, especially the less educated. However, economic

research does not support this popular opinion. A comprehensive analysis of history of both

Italian emigration and immigration is presented in Venturini&Del Boca(2003). The empirical

analysis (based on repeating cross-section data covering the period from 1989 until 1995)

estimates the influence of immigration shocks on wages and employment of natives by

branches and Italian regions. The share of immigrants flow with respect to natives has

positive impact on the wage growth of natives. The complimentarity effect is even stronger

when the estimation is performed for data referring to Northern regions of Italy, blue collars

or small companies. However, the response of Italian labor market is nor linear. The negative

estimates suggest that once the share of foreign workers reaches 10-14 per cent in regions and

branches, foreigners will turn from compliments to substitutes and will compete with natives

in Italian labor market (Venturini&Del Boca,2003). Similar results are documented in

Venturini&Villosio (2002); the effect of foreigners’ concentration on the employment

opportunities of natives is discussed. The estimation results state that the probability of

moving from employed to unemployed either decreases once the percentage of foreigners

increases, or is statistically not significant. The effect of percentage of foreigners on the

probability to find new job for natives is estimated as positive. Yet, the probability of finding

job for young people looking for their first job seems slightly negative ( Venturini&Villosio,

2002).

Overall, the estimation based on the data referring to legally present immigrants and

regular Italian labor force shows that immigrants and natives are compliments in the labor

market. However, the extent of empirical research is usually constraint by availability of data.

Most of above presented results are obtained using the number of legally present immigrants

as a measure of foreign presence in Italy. Yet, continuous increase of irregularly present

foreigners who are eager to accept the lower wages as well can stimulate enlargement of 6 See for example 1995 ISSP , National Identity Module or “Demoskopia” survey held by Fondazione Rodolfo DEBENEDETTI(fRDB)in 2003

12


shadow economy and cause flow of capital from legal sector to illegal on. In this way

irregular immigrants can indirectly compete with natives both in irregular labor market and

with regularly employed natives through stimulating the growth of the irregular economy

injuring the regular one(Venturini&Del Boca, 2003). Even if focus is turned to irregular labor

market where immigrants can undermine the legal employment by increasing the scale of

shadow economy, Italian immigrants do not seem be in completion with natives

(Venturini,1999). The general pattern for Italian immigrant workers appears to be a

fragmented career; restricted to seasonal or temporary jobs as well as alternating between

legal and illegal employment (Venturini&Villosio, 2008). Consequently, one can concluded

that there is no direct competition between natives and immigrant workers. The above-

described results, coupled with Italian labor force immobility, allow supposing that at least on

provincial level, there is no displacement of natives cause immigrants through competition in

Italian labor market.

13


3 Methodological approach

3.1 OLS estimation

In this section, I present the empirical approach used to estimate the impact of

immigration on prices in provincial housing markets. As a dependent variable, I consider the

logarithm of prices per square meter of housing units. The main explanatory variable is the

number of legally present immigrants in province in a given period. Housing data comes

from biennial survey and it is available for every second year. The peculiarities of data on

housing prices are such that the data is available for every second year.

To derive the regression model I first present the simple model in levels.

Equation (1)

ln ln ln

Where,

ln(Pit) is the dependent variable: mean log price of residential unit per square meter in

province i and in time period t.

ln(IMMit) is the main independent variable, which is the stock of immigrants in province

measured as the of number of valid residence permits in province i in time period t.

ln(POPit) measures the population in province i in time period t.

Wit is the set of macroeconomic variables such as employment rate or GDP per capita in

province i in time period t. The set of macroeconomic controls is supposed to capture

differences in housing prices due to differences in economic conditions between provinces.

μt is the set of year dummies which captures national trends in inflation and other economic

variables.

φi is the set of province dummies, which captures time-permanent and province-specific

characteristics.

Finally, εit is the error term. The main estimate of interest is β, which captures the effect of concentration of

immigrants on the price of residential units. The model constructed in this way let β capture

the effect of increased immigrant population net of increase in overall population in the

province, while γ will capture the effect of overall change in population.

14


To get rid of all province-specific time-invariant factors, which can influence both the

immigration flows and dynamics of housing markets I take the first difference of the model

described in equation (1). After first-difference transformation, the model takes the following

form

Equation (2)

∆ln ∆ ln ∆ln ∆ ∆ ∆

β coefficient have the following interpretation: the percentage change in housing price

as a result of 1 per cent increase of foreign population in province. It can be the case that the

effect of immigration shocks on housing market dynamics has not linear patters. To capture

it, I propose to compare the estimation results from Equation 2 with the estimation that

include the squared term of number of immigrants7.

Equa )tion (3

∆ln ∆ ln ∆ ln ∆ln ∆ ∆ ∆

3.2 Potential problems

Empirical estimation of impact of immigration shocks on housing market contains a

number of challenges. In this section, I am trying to present these problems and possible

remedies to tackle them.

First, the measure of immigrants’ concentration I use in the estimation is based on the

number of valid residence permits issued by Ministry of Interior Affairs of Italy. Hence, this

measure of foreign presence in Italy refers only to legally present immigrant and says nothing

about illegal immigrants which are not necessarily have the same distribution pattern as legal

ones across Italian provinces. However, illegal immigrants influence the housing market as

well. Measuring immigration stock by using the number of officially present foreigners ,

which does not take into account illegal immigrants, can lead to a bias in the estimates of

immigration concentration influence on housing market outcomes.

15

7 As an alternative, I estimate also using as a dependent variable the concentration of immigrants and its squared term. ∆ln ∆ ∆ ∆ ∆ ∆


To tackle the this problem it is necessary to understand where the illegal immigration

arises from and whether and how it is correlated to legal one in Italy. Immigration is a quite

recent phenomenon for Italy, which started in 1980s. The initial unplanned immigration flow

and some control problems created numerous illegal immigrants in Italy. The first law

regulating the inflow and the presence of foreigners was approved and implemented in 1990,

with future amendments in 1995, 1998 and 20028. The Italian immigration policy is based on

residence permits issued by Ministry of Internal Affairs.

