Research Article |
Corresponding author: Valeriia G. Oksinenko ( leraoksinenko@gmail.com ) © 2023 Valeriia G. Oksinenko.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Oksinenko VG (2023) The effect of creating a single EAEU labor market on the level of financial well-being of migrant households (exemplified by the Republic of Armenia). Population and Economics 7(1): 21-32. https://doi.org/10.3897/popecon.7.e87175
|
The article analyzes the effect of creating a single EAEU labor market on the financial well-being of households in countries that mainly send migrants, exemplified by the Republic of Armenia. Based on data of the Integrated Living Conditions Survey (ILCS) 2012-2017 and using the difference-in-differences method, the study has identified a significant causal relationship between the country’s accession to the EAEU and reduced income gap between households with migrants in Russia and households with migrants in other countries and regions or without migrants at all. Thus, after Armenia’s accession to the EAEU, income of the households sending migrants to Russia became 6-11% higher than the one of the households sending migrants to other countries, despite the fact that until 2015 there was hardly any difference in income; the income gap with the households without migrants decreased from 6% in 2012-2014 to 2% in 2015-2016. The confirmed positive impact of Armenia’s accession to the EAEU makes it possible to consider it a tool for reducing poverty in the Republic of Armenia as well as recommend to continue development and harmonization of the labor migration policy of the EAUE member states.
Eurasian Economic Union (EAEU), labor migration, Republic of Armenia, single labor market, household income
International migration is a powerful mechanism for smoothing effects of the short- and long-term economic shocks and imbalances in both hosting and sending countries, not only in terms of the economy in general, but at the level of individual households as well. The effect on individual households is particularly interesting for countries that send migrants. Researchers have identified the following two channels of this compensatory effect. The first one is remittances of migrant workers from abroad to their homeland, which positively affects the growth of the household financial well—being (and not only households with migrants, but also those without them) (
For Russia and other countries of the former USSR, the issue of migration is also relevant: the countries are historically interconnected by intense migration flows, with Russia being the main center of attraction (
The purpose of this study is to assess the impact of a single labor market within the EAEU on the financial well-being of migrant households in the Republic of Armenia. The study is a further continuation of the study (
Choosing to study one market and one country allows us to analyze results of the EAEU creation at the micro level and identify emerging trends and evaluate its effects on individual citizens rather than the economy in general.
The Republic of Armenia was also selected for several reasons. On the one hand, it is one of the countries that most intensively sends labor migrants to Russia (
Finally, the fight against poverty is an acute social issue in the Republic: at the beginning of 2022, 30% of the population lived below the poverty line (COFACE). It has been shown (
Logarithm of the average per capita household income in the Republic of Armenia, expressed in drams, depending on the presence or absence of migrants and the direction of migration, 2013-2017
2013 | 2014 | 2015 | 2016 | 2017 | |
Migrant in Russia | 10.6908 | 10.8485 | 10.8252 | 10.7597 | 10.6208 |
Migrant outside Russia | 10.6605 | 10.9186 | 10.8851* | 10.9380** | 10.8588** |
(0.335) | (0.148) | (0.064) | (0.000) | (0.000) | |
No migrants | 10.6933 | 10.8011* | 10.8838* | 10.8598** | 10.8350** |
(0.934) | (0.087) | (0.059) | (0.001) | (0.000) |
An upward trend in the income of the poorest households as a result of the simplified labor migration from Armenia to Russia is a perfect proof of a positive impact of the single labor market on countries that send migrants, substantiating the need for its further liberalization.
The article consists of the following sections:
The empirical basis of the study is data of the Integrated Living Conditions Survey in the Republic of Armenia (hereinafter referred to as the Survey), which has been conducted annually since 2001 (the latest full results are available for 2018).
The data uniqueness in comparison with similar surveys in other EAEU countries, including Russia, is the level of detail (the Survey presents questions that comprehensively cover the household well-being, including participation in migration), continuity of methodology and significant coverage of households in terms of their socio-economic characteristics. The sample size of the Survey varies from 5,000 to 8,000 households living in both urban and rural areas, which is about 20,000 respondents annually, the data are cross-sectional (the sample is formed annually according to the same methodology). The proportional data collection throughout the year makes it possible to level out seasonal shifts, including migration ones.
