Impact of Low‐Skilled Immigration on Female Labour Supply

• Published date: 01 April 2015
• Date: 01 April 2015
• Author: Emanuele Forlani,Elisabetta Lodigiani,Concetta Mendolicchio
• DOI: http://doi.org/10.1111/sjoe.12101

Scand. J. of Economics 117(2), 452–492, 2015

DOI: 10.1111/sjoe.12101

Impact of Low-Skilled Immigration

on Female Labour Supply∗

Emanuele Forlani

University of Pavia, 27100 Pavia, Italy

emanuele.forlani@eco.unipv.it

Elisabetta Lodigiani

Ca’ Foscari University of Venice, 30121 Venice, Italy

elisabetta.lodigiani@unive.it

Concetta Mendolicchio

Institute for Employment Research, DE-90478 Nuremberg, Germany

concetta.mendolicchio@iab.de

Abstract

In this paper, segmenting the market by educational levels, we investigate which native-born

women are more affected by an increase of low-skilled immigrants working in the household

service sector. We present a model of individual choice with home production and, using a

harmonized dataset (the Cross-National Equivalent File), we estimate its main comparative

static results. The results suggest that the share of immigrants working in services is positively

associated with an increase of native-born women’s labour supply at the intensive margin,

if skilled, and at the extensive margin, if unskilled. Moreover, the results show that these

effects are larger in countries with less-supportive family policies.

Keywords: Family policy; female labour decision; international migration

JEL classification:J22; J61

I. Introduction

International migration can significantly alter the labour market conditions

in destination countries. As a first-order effect, it might change the pre-

vailing wage rates in sectors where a large number of migrants look for

a job. In particular, unskilled migrants represent a significant fraction of

the employed in sectors that provide services to households. Several recent

studies (Barone and Mocetti, 2011; Cort´

es and Tessada, 2011, hereafter CT;

∗We acknowledge the financial support of the IAB, the data contribution of Australian Bureau

of Statistics, HILDA, SHP, SFSO, SOEP, BHPS, Office for National Statistics (UK), PSID,

IPUMS, and Bertelsmann Foundation. We are grateful to the editor and two anonymous

referees for insightful comments. The responsibility for any mistakes is our own. E. Forlani

acknowledges support from the Cariplo Foundation project no. 2010-2372.

CThe editors of The Scandinavian Journal of Economics 2014.

E. Forlani, E. Lodigiani, and C. Mendolicchio 453

Far r´

eet al., 2011) show that low-skilled immigrants have contributed to a

decrease in the prices of household services where they specialize, or in

sectors with a high concentration of low-wage workers, such as housekeep-

ing, childrearing, or caring for the elderly; see Cort´

es (2008) and Frattini

(2012) for the US and UK, respectively. Given that these services are,

typically, a substitute for time-consuming activities carried out, mostly by

women, within the family, there might be a second-order effect on the

labour supply of native women. This could be especially true in countries

with weak family policies.

In this paper, we aim to study the impact that unskilled immigrants in the

household service sector have on native female labour supply for a group of

developed countries, which differ in terms of family policies, considering

both their effects on the employment decision and on the number of hours

worked.

Previous studies closely related to ours have focused on a single country,

usually one with weak family-supporting policies. In particular, for the US,

CT find that the labour supply of working women at the top quartile of

the female wage distribution is positively affected at the intensive margin

by low-skilled immigration. This effect is particularly relevant for highly

skilled women. For Spain, Farr´

eet al. (2011) show that the positive effect

of female immigration on the labour supply of women depends on the edu-

cational level and the family responsibilities of the native women. For Italy,

Barone and Mocetti (2011) study how the inflow of female immigrants af-

fects the labour supply of both high-skilled and low-skilled women. As in

previous papers, they find a positive impact only on the hours worked by

high-skilled natives. They also find a stronger effect in Italian municipali-

ties where local welfare services are less developed. In a slightly different

perspective, Furtado and Hock (2010) evaluate how immigration affects the

employment–maternity trade-off in the US.

Our paper adds to these results from several points of view. First, we

perform a multilevel, cross-country analysis. This allows us, using the

cross-country dimension of our dataset, to consider how unskilled migrants’

labour supply interacts with the policies adopted to support the families.

This is of particular importance because differences in family policies

across OECD countries contribute to the significant variations in women’s

labour supply behaviour (e.g., Del Boca and Sauer, 2009; Del Boca et al.,

2009). Second, these family policies are usually more effective for unskilled

women, encouraging part-time work and jobs in lower-level positions (Blau

and Kahn, 2013). For this reason, in the empirical analysis, we test whether

this effect is stronger for skilled or unskilled natives, conditional on the

degree of family policy support. Third, using the International Standard

Industrial Classification, we adopt an occupation-based definition to study

the role of unskilled migrants employed in a subgroup of service sectors

CThe editors of The Scandinavian Journal of Economics 2014.

454 Impact of low-skilled immigration on female labour supply

compared to the total labour force size.1Finally, the theoretical model

we adopt to analyse the labour supply decisions extends and generalizes

that of CT. Here, one of the key assumptions is that an increase in the

concentration of unskilled migrants in services puts downward pressure on

the local prices of household services. This assumption is supported by the

literature and confirmed in our empirical analysis. In the model, household

services bought on the market and own time are inputs in a household

production function. Under appropriate restrictions on the utility and home

production functions, as a general result, we find that a decrease in the

price of these services increases the native labour supply. This effect is

larger, for low-wage native women, if the country implements weak family

policies.

The empirical analysis aims to test the comparative statics properties of

the model and is based on the Cross-National Equivalent File (CNEF), ob-

tained with harmonized data from different national surveys. Specifically,

we use the compatible surveys for three years (2001, 2003, and 2005)

and for five countries (Australia, Germany, Switzerland, UK, and US). Our

identification strategy exploits the variability within region–year in the con-

centration of low-skilled immigrants in the service sector. To address the

potential endogeneity issues caused by the location choices of immigrants

(because their distribution across regions is not random), and by measure-

ment errors (because of undocumented migrants), we exploit the tendency

of migrants to move to regions with a large share of migrants of the same

origin and use the past distribution of migrants across regions (Card, 2001).

We present three sets of results. First, we estimate the impact of immi-

gration on the probability to be employed for native women aged 22–45 by

skill levels. In contrast to what has been obtained in previous research, we

find that there is a positive and statistically significant effect on the average

probability of working for unskilled native women. Consistently with the

previous body of literature, there is no detectable impact on the probability

of being employed for skilled women. Second, we test the empirical rela-

tionship between unskilled migration and the number of hours worked per

week. We observe that, compatible with the theoretical results, migration

increases the number of weekly hours worked by skilled women. How-

ever, there is no effect on the choices at the intensive margin of unskilled

women. All these findings are particularly strong for women in households

with children. The results are robust to different sample composition and

identification tests. Finally, we show that, consistent with the theoretical

1Barone and Mocetti (2011) also focus on female immigration specialized in household

services. Unlike us, they consider migrants from a subgroup of countries with the highest

share of the employed in domestic services in Italy. Therefore, they define specialization in

terms of home-country basis and not in terms of actual occupation.

CThe editors of The Scandinavian Journal of Economics 2014.