Bianci et al.(2008) using from regularization in 1995, 1998 and 2002 prove

empirically that the combination of logarithm fixed effects may attenuate measurement

errors related to illegal immigration.9 Particularly, if the number of officially present

immigrants is proportional to total immigrants and the constant proportionality is the product

of province and year specific c nt g ue 10 onsta s, then the followin is tr

Wher

and are respectively the logarithm of total (official plus the number of

application for regularization) and official share of immigrants with respect to total

population in Italian province i at time period t. The above-mentioned arguments allow

assuming that using logarithm of ratio of number of permits over population (once time and

province dummies are included) helps to avoid bias created by measurement error due to

illegally present immigrants. For the reasons described above, I introduce an alternative

specification, where instead of using number of immigrants and population of province I use

the concentration of immigrants in province; that is the logarithm of immigrants over

population ratio. The first differenced model of this specification has the following form.

e,

Equation (4)

∆ln ∆ ln ∆ ∆ ∆

8 Regularizations in 1995, 1998, and 2002 in ed 246, 217 and 700 thousand individuals, respectively (Bianci et al., 2008). volv9 The relationship between actual and official immigration once both the province and year fixed effect are taken into account are the following: the OLS estimated coefficient of is 0.92 and the R2 is 99%.. For further details see Bianci et al.(2008). 10 For more detailed derivation see Bianciet al.(2008).

16


Another problem that can undermine the reliability of the estimates is omitted variable

problem. Indeed, the location choice of immigrants can be motivated by unobserved factors

that cause changes in the housing prices as well. Suppose that for some reason some

provinces became more attractive (for example, expectation of future improvement of

economic conditions or amenities). It will lead to more intensive flow of immigrants and

natives; hence to higher housing prices. In this case, the omitted variable would lead to

overestimation of immigration impact on housing prices.

Endogeneity creates additional problems as well. Indeed immigration flow can be

endogenous to housing prices. On the one hand, immigrants may tend to avoid the regions

where, given the similar employment rate and GDP, housing prices are higher. In this case,

the OLS estimation would lead to downward bias. Immigrants might also be heavily

influenced by cheap available housing in depressed areas or may be especially attracted to or

attractive for declining industries (Filer, 1992). On the other hand, the inflow of immigrants

may make natives avoid area or move out because of the competition in the local labor

market. The possibility of native displacement by immigrants is intensively discussed in

Filer(1992), Card(2001), Card&DiNardo(2000) etc. In fact, if immigrants cause “one-to-one”

offsetting outflow of natives there will be no shift in local housing demand, while positive

effect will suggest that even if there is any displacement it is does not have “one-to-one”

nature. It is necessary to mention that immigrants may have different preferences in housing

market. They might evaluate network of their compatriots or other amenities as more

important factors than the housing prices. Immigration shock can lead to a housing demand

shock and as result outflow of natives that are more sensitive to increase in housing rents or

prices. Hence, the outflows of native due to fear to face competition in the labor market can

weaker the effect of immigrants on housing prices; part of the effect would take place

through native displacement. As a result, the OLS estimate will tend to underestimate the

effect of immigration shocks.

The discussion above suggests that the sign of bias depend on many factors and is not

easy to predict; obtained results should be interpreted with caution.

17


3.2 Instrumental variable approach

Instrumental variable approach can serve as a plausible strategy to solve the

endogeneity problem described in the previous section. I use the historical settlement pattern

of immigrants to instrument the current inflow of immigrants in Italian provinces. The current

geographical distribution of immigrants across provinces is supposed to be correlated with

the historical one and uncorrelated with the current province specific housing market shocks.

It is well stated in the economic literature that a number of noneconomic factors

determine the decision upon the destination for international immigrants. The prior existence

of enclaves of immigrants from a country is an important magnet for future flow from a

country. It turned out that immigrants rely on social and ethnic factors and do not follow

simple utility maximizing approach based only on the economic conditions of the destination.

Indeed, it is reasonable to expect that network effect plays significant role in for immigrants

in destination decision making process; new immigrants will tend to settle in the areas

relatively more densely populated by people from the same country of origin to be able to

benefit from the support of their compatriots. Economic literature on migration provides rich

evidence of this prediction. The possibility to live among people speaking the same language,

having similar cultural traditions makes particular regions more attractive for new comers

(Pedersen et al.(2008), Carrington et al.(1996)). These arguments frequently motivated

economists to construct instrumental variables for actual immigration flows based on the

historical information (see, for example, Saiz (2007), Cortes (2008), Ottaviano and

Peri(2007), Card (2000), McKenzie and Rapoport, 2007).

The first instrument is bases on the overall changes in immigration flows from one

period to another and the initial distribution of immigrants across Italian provinces. The

instrument is based on two assumptions. First, the initial distribution of immigrants in t=0 is

not correlated with the omitted variables which can have influence on dynamics of housing

market in the future. More explicitly, this assumption imply that based on available

information set immigrants cannot forecast the future of housing market in any province i

better than natives. Housing price is the capitalized discounted value of the stream of future

rents. If one believes that immigrants are able to pick the future “winner” locations based on

the set of information available in the present then one has to explain why natives did not

capitalize on the same set available information (Saiz,2006). Second, changes in immigrants’

18


flows on national level are exogenous with respect changes in the province–specific

amenities. The instrument is calculated according to the following formula

, , · ∆ ,

The predicted number of immigrants in province i period t is calculated based on the

initial distribution of immigrants across provinces in year t=0 and the total number of permits

of stay issued during year t. θi,t=0 is the share of immigrants who settled in province i in t=0.

I take year 1990 as t=0 since it is the first year for which the data on residence permits is

available on provincial disaggregation level. This estimate of immigrants’ presence is free

from province and time specific shocks.

The second instrument relies not only on the historical settlement of overall

immigration stock but also on its composition based on the country of origin. Here I rely on

assumption that country of origin-province initial distribution is not correlated with the

demand shocks which the provinces face in the later periods.

, , , · ∆ , ,

W

, , , is the fraction of immigrants from country or area c which settled in province i

in the period t=0. , , is the number of immigrants from country or are c that lives in

Italy at time period t.

here

To construct this instrument I used number of resident permits issued. The

information about the country of origin of immigrants is available on province disaggregation

level. First, I took thirty-seven origin countries separately. The rest of countries were grouped

based on geographic criteria. The resulting nine geographic groups are the following: Central

Europe, Other Europe, Former Soviet Union, Asia, Northern Africa, Southern Africa,

Southern America, Central America, Australia and Oceania. More detailed description can be

found in Table 2 of Appendix.

The result of the univariate regression confirms that the instrument fits the actual

changes of immigrant population.