Since the Republic of Armenia joined the EAEU in 2015, which could potentially affect the household income, we take data for the period 2012-2017, including 2015, to identify the expected effects. We choose 2012 as the initial year due to a change in the methodology: before 2012, the sample size was about 8,000 households, while since 2012 it has been reduced to slightly more than 5,000. The sample was expanded again to 8,000 respondents in 2017, therefore we consider it not rather relevant, and in some cases, we limited our study to 2016 only.
To identify the effect of creating a single labor market within the EAEU on the financial well-being of the Armenian migrant households, the difference-in-differences method is used. This method makes it possible to obtain an unbiased estimate of an event or program by considering the non-random selection. A detailed justification of the relevance of the methodology to the data used and its description are presented in the article (
The difference-in-differences method allows assessing the effect of an event or program on the target group (intervention) by comparing the indicator that evaluates successes or failures of the program (in our case, average per capita income), before and after the introduction of changes. To adjust for other changes between two points in time, we use data on the group that is outside the program (control). Assuming parallel trends in the selected parameter in the control and intervention groups before the program, the effect of the program is identified by a double difference (formula 1) (
E[Y1(1) − Y1(0)|X] =
= {E(Y1|X, W = 1) − E(Y1|X, W = 0)} − {E(Y0|X, W = 1) − E(Y0|X, W = 0)},(1)
where W is an indicative variable of affiliation with the intervention (W = 1) or control (W = 0) groups, Y1 is a measurement of the result after and Y0 is a measurement of the result before, X – covariates, controlling which allows to ensure the condition of parallel trends.
Following the methodology (
For the purposes of the study, we also defined the intervention group in the following two ways: 1) households with migrants in Russia at the time of the Survey; 2) households with migrants only in Russia at the time of the Survey (without migrants in other countries or regions of the Republic of Armenia). Table
Number of Armenian households with and without migrants in their composition, with due regard to the country of residence of the migrant, 2012-2017
Households with migrants currently outside Russia ONLY | Households with migrants currently in Russia ONLY | Households with migrants currently in Russia | Households without migrants | |
2012 | 370 | 535 | 624 | 2 352 |
2013 | 279 | 568 | 652 | 3929 |
2014 | 293 | 529 | 606 | 3 846 |
2015 | 203 | 512 | 571 | 3960 |
2016 | 224 | 557 | 619 | 3837 |
2017 | 303 | 678 | 786 | 5 951 |
To adjust for differences between the intervention and control groups regarding characteristics under study, we included a set of covariates: sex, age (in the form of five-year groups for ages 16 to 80, a 15-year interval for ages 0-15, and an open interval for ages older than 80), marital status, health status, the share of females in the household, a household member with higher education, the share of working household members among working-age citizens, the share of children and the elderly in the household, type of settlement (Yerevan, urban or rural area). In addition, to consider the cross-sectional nature of the data, the main specifications include a set of year dummy variables.
The interval 2012–2014 was taken as the period before the program, and the interval 2015–2017 (2015-2016) was taken as the period after. Also, to test the sensitivity of the results to the assumption of parallel trends in the intervention and control groups, we weakened the assumption and assumed that (linear) trends before joining the EAEU differ across the intervention and control group. The corresponding cross term is statistically significant in all specifications, results of replacing the year dummy variables with the time trend are presented in Table
We also conducted placebo tests. In one of them, we took a placebo period – 2013 instead of 2015 (specifications 7-10, table 6), and in the other – a placebo group: we replaced the intervention group with households that currently have labor migrants in other countries or regions of Armenia (outside Russia), while households that do not include labor migrants were taken as the control group (specifications 11-12, table 6).
The problem of isolating the effect asssociated with joining the EAEU is complicated by the fact that during the same period the Russian ruble (RR) devaluated against the U.S. dollar (USD) and the Armenian dram (AMD) (the national currency of the Republic of Armenia) (Fig.
