by Emanuela Ghignoni[*], Marilena Giannetti[†] and Vincenzo Salvucci[‡]
This paper explores the evolution of wage gaps in Italy by gender and citizenship. Using Labour Force Survey (LFS) data over the period 2009-2020 we apply two different matching comparison methodologies, the Ñopo decomposition and the Inverse Probability Weighted Regression Adjustment (IPWRA) technique, that allow “like for like” comparisons between individuals and are able to take into account how gender interacts with citizenship in shaping wages. Our findings show that the general gender wage gap in Italy is rather low. This gap is largely explained by workers’ observable characteristics. Conversely, the citizenship wage gap appears to be much larger. Moreover, most of the reported wage gap seem to be explained by unobservable characteristics. We finally estimate the double-negative effect of being both a woman and a foreigner. Non-Italian women earned on average 44.3% per hour less than Italian men in 2009 and 46.5% less than Italian men in 2020.
This article explores the extent and evolution of wage gaps in Italy by gender and citizenship. In the last few decades, research on the gender wage gap has produced a large body of literature. Various explanations for the existence of this gap have been advanced such as different experience, maternity penalty, and lower qualifications (see Blau and Kahn, 2017). On top of that, wage gaps between migrant[1] and native workers are registered with the same frequency as gender gaps, if not more so. Most of the studies on these issues focus on either male-female or migrant-native wage gap, and the related empirical literature often uses widely the Oaxaca-Blinder methodology (Blinder, 1973; Oaxaca, 1973) to estimate that part of the wage gap that cannot be explained by differences in worker’s characteristics and that may, therefore, be attributable to discrimination and/or unobservable characteristics.
Moreover, most of the empirical analysis, especially those concerning immigrants, uses data from the US and a few other countries. Only a handful of studies focus on the immigrant wage gap in Italy (Venturini and Villosio, 2008; Dell'Aringa et al., 2015; Venturini et al., 2018; Piazzalunga, 2015, among others). Nonetheless, the relevance of the immigrant population, and of immigrant women in particular, increased in the Italian labour market. According to OECD data (OECD, 2021), in 2020 permanent immigrants represented more than 10% of the Italian active population and 46% of these were women. Immigrant workers reached 14,2% of total employment, 40 per cent of which were women. Hence, even though the relevance of migrant women in Italy is increasing, the economic literature has not devoted much attention to them.
The analysis presented in this paper aims at filling this gap, shedding light on if and how much migrant workers, women and migrant women are penalized in the Italian labour market. Differently from Piazzalunga (2015), which also focuses on “double discrimination” using 2009 data, we analyse the entire 2009-2020 period using different matching techniques. This period has been characterized by several changes in Italian politics, social struggles and a decade-long economic crisis from which the economy has not yet fully recovered. Moreover, Italy experienced an increased influx of migrants during these years, which reached unprecedented levels (OECD, 2021). Moreover, Piazzalunga (2015) applies the Oaxaca-Blinder and the Shamsuddin decomposition (Shamsuddin, 1998), whereas we decompose the gender wage gap that exists between the different categories using two different decomposition approaches: the Ñopo decomposition (Ñopo, 2004) and the Inverse Probability Weighted Regression Adjustment (IPWRA) decomposition (Cattaneo, 2010).
Based on quarterly Labour Force Survey (LFS) sample data, produced and collected by the Italian Statistical Office (ISTAT), we proceed first with the estimation of hourly wages for both Italian and foreign men/women workers. The data used in the empirical section are drawn from Istat Labour Force surveys during 2009- Q1; 2020 Q4. For the aim of this paper, we consider only dependent workers, who work from 20 to 80 hours per week. Indeed, our data source does not contain information about the earnings of self-employed workers. Our sample is made of 1,629,708 dependent workers, 172,436 of which are foreigners.
We decompose the gender wage gap that exists between the different categories. In doing so, we use the Ñopo decomposition (Ñopo, 2004), which is based on matching and makes it possible to eliminate the estimation bias due to the different distribution of the characteristics among male and female workers; and the Inverse Probability Weighted Regression Adjustment (IPWRA) decomposition (Cattaneo, 2010), that confirms the robustness of our results. To the best of our knowledge, this is the first time that an IPWRA procedure is applied to the study of the citizenship/gender wage gap.
Concerning the main results, using the Ñopo decomposition we get that:
- When the gender gap is considered (regardless of nationality), results show that the total gap in hourly wages is rather low. Moreover, the part which can be attributed to “discrimination” is about zero when labour market variables are included in the analysis, as women are (partly) concentrated in lower paid types of jobs.
- When the citizenship gap is taken into account (regardless of gender), the total gap in hourly wages grows to 33% (in favour of Italian workers). The part attributable to “discrimination” is very high, and it even increases when labour market variables are included in the model. That is, foreign workers not only are concentrated in lower paid jobs, but they are also paid less than Italians in all types of jobs.
- The wage gap between Italian men and foreign women is therefore extremely high (47%) and discrimination accounts for most of it. Moreover, the latter increases (from 43.4% to 47.9%) when labour market variables are included in the analysis, as foreign women are not only employed in lower paid jobs, but are also paid less than Italian men in all types of jobs.
As introduced, we also analyse the evolution of the gender/citizenship gap during the whole period under observation by applying the Ñopo decomposition to different years (Figure 1). As also found in Piazzalunga and Di Tommaso (2019) the total gender gap seems to increase during the financial crisis. Nevertheless, the increase we register appears to be much less pronounced than in other studies. This may be due to the fact that our results are not fully comparable with those obtained using a Oaxaca-Blinder decomposition, as we compare “like with like” individuals through matching techniques that should produce lower estimated gaps.