19


∆ 1410,257 0,671∆

The coefficient at the instrument is significant at 1 percent level. The F-statistic meets

the requirements and is equal to 855.74. The result is robust to inclusion of time dummies as

well.

20


4. Data description

The data used for estimation come from different sources. The descriptive statistic of

the data used in estimation is presented in Table1, Appendix. In this section, I present in

details the peculiarities of information used in estimation.

4.1 Immigration flow

The information source for immigration flows comes from ISTAT11 , which in

collaboration with Ministry of Internal Affairs develops and delivers data on foreign nationals

legally present in Italy since early 90s. The Ministry of Internal Affairs provides the initial

information; that is the number of existing valid residence permits on January 1 of each year.

The estimated number of permits is based on the information taken at least 6 month after the

reference date. This allows first, to take into account those foreigners whose permit of stay

was expired by January 1, but who had applied for renewal, hence where legally present in

Italy. Second, in addition to residence permits valid on January 1 of a given year, it allows to

include also those foreigners who though were present legally in Italy, but without residence

permit due to long time required to for completion of the practice of the first release.

The statistical data are available from 1992 with yearly frequency with detailed

description of demographic characteristics: gender, age, marital status, country of origin and

reason for presence in Italy. The territorial units are not necessarily limited to the provincial

level since the residence permits are issued by State Police at the Police Headquarters.

However, the inconsistency of number of Italian province through time requires some

work to be done. Particularly, due to the creation of four new provinces (Olbia-Tempo,

Ogliastra, Medio Campidano and Carbonia-Iglesias) in region Sardinia the number of Italian

provinces grew from 103 to 107. At the same time, the housing market data are available in

“103 provinces” format. Hence, to be consistent with geographic units considered in housing

market data I adjusted the data on immigrants to “103 provinces” as well. I added the number

of immigrants reported for new provinces to the provinces which they geographically

belonged before separation. The number reported for Olbia-Tempo was added the one

reported to Sassari and reported as Sassari, Ogliastra was added to Nuoro and reported as

11 ISTAT-“The Italian National Institute of Statistics is a public research organization. It has been present in Italy since 1926, and is the main producer of official statistics in the service of citizens and policy-makers. It operates in complete independence and continuous interaction with the academic and scientific communities.”

21


Nuoro , Medio Campidano together with Carbonia-Iglesias were added to Cagliari and

reported as Cagliari.

4.2 Population

The data on population in Italian provinces comes from ISTAT as well. Particularly,

the Demographic balance of yearly resident population provides the results of the monthly

data collection called “Movement and calculation of resident population”12 It is implemented

by ISTAT in collaboration with the Population Register offices (anagrafi) of the Italian

municipalities (Comuni). The information is available from 1992. It provides with data on

population on January 1 of each year. Resident population encompass Italian and foreign

citizens usually living on the national territory, even if temporarily absent.

Each person having usual residence has to register, by law, in the population register

of the municipality where usually lives. The legal population is being determined on the base

of the Population Census. I adjusted the population data for Italian provinces to “103

provinces” format using the same above described methodology used for data for immigrants.

4.3 Housing price

The data on housing value come from the Italian Survey on Household Income and

Wealth (SHIW13). The sample used in the most recent surveys comprises about 8,000

households (24,000 individuals), distributed over about 300 Italian municipalities. The

variable used is the average value of housing value per square meter. It is calculated based on

the answer to the following question asked during the interview:

“In your opinion, what price could you ask for the dwelling in which you live (unoccupied).

In other words, how much is it worth (including any cellar, garage or attic)? Please give

your best estimate”.

It is worth to mention the disadvantages of the data. First, the value of housing is self-

reported by owner of the house or person who occupies it. It is not always the case that

person is aware about current market price of the dwelling. Second, the number of 12 form ISTAT P.2 13 SHIW - began in the 1960s with the aim of gathering data on the incomes and savings of Italian households. Over the years, the scope of the survey has grown and now includes wealth and other aspects of households' economic and financial behavior.

22


observations is around eight thousands, while the number of Italian provinces is equal to 103.

Hence, the number of observation is around 80 per province14, which cannot give very

precise estimate of current market price.

4.5 Gross domestic product

The data on Gross Domestic Product come from Regional Statistics provided by

EUROSTAT15 Statistical office of European Communities. Data are available from 1995

with annual periodicity. The data are calculated by EUROSTAT based on data from

European System of Accounts ESA 1995 initially sent by National Statistical Institute. Italian

provinces correspond to NUTS 3 level regional breakdown. Particularly, the provincial GDP

per capita in Euros (EUR_HAB) and in Purchasing Power Standards are used (PPS_HAB).

The data covers 107 Italian provinces and because of already mentioned reasons, the

data was adjusted to “103 provinces” format. Actually, the only problem is related to

Sardinia. The data on NUTS3 level are available only from 2001 and in “107 format”. To

bring the data to “103 format” I performed several steps. First, to obtain GDP data of “old”

provinces I weighted them by population of “new” ones. It was done for years from 2001 to

2006. Second, I tried to fill the missing values of GDP Sardinian provinces from 1995 till

1999 in the following way. I calculated the average ratio of GDP of Sardinian province i to

regional GDP for period when data are available (from 2001 till 2006). The obtained ratio

was used to calculate the GDP for four “old” provinces for years from 1995 till 1999.

Particularly, to obtain GDP estimate for year t the above-mentioned ratio was multiplied by

the regional data of year t.

4.6 Unemployment rate

The data on unemployment come from EUROSTAT. It is based on LFS (Labor Force

Survey, “age 25 and over” is considered) quarterly household sample survey and is available

on NUTS3 geographical disaggregation level. The first data set is Regional labor market data

based on pre-2003methodology (LFS adjusted series) and covers period from 1995 to 2001.

The second data set comes from Regional unemployment LFS series and covers period from

14 The number of observation varies from five to more than five hundreds. 15 Eurostat is the statistical office of the European Union situated in Luxembourg. Its task is to provide the European Union with statistics at European level that enable comparisons between countries and regions.

23


1999 to 2008.The data is given in two datasets because of methodological change made at

some point of time. Because of changes made in definitions there is some discrepancy once

the overlapping period from 1999 to 2001 compared, which points on the fact that these two

datasets are not comparable in the raw way. To fix this problem I took the data from Regional

unemployment LFS series which covers period from 1999 to 2008 as a benchmark and

adjusted the remaining period (from 1995 to 1998) to the recent methodology. I used the data

for overlapping years to calculate the average ratio of “new” to “old” unemployment rate for

every province. Then the “old” values (for period beginning from 1995 to 1998) were

multiplies by already calculated average ratio to obtain “new” values for the mentioned

period.

Beginning from 2001 the data for provincial unemployment is presented in “107

provinces” format. I adjusted the unemployment data to “103 provinces” format as well. To

bring the data to “103 format” I used the population of “new” ones as weights. In the next

step, I filled the missing values of Unemployment for Sardinian provinces from 1995 to 1999

in the following way. I have unemployment rate for all eight Sardinian provinces for year

2008 only and regional data for all years. I brought the 2008 data to “103 format” and

calculated the ratio of provincial to regional data for all “old” provinces. After that, I

calculated the estimated provincial unemployment rate for Sardinian provinces multiplying

the above-mentioned ratio by the regional data of corresponding year.

24


5 Results OLS estimation results

In this section, I present estimation results. I begin from ones based on OLS

estimations.

Table A reports estimation results based on model described in Equation (2). Column

(1) shows that with no additional controls growth in immigrant population is positively and

statistically significantly correlated with the growth in housing prices. This makes sense,

because increase in population due to the inflow of immigrants leads to an increase in

demand for housing units, which in its turn pushes housing prices up. The population growth

control is essential in this case, because it captures the effect of larger population, which can

be result of the inflow of both natives and immigrants. Neglecting of control for total

population growth would lead to certainly positive coefficient at the number of immigrants,

as immigrants’ inflow results in larger population, therefore higher demand for housing units.

Column (2) reports the results of basic specification with the set of time dummies that are

supposed to capture national trends in inflation and other economic variables. Although

adding time dummies decreases the magnitude of β and makes it statically insignificant, it

notably improves the explanatory power of the model; R2 increases from 0.03 to 0.60.

Column(3) demonstrates results once the estimation includes controls for changes in

economic conditions at provincial level; i.e. unemployment rate and GDP per capita. These

macroeconomic controls are supposed to capture differences in housing prices due to

differences in economic conditions between provinces. The coefficient at immigration is

equal 0.058, which is significant only at 10 per cent level. The results are robust to inclusion

of time-part dummies, which allows controlling for differences in business cycles across

Italian geographic areas. Both Column (3) and Column (5) report very similar coefficient at

the number of immigrants. The results suggest that 1 percent increase in number of

immigrants increases housing prices in province by about 0.06 percent, which indicates quite

modest effect of immigrants on average housing prices on provincial level. While interpreting

the results, it is necessary to keep in mind some facts. The first one is the possibility of

existence of previously discussed displacement effect. In the presence of displacement, the

estimation results give only a lower bound of the full magnitude of the effect. Second, it can

be the case the effect of immigration shocks on housing market dynamics has non-linear

25


patters. To capture non-linearity I repeat the estimations of Column (3) and Column (5)

including squared term logarithm of number of immigrants.(Can I do that ??? Bo). The

results are reported in Column (4) and Column (6) respectively. Indeed, the estimation

confirms the initial suspect about non-linear pattern. The coefficient at immigration increases

drastically and become statistically more significant. For example, the coefficient in Column

(4) increased to 0.394 and became statistically significant at 1 per cent level compared to

0.058 significant at only 10 percent significance level in Column (3). In column (4) and (6),

the coefficients at squared term of number of immigrants are negative and statistically

significant at 5 and 10 per cent respectively. This might mean that initial model without

inclusion of squared term of log-change of number of immigrants overlooked some

potentially important nonlinearities. The number of immigrants no longer has a positive

effect on housing prices of province: the relationship between log-change of housing prices

and log-change of number of immigrants is positive up until log number of immigrants is

equal 8.2. This value corresponds to 3640 immigrants in a province. All estimations reported

in columns (3)-(6) include control for changes in density of population in provinces.

I performed estimations using alternative measure of concentration of foreign

population; that is the ratio of number of immigrants to population in province i at time t. The

results are presented in Table B. Column (1) presents the results of simple first-difference

estimation without any additional controls. As in the previous case, the coefficients at

immigrants’ concentration are positive and statistically significant. Obviously, without other

controls the explanatory power of the model is very low. Adding time-dummies and

macroeconomic controls (see column (2) and (3) respectively) significantly increase the

explanatory power of model however turns the coefficients at foreign concentration to

negative and statistically insignificant. Column (5) shows that addition of time-part dummies

leaves coefficient of main interest negative and statistically insignificant. However, following

the same reasoning as in the previous section I add the squared-term of immigrants`

concentration. Column (4) and (6) show that also in this case the non-linearity of housing

price response to immigrant’s inflow is confirmed. The coefficient of main interest turns to

positive and statistically significant at 10 per cent level. Even more, the coefficient at squared

term of immigrants` concentration is negative and statistically highly significant. From the

estimated coefficients it can be concluded that though concentration of immigrants leads to

appreciation of housing prices the effect turns to negative once concentration reaches some

26


critical level. The critical value of immigrants’ concentration after which the relationship

turns to negative one is estimates close to 3 percent.

As it was already mentioned before the measure of foreign presence I use is based on

number of residence permits issued by Ministry of Internal Affairs, which accounts only

legally present immigrants. The possible solution to correct mismeasurement was proposed in

Bianci(2008). Following the methodology presented in this paper, I estimated the model

described by in Equation (4). The results are presented in Table C. As in two previous cases

the simple first–difference estimation shows positive association between increase of foreign

concentration growth of housing prices. The addition of time dummies significantly increases

the explanatory power of model, but turns the coefficient of main interest to statistically

insignificant. Inclusion of both macro controls and time-part dummies leaves the magnitude

of β almost unaltered (see Column (3) and (5) respectively). The coefficient at immigrants

concentration is similar in its magnitude to the once presented in Table A; i.e. that 1 percent

increase in number of immigrants increases housing prices in province by about 0.06 percent.

However, the coefficient remains significant only at 10 percent level.

Instrumental variable estimation result (To be done after discussion with prof. Sembenelli)

6 Conclusion

A large body of literature analyzes the impact of immigration on the employment

opportunities of native population. With few exceptions, this literature addresses the issues

related to the labor market outcome or cost or benefits imposed on native taxpayers because

of immigrants’ inflow. If the purpose of economic studies is investigation of economic

processes for precise policy design, then the final judgment can be made only after careful

consideration of a wide range of phenomenon related results. An enriched picture of

immigrants influence on the well-being of natives can be obtained if along with the effect

immigrants have on production process, through altering composition of labor force, the

impact on local prices through consumption process is taken into account as well.