The gender gap reduces even further in 2020 (Figure 1, panel a). This is not in contrast with the findings of Bonacini et al. (2021) who revealed a positive association between the level of “working from home” attitude of occupations and the gender wage gap only for (older and married) female employees working in the private sector. As in Italy women are more likely to be employed in the public sector, this may have contributed to protect female wages during the pandemic. Also, the part attributable to discrimination appears very low.
The total citizenship gap, even if it is much higher than the gender gap, follows the same pattern (Figure 1, panel b). It increases until the peak of the crisis and decreases afterwards in 2019 and 2020. As a matter of fact, foreign workers lost their jobs more frequently than Italians both during the financial crisis and the Covid pandemic. Nevertheless, the citizenship gap grew fast during the crisis, when some immigrants may have accepted wage cuts to remain employed, whereas it remained stable between 2019 and 2020 (Figure 1, panel b). Probably, in this case, foreign workers were not able to trade-off wage for job stability and simply lost their job, due to the implementation of social distancing measures and the low level of “working from home” attitude of the types of occupation they most frequently hold. This seems to be particularly valid to explain the evolution of the total gap in hourly wages between Italian men and foreign women (Figure 1, panel c). This gap increased from 2009 to 2015 and, after a temporary reduction in 2016, it stable from 2017 onwards. Even in this case, migrant women may have been able to maintain their job during the financial crisis, at the cost of a wage cut, but not during the Covid pandemic. Indeed, the employment rate for foreign women fell by 4.9%, in 2020. This is more than double the drop found for foreign men (2.2%) and eight times more than that of Italian women (-0.6%, in line with that of Italian men). This can be explained by the fact that over half of foreign women are employed as domestic workers, caregivers and cleaning staff for offices and businesses, so that a large share of them could not continue working and simply lost their job when severe constraints on mobility were imposed to contrast the pandemic. IPWRA estimations are very close to results obtained by Ñopo decomposition and both methods of estimation provide evidence of an increase in the gaps (in particular, in the citizenship gaps) during the crisis, proving the robustness of our results.
Overall, and in line with previous studies, we find that in Italy the gender wage gap exists but is rather low, and largely explained by observable characteristics of demographics, location, education, work experience and job characteristics. Moreover, it slightly increases up to the peak of the financial crisis and declines afterwards. Conversely, the citizenship wage gap is much larger than the gender gap and the unobservable component accounts for most of the observed gaps. Regarding the time trend, we find that it also increases until 2015 and declined after the peak of the financial crisis. Conversely, when we estimate the double-negative effect of being both a woman and a foreigner worker, using the same decomposition approaches highlighted above, it appears that non-Italian women earned on average 44.3% per hour less than Italian men in 2009 and they earned 46.5% less than Italian men in 2020. Even in this case, the gap is larger –and equal to 50%– at the peak of the financial crisis. The decomposition results show that differences in unobservable characteristics play a crucial role in explaining citizenship wage gaps.
As it occurs in other studies on the topic, the results obtained might in part be biased by some unobservable worker characteristic; nonetheless they give some insight on the behaviour of the Italian labour market and call for policy intervention. Given the focus on the double discrimination of foreign women, policy recommendations include both measures to reduce the immigrant-native and the gender gap, such as reducing the language barriers between immigrant and native workers, predominantly targeting women; raising immigrant workers’ –and especially female immigrant workers’– awareness regarding local legislation and individual rights; increasing the transferability of immigrant workers’ qualifications and skills, and promoting active labour policies, specifically targeting women.
References
Blau, F. D., & Kahn, L. M. (2017). The gender wage gap: Extent, trends, and explanations. Journal of economic literature, 55(3), 789-865.
Blinder, A. S. (1973). Wage discrimination: reduced form and structural estimates. Journal of Human resources, 436-455.
Bonacini, L., Gallo, G., & Scicchitano, S. (2021). Will it be a shecession? The unintended influence of working from home on the gender wage gap related to the COVID-19 pandemic (No. 771). GLO Discussion Paper.
Cattaneo, M. D. (2010). Efficient semiparametric estimation of multi-valued treatment effects under ignorability. Journal of Econometrics, 155(2), 138-154.
Dell’Aringa, C., Lucifora, C., & Pagani, L. (2015). Earnings differentials between immigrants and natives: the role of occupational attainment. IZA Journal of Migration, 4(1), 1-18.
Ñopo, H. (2004). Matching as a Tool Decompose Wage Gaps. IZA Discussion Papers 981. Institute for the Study of Labor (IZA).
Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International economic review, 693-709.
Piazzalunga, D. (2015). Is There a Double‐Negative Effect? Gender and Ethnic Wage Differentials in I taly. Labour, 29(3), 243-269.
Piazzalunga, D., & Di Tommaso, M. L. (2019). The increase of the gender wage gap in Italy during the 2008-2012 economic crisis. The Journal of Economic Inequality, 17(2), 171-193.
Shamsuddin, A. F. (1998). The double-negative effect on the earnings of foreign-born females in Canada. Applied Economics, 30(9), 1187-1201.
Venturini, A., & Villosio, C. (2008). Labour-market assimilation of foreign workers in Italy. Oxford Review of Economic Policy, 24(3), 517-541.
Venturini, A., Strom, S., Piazzalunga, D., & Villosio, C. (2018). Wage assimilation of immigrants and internal migrants: the role of linguistic distance.
University of Rome La Sapienza,
[†] University of Rome La Sapienza,
[‡] University of Copenhagen,
[1] In the article we use the term “migrant workers” to refer to workers who do not have Italian citizenship, as found in the Labour Force Survey (LFS)