Consideration of price effect can enhance the influence on real income and real wealth of

population.

27


Housing market through changes in rents and prices may present one of the main non-

labor channel by which immigrants can influence the well being of natives. However, the

attempts to estimate the influence of immigrants on housing market outcome were made

mostly for the USA. With the exception of Gonzales&Ortega(2009), the influence of

immigrants on European housing markets remains unexplored.

In this paper, I estimate the estimate the impact of immigration flows on Italian

housing market from 199. The paper contributes to the immigration literature in the following

ways. First, it enhances the image of influence of recent intensive immigration flow to Italian

economy by estimating the impact of immigration to Italian housing market. Italian housing

market has never been considered in connection to immigration flows Second, the fact that

estimation is performed on the subject of European housing market makes it remarkable in a

wider context; i.e. it gives chance to estimate the influence of immigration on housing market

in European region in the future. Third, it contributes to recently emerging branch of

literature on influence of immigration on price in general.

The OLS estimation results show that immigration has positive, however, declining

effect on the growth of housing prices in Italian provinces. The estimated results suggest that,

ceteris paribus, as the growth of immigrants concentration in province reaches approximately

3 percent, the further increase of it leads to decrease in the rate of housing price appreciation.

The instrumental variable estimations in all specifications show that immigration presence in

provinces has positive effect on the growth of housing prices. However, the results are

statistically insignificant.

To add some discussion here

28


References

Bertoli, S., Bruecker, H., Facchini, G., Mayda, A. M. and G. Peri (2009). “The battle for brains: How to attract talents.” Report presented at the Fondazione Rodolfo de Benedetti XI European Conference on Brain drain and brain gain, Pisa.

Bianchi, M., Bounanno, M., and P. Pinotti (2008). “Immigration and Crime: an Empirical Analysis”. Temi di discussione, Banca d'Italia , No. 698.

Borjas, G.J. (1994),“The Economics of Immigration” Journal of Economic Literature, 32 (4), 1667-1717.

Borjas, G.J. (1995),“The Economic Benefits from Immigration.” Journal of Economics Perspectives, 9 (2), 3-22.

Borjas, G.J. (2003), “The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market” The Quarterly Journal of Economics Vol. 118(4), 1335-1374

Burnley, I., Murphy, P.,and R. Fagan (1997), “Immigration and Australian Cities”, The Federation Press .

Brandolini, A., Cannari, L., D’Alessio, G. and I. Faiella (2004), “Household Wealth Distribution in Italy”, Banca d’Italia, Temi di discussione, No. 530.

Carrington, W.J., Detriagiache, E. and T. Vishwanath (1996), “Migration with Endogenous Moving Costs” American Economic Review, 86(4), 909-930.

Card, D. (2001), “Immigrant Inflows, Native Outflows, and the Local Labor Market Impacts of Higher Immigration”, Journal of Labor Economics 19(1), 22-64.

Card, D. (1990, “The Impact of the Mariel Boatlift on the Miami Labor Market”, Industrial and Labor Relations Review , 43 (2), 245–257.

Cortes, P. (2008 ), “The Effect of Low-skilled Immigration on U.S. Prices: Evidence from CPI Data. Journal of Political Economy , 381-422.

Del Boca, D. and A.Venturini, (2003). “Italian Migration”, IZA Discussion Paper, No. 938

Levin, A. and C.F. Lin, “Unit root test in panel data: Asymptotic and finite sample properties.” University of California at San Diego Discussion paper 92-93.

Di Addario, S. (2002), “Italian Household Tenure Choices and Housing Demand”, Banca d’Italia, mimeo.

D. Ley, J. Tuchener. (1999), “Immigration and metropolitan house prices in Canada”. Research on immigration and integration.Working paper, No.99-09.

McKenzie, D. and H. Rapoport (2007), “Self-selection patterns in Mexico-U.S. migration: The role of migration networks” CReAM Discussion Paper No. 0701.

29

http://www.jstor.org/stable/25053941

http://www.jstor.org/stable/25053941

http://www.jstor.org/action/showPublication?journalCode=quarjecon


Gonzalez, L.and F.Ortega (2009), “Immigration and Housing Booms: Evidence from Spain”, Centre for Research and Analysis of Migration. , CReAM Discussion Paper No 19/09.

Nannestad.P. (2007), “Immigration and welfare states: A survey of 15 years of research”, European Journal of Political Economy 23 , 512-532.

Ottaviano,G. and G. Peri (2006), “Rethinking the Effect of Immigration on Wages” NBER Working Paper Series , Working Paper 12497.

Ottaviano, G. I. and Peri, G. (2007), “The Effects of Immigration on U.S. Wages and Rents: A General Equilibrium Approach”, CReAM Discussion Paper No 13/07 .

Saiz, A. (2003), “Room in the kitchen for the melting pot: Immigration and rental prices”. Review of Economics and Statistics , 85 (3), 502–521.

Saiz, A. (2007). Immigration and housing rents in American cities. Journal of Urban Economics , 61 , 345–371.

Muller, T and T.J. Espenshade. (1985), “The Fourth Wave: California’s Newest Immigrants”. The Urban Institute Press .

Dustmann, C., Glitz, A., and T.Frattini, (2008), “The labour market impact of immigration” Oxford Review of Economic Policy 24(3), 477-494

Cannari, L., Faiella, I., Grazia, M. and A. Neri (2008) “The Real Assets of Italian Households”,Household wealth in Italy. Papers presented at the conference held in Perugia, 16-17 October 2007

Card, D. (2001), “Immigrant Inflows, Native Outflows, and the Local Labor Market Impacts of Higher Immigration” Journal of Labor Economics, 19(1)

Card, D. (1990), “The Impact of the Mariel Boatlift on the Miami Labor Market” Industrial and Labor Relations Review. 43 (October): 245–57.

Card, David (2005). “Is the New Immigration Really So Bad?” Economic Journal 115 (November)300–323.

Card, D. and J.E. Di Nardo (2000),“ Do immigrants inflow leads to native outflows?” National Bureau of Economic Research, Working Paper 7578.

Chaloff, J. and G. Lemaitre (2009), “Managing highly-skilled labour migration: a comparative analysis of migration policies and challenges in OECD countries” OECD social , employment and migration working papers , No.79

Faiella, I. and A. Neri (2004) “La ricchezza delle famiglie italiane e americane”, Banca d’Italia, Temi di discussione, No. 501

Federman, M.N., Harrington, D.E. and K.J. Krynski (2006) “Vietnamese Manicurists: Displacing Natives or Finding New Nails to Polish?” Industrial. And Labor Relations Review. 59 (January), 302–18.

30

http://oxrep.oxfordjournals.org/search?author1=Christian+Dustmann&sortspec=date&submit=Submit

http://oxrep.oxfordjournals.org/search?author1=Albrecht+Glitz&sortspec=date&submit=Submit

http://oxrep.oxfordjournals.org/

http://oxrep.oxfordjournals.org/content/24/3.toc


Filer, R.K.(1992), “Immigrant Arrivals and the Migratory Patterns of Native Workers.” in George J. Borjas and Richard B. Freeman (Eds.), Immigration and the Workforce, Chicago: The University of Chicago Press.

Lewis, E. (2004), “How did the Miami labor market absorb the Mariel immigrants?”, Federal Reserve, Bank of Philadelphia, Working paper 04-3

Manacorda, M., Manning, A. and J. Wadsworth (2006), “The Impact of Immigration on the Structure of Male Wages: Theory and Evidence from Britain”, Manuscript, London School of Economics,.

Malpezzi, S.(1996),“Housing Prices, Externalities, and Regulation in U.S. Metropolitan Areas”, Journal of Housing Research 7 (2) 209–41.

Nannestad, P.(2007), “Immigration and welfare states: A survey of 15 years of research”, European Journal of Political Economy 23 , 512–532

Ohls, J.(1975) “Public Policy towards Low Income Housing and Filtering in Housing Markets”, Journal of Urban Economics 2 (April) 144–71.

Pedersen J.P., Pytlikova, M. and N. Smith (2008),“Selection and network effects—Migration flows into OECD countries 1990–2000”, European Economic Review 52 , 1160–1186.

Paiella, M. (2001), “Demand for Housing, Saving Behavior and Wealth Allocation”, Banca d’Italia.

Poot, J. and W. Cochrane (2005), “Measuring the economic impact of immigration: A scoping paper”, Hamilton, New Zealand: University of Waikato, Population Studies Centre. Population Studies Centre Discussion Paper 48

Susin, S. (2001), “The Impact of the Mariel Boatlift on the Miami Housing Market,” U. S. Census Bureau.

Stillman, S. and C. Maré David (2008), “Housing Markets and Migration: Evidence from New Zealand”, Motu Economic and Public Policy Research, Working Paper 08-06

Venturini A. and C.Villosio (2008), “Labour-market assimilation of foreign workers in Italy”, Oxford Review of Economic Policy, Volume 24, Number 3,517–541

31


Appendix A

Table1 Descriptive statistics Variable Population weighted Non-weighted Min Max Mean Std.dev. Mean Std.dev.

Housing value (euro per sq. m) 1996 1800.12 654.27 1666.9 670.07 640.99 4918.5 2007 2039.96 702.8 1855.29 646.78 823.46 4539.81 Number of residence permits issued 1996 22,127 38,313 7,079 16,602 54 142,780 2007 57,945 77,454 23,446 36,681 942 257,779

Population 1,227,823 1,194,315 558,225 611,628 89,043 4,013,057

Immigrants concentration (permits/population) 1996 0.0128 0.0097 0.0104 0.0077 0.0003 0.0434 2007 0.0408 0.0231 0.0389 0.0222 0.0054 0.0981

GDP per capita 21,628 6,650 20,560 5,504 8500 37300

Unemployment rate (%) 8.04 5.8 7.48 5.52 0.7 29

Province area (sq.km) 3296 1769 2845 1600 212 7400 7400 Notes: All variables are defined at provincial level. All variables are defined at the annual level. Province area is time invariant.

32


Figure1 0

1020

3040

Freq

uenc

y

0 50000 100000 150000Number of immigrants

Distribution of immigrants across provinces,1996

05

1015

2025

Freq

uenc

y

0 50000 100000 150000 200000 250000Number of immigrants

Distribution of immigrants across provinces,2007

05

1015

20Fr

eque

ncy

0 .01 .02 .03 .04Immigrants/population ratio

Share of immigrants in total population across province,1996

05

10Fr

eque

ncy

0 .02 .04 .06 .08 .1Immigrants/population ratio

Share of immigrants in total population of province,2007

Source: ISTAT

This figure presents the evolution of distribution of immigrants across Italian provinces during period from 1996 to 2007. The number of immigrants is equal to number of valid residence permits issued by Ministry of Internal Affairs in provinces at the beginning of calendar year. The share of immigrants in total population in provinces is measured as ratio of number of valid residence permits over total population at the beginning of calendar year

33


Figure 2. Concentration of immigrants vs. Housing prices

010

0020

0030

0040

0050

00H

ousi

ng p

rice

of a

squ

are

met

er

0 .02 .04 .06 .08 .1Immigrants/population ratio

Fitted values Housing Price of a square meter

Note: the unit of observation is an Italian province

Correlation between the Concentration of Immigrants and Housing Prices

This graph presents correlation between the concentration of immigrants and housing prices per square meter. The horizontal axis is the share of immigrants in total population in provinces measured as ratio of number of valid residence permits over total population at the beginning of calendar year. The vertical axis is the average housing price per square meter in Italian provinces at the beginning of calendar year.

34


Figure 3. Log-change of actual vs. predicted number of immigrants

-20

24

6Lo

g-ch

ange

of

acua

l num

ber o

f im

mig

rant

s

-3 -2 -1 0 1 2Log-change of predicted number of immigrants

Fitted values Log-change of number of immigrants

Correlation between log-change of actual and predicted number of immigrants

This graph presents correlation between the log-change of actual number of immigrants and predicted number of immigrants. The vertical axis is the log-change of number of immigrants measured as the number of valid residence permits at the beginning of calendar year. The horizontal axis is log-change in predicted number of immigrants based on initial settlement pattern of immigrants by country of origin.

35


Results of OLS estimation (based on number of immigrants)

Table A

(1) (2) (3) (4) (5) (6) Δln(Number of Immigrants) 0.204*** 0.051 0.058* 0.394** 0.060* 0.370*

(0.001) (0.111) (0.075) (0.014) (0.080) (0.041)

Δln(Number of Immigrants)2 -0.024** -0.022* (0.041) (0.093)

Δln(Population) 1.139 -0.188 -1.084 0.109 -0.724 0.050 (0.263) (0.768) (0.439) (0.938) ( 0.617) (0.972)

ΔUnemployment rate 0.009 0.010 0.009 0.009 (0.117) (0.091) ( 0.168) (0.154)

Δln (GDP per capita) 0.417 0.523 0.375 0.419 ( 0.297) (0.186) (0.359) (0.307)

Δ Density -80.340 -23.420 -60.561 -19.372 ( 0.394) (0.808) (0.525) (0.840)

Number of obs. 595 686 500 500 500 500 Number of prov. 103 103 103 103 103 103 Constant Yes Yes Yes Yes Yes Yes Year dummy No Yes Yes Yes Yes Yes Time-part dummy No No No No Yes Yes R-squared 0.0343 0.5955 0.6141 0.6172 0.6285 0.6306 F-stat 6.74 83.38 73.78 68.64 34.38 34.01 Notes: The table presents results of OLS estimations on a panel of biennial observations for 103 Italian provinces during the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The log change of number of immigrants (i.e. residence permits) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and covers 103 Italian provinces. . Regression also controls for biennial changes in log population, log income, unemployment rates and density of population. The robust standard errors are presented in parenthesis (for now p-values). *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at10%, 5% and 1% significance level, respectively.

36


Results of OLS estimation ( based on concentration of immigrants ) Table B

(1) (2) (3) (4) (5) (6) Δ(Concentration of imm.) 12.271*** -2.604 -3.564 7.614* -3.401 7.468*

(0.000) (0.248) (0.117 ) (0.072) (0.203) (0.068)

Δ(Concentration of imm.)2 -127.315*** -127.381*** (0.002) (0.002)

ΔUnemployment 0.010 0.011 0.009 0.009 (0.099) (0.061) (0.186) (0.172)

Δln (GDP per capita) 0.445 0.399 0.414 0.329 (0.262) (0.315) (0.307) (0.421)

ΔDensity -29.837 -29.342 -19.928 -18.809 (0.530) (0.512) (0.655) (0.663)

Number of obs. 595 595 500 500 500 500 Number of prov. 103 103 103 103 103 103 Constant Yes Yes Yes Yes Yes Yes Year dummy No Yes Yes Yes Yes Yes Time-part dummy No No No No Yes Yes R-squared 0.0353 0.5952 0.6142 0.6203 0.6278 0.6332 F-stat 22.07 111.45 84.94 75.50 37.04 35.68 Notes: The table presents results of OLS estimations on a panel of biennial observations for 103 Italianprovinces for the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The change in concentration of immigrants (i.e. number of valid residence permits over population) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and cover 103 Italian provinces. Regression also controls for biennial changes in log income,unemployment rates and density of population. The robust standard errors are presented in parenthesis (for now p-values). *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at 10%, 5%and 1% significance level, respectively16.

16 For now (in the draft version )I present p-value instead , but it will be changes for later drafts.

37


Results of OLS estimation (based on log-concentration of immigrants)

Table C

(1) (2) (3) (4) (5) (6) Δln(Concentration of imm.) 0.210*** 0.050 0.053* 0.057***

(0.001) (0.116) (0.100) (0.090)

Δ Unemployment 0.009 0.009 (0.138) (0.172)

Δln (GDP per capita) 0.445 0.389 (0.262) (0.339)

Δ Density -12.101 -18.852 (0.783) (0.664)

Number of obs. 585 585 500 500 Number of prov. 103 103 103 103 Constant Yes Yes Yes Yes Year dummy No Yes Yes Yes Time-part dummy No No No Yes R-squared 0.0316 0.5954 0.6137 0.6283 F-stat 11.47 110.94 83.25 36.10 Notes: The table presents results of OLS estimations on a panel of biennial observations for 103 Italian provinces for the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The log-change in concentration of immigrants (i.e. log-change of number of valid residence permits over population) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and covers 103 Italian provinces. Regression also controls for biennial changes in log income, unemployment rates and density of population. The robust standard errors are presented in parenthesis (for now p-values). *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at 10%, 5% and 1% significance level, respectively.

38


Results of the IV estimation (based on number of immigrants) Table A1

(1) (2) (3) (4) (5) (6) Δln(Number of Immigrants) 0.962 0.166 0.175 1.242 0.180 1.062

(0.000) (0.519) (0.455) (0.461) (0.472) (0.392)

Δln(Number of Immigrants)2 -0.069 -0.059 (0.459) (0.393)

Δln(Population) -0.706 -0.125 -1.088 1.893 -0.450 1.468 (0.545) (0.875) (0.539) (0.578) (0.810) ( 0.510)

ΔUnemployment rate 0.012 0.013 0.013 0.012 (0.080) (0.049) (0.063) (0.112)

Δln (GDP per capita) 0.350 0.521 0.298 0.433 (0.423) (0.216) (0.517) (0.322)

Δ Density -75.950 73.136 -40.904 63.653 (0.495) (0.709) (0.759) (0.636)

Number of obs. 556 556 461 461 461 461 Number of prov. 103 103 103 103 103 103 Constant Yes Yes Yes Yes Yes Yes Year dummy No Yes Yes Yes Yes Yes Time-part dummy No No No No Yes Yes R-squared 0.5940 0.6141 0.6136 0.6289 0.6286 F-stat 6.74 87.35 68.48 72.04 30.98 34.01 Notes: The table presents IV (second-stage) estimates on a panel of biennial observations for 103 Italian provinces for the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The log change of number of immigrants (i.e. residence permits) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and covers 103 Italian provinces. Regression also controls for biennial changes in log income, unemployment rates and density of population. The bottom panel reports first-stage estimates of IV regressions(to be added , based on what Sembenelli tells). The first stage instrument is the weighted sum of the changes of immigrant population flow by nationality to whole Italy. The weights are the shares of permits held by each nationality over total permits in a province in 1990 (see equation Section 3 in the main text). The F-statistic for excluded instruments refers to the null hypothesis that the coefficient of the excluded instrument is equal to zero in the first stage. Robust standard errors are presented in parenthesis. *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at 10%, 5% and 1% significance level, respectively.

39


Results of the IV estimation (based on concentration of immigrants)

Table B1 (1) (2) (3) (4) (5) (6) Δ(Concentration of imm.) 39.282 7.615 2.211 21.180

(0.000) (0.560) (0.864) (0.529)

Δ(Concentration of imm.)2 -251.235

(0.411)

ΔUnemployment 0.011 0.015 (0.193) (0.035)


ΔDensity -6.032 -22.848 (0.931) (0.669)

Number of obs. 556 556 461 461 461 461 Number of prov. 103 103 103 103 103 103 Constant Yes Yes Yes Yes Yes Yes Year dummy No Yes Yes Yes Yes Yes Time-part dummy No No No No Yes Yes R-squared . 0.5867 0.6153 0.6206 0.6278 0.6332 F-stat 32.93 101.69 78.95 68.91 37.04 35.68 Notes: The table presents IV (second-stage) estimates on a panel of biennial observations for 103 Italian provinces for the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The change in concentration of immigrants (i.e. number of valid residence permits over population) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and covers 103 Italian provinces. Regression also controls for biennial changes in log income, unemployment rates and density of population. The bottom panel reports first-stage estimates of IV regressions(to be added , based on what Sembenelli tells). The first stage instrument is the weighted sum of the changes of immigrant population flow by nationality to whole Italy. The weights are the shares of permits held by each nationality over total permits in a province in 1990 (see equation Section 3 in the main text). The F-statistic for excluded instruments refers to the null hypothesis that the coefficient of the excluded instrument is equal to zero in the first stage. Robust standard errors are presented in parenthesis. *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at 10%, 5% and 1% significance level, respectively.

40


Results of IV estimation (based on log-concentration of immigrants)

Table C1 (1) (2) (3) (4) (5) (6) Δln(Concentration of imm.) 0.963 0.162 0.212 0.190

(0.000) (0.559) (0.428) (0.460)

Δ Unemployment 0.012 0.0125 (0.089) (0.089)


Δ Density -16.067 -24.508 (0.721) (0.579)

Number of obs. 556 556 461 461 461 461 Number of prov. 103 103 103 103 Constant Yes Yes Yes Yes Year dummy No Yes Yes Yes Time-part dummy No No No Yes R-squared 0.0316 0.5941 0.6122 0.6284 F-stat 11.47 102.01 76.80 35.30 Notes: The table presents IV (second-stage) estimates on a panel of biennial observations for 103 Italian provinces for the period 1996-2007. The biennial log-change of average housing prices in provinces is the dependent variable. The log-change in concentration of immigrants (i.e. log-change of number of valid residence permits over population) is the main explanatory variable of interest. Housing prices data are from 1996, 1999, 2001, 2003, 2005, 2007 and covers 103 Italian provinces. Regression also controls for biennial changes in log income, unemployment rates and density of population. The bottom panel reports first-stage estimates of IV regressions(to be added , based on what Sembenelli tells). The first stage instrument is the weighted sum of the changes of immigrant population flow by nationality to whole Italy. The weights are the shares of permits held by each nationality over total permits in a province in 1990 (see equation Section 3 in the main text). The F-statistic for excluded instruments refers to the null hypothesis that the coefficient of the excluded instrument is equal to zero in the first stage. Robust standard errors are presented in parenthesis. *, ** and *** denote rejection of the null hypothesis of the coefficient being equal to 0 at 10%, 5% and 1% significance level, respectively.

41


42

Table 2 The list of origin countries and their groups used to construct instrument for actual immigration flows. Individual countries Albania, Algeria, Argentina, Australia, Bangladesh,

Brasilia, China, Columbia, Cost D’Avour, Cuba, Dominican Republic, Ecuador, Egypt, Ethiopia, Philippines, Pakistan, Peru, Poland, Romania, Senegal, Somalia, Sri Lanka, USA, Switzerland, Tunis, Turkey, Canada, Ghana, Japan, Jordan, India, Iran, Lebanon, Morocco, Mauritius, Nigeria, Without citizenship

Central Europe Bulgaria, Czech Republic, Slovakia, Hungary Europe (other) UK, San Marino, Spain, Sweden, Andorra, Austria,

Belgium, Cyprus, Vatican, Denmark, Finland, France, Greece, Ireland, Island, Lichtenstein, Luxemburg, Malta, Monaco, Norway, The Netherlands, Portugal

Former Soviet Union Armenia, Azerbaijan, Belarus, Estonia, Georgia, Kazakhstan, Kirgizia, Lithuania, Latvia, Moldavia, Russia, Tajikistan, Turkmenistan, Ukraine, Uzbekistan

Asia Afghanistan, Saudi Arabia, Bhutan, Northern Korea, Southern Korea, Laos, Syria, Yemen, Bahrain, Cambodia, Un. Arabic Emirates, Iraq, Israel, Kuwait, Mongolia, Myanmar, Nepal, Oman, Palestine, Qatar, Taiwan, Thailand, Vietnam

America (South ) Bolivia, Chile, Guyana, Nicaragua, Paraguay, Suriname ,Uruguay

America (Central ) Antigua and Barbuda, Bahamas, Barbados, Belize, Costa Rica, Dominica, Salvador, Jamaica, Grenada, Haiti, Honduras, Mexico, Panama, Santa Lucia, Trinidad and Tobago, Venezuela, Guatemala

Africa (North) Cameroon, Capo Verde, Central Africa, Chad, Eritrea, Gambia, Gibraltar, Djibouti, Guinea, Guinea Bissau, Equatorial Guinea, Kenya, Liberia, Libya, Madagascar, Mali, Mauritania, Nigeria, Ruanda, San Tome and Principe, Sierra Leone, Sudan, Togo, Uganda, Benin, Burkina Faso

Africa (South) Angola, Botswana, Burundi, Comoro, Congo, Gabon, Lesotho, Malawi, Mozambique, Namibia, Democratic Republic of Congo, South Africa, Swaziland, Tanzania, Zaire, Zambia, Zimbabwe

Australia and Oceania Brunei, Fiji, Indonesia, Kiribati, Malaysia, Maldives, Marshal, Micronesia, New Zealand, Palau, Papua New Guinea, San Vincent and Grenadine, San Christ and Nevis, Salomon, Samoa, Seychelles, Singapore, Timor, Kingdom of Tonga, Tuvalu, Vanuatu

Documents

Housing Market Responses to Immigration Shocks; Evidence ... · PDF fileSona Kalantaryan Housing Market Responses to Immigration Shocks: Evidence from Italy. and wage growth. Finally,