RR to AMD and USD Exchange rates, 2012-2017. Source: Central Bank of the Russian Federation
The monetary nature of the well-being indicator and the availability of detailed information about the constituent parts of this indicator make it easier to separate one effect from another. Thus, we adjust the amounts of incomes received by households from Russia to the changes in RR to AMD exchange rate, so that all income received from Russia are considered at the 2013 ruble-dram exchange rate. The remaining household income received within the country (that is, in drams) or from other countries (in foreign currencies, mainly in USD and EUR) remained unchanged, calculated at the exchange rate of the year in question. As a result, we get the income values as they would be if the dram-ruble exchange rate were stable throughout the period considered in the study.
As a measure of a single labor market creation effect on the level of the migrant household financial well-being, the logarithm of the average per capita income of the present household members was taken. Since all household incomes were indicated in current prices with due regard to the level of average annual inflation of the Armenian dram, all incomes were brought to the price level of 2013.
Results of the difference-in-differences evaluation for the 2012-2017 interval (Table
Estimation by the difference-in-differences method, changes in the logarithms of the average per capita household income in the Republic of Armenia, expressed in drams, depending on the presence or absence of migrants and the direction of migration, 2012-2017
Intervention | Households with migrants currently in Russia | Households with migrants currently in Russia ONLY | Households with migrants currently in Russia ONLY | |||
Control | Households with migrants currently outside Russia ONLY | Households without migrants | ||||
(1) | (2) | (3) | (4) | (5) | (6) | |
Before | 2012-2014 | |||||
Control | 10.247 | 10.124 | 10.247 | 10.031 | 10.403 | 10.018 |
Intervention | 10.151 | 10.109 | 10.160 | 10.043 | 10.160 | 9.950 |
Diff (T-C) | -0.0959 | -0.0153 | -0.0875 | 0.0125 | -0.243 | -0.0680 |
After | 2015-2017 | |||||
Control | 10.415 | 10.193 | 10.415 | 10.251 | 10.572 | 10.133 |
Intervention | 10.454 | 10.218 | 10.485 | 10.317 | 10.485 | 10.094 |
Diff (T-C) | 0.0400 | 0.0252 | 0.0708 | 0.0662 | -0.0866 | -0.0396 |
Diff-in-diff | 0.136** | 0.0405 | 0.158** | 0.0537* | 0.156** | 0.0284 |
(0.0217) | (0.0250) | (0.0222) | (0.0256) | (0.0144) | (0.0175) | |
Covariates | no | yes | no | yes | no | yes |
Year dummy variables | yes | yes | yes | yes | yes | yes |
Observations | 25629 | 14455 | 23215 | 13116 | 97337 | 73010 |
Control | 7182 | 4574 | 7182 | 4574 | 81304 | 64468 |
Intervention | 18447 | 9881 | 16033 | 8542 | 16033 | 8542 |
R-squared | 0.027 | 0.204 | 0.030 | 0.206 | 0.023 | 0.325 |
Means and standard errors were estimated using linear regression. Robustness of Standard Errors: Significance: ** p<0.01; * p<0.05. |
Estimation by the difference-in-differences method, changes in the logarithms of the average per capita household income in the Republic of Armenia, expressed in drams, depending on the presence or absence of migrants and the direction of migration, 2012-2016
Intervention | Households with migrants currently in Russia | Households with migrants currently in Russia ONLY | Households with migrants currently in Russia ONLY | |||
Control | Households with migrants currently outside Russia ONLY | Households without migrants | ||||
(1*) | (2*) | (3*) | (4*) | (5*) | (6*) | |
Before | 2012-2014 | |||||
Control | 10.247 | 9.987 | 10.247 | 10.061 | 10.403 | 10.014 |
Intervention | 10.151 | 9.976 | 10.160 | 10.081 | 10.160 | 9.950 |
Diff (T-C) | -0.0959 | -0.0116 | -0.0875 | 0.0194 | -0.243 | -0.0635 |
After | 2015-2016 | |||||
Control | 10.375 | 10.114 | 10.375 | 10.270 | 10.527 | 10.097 |
Intervention | 10.455 | 10.178 | 10.491 | 10.387 | 10.491 | 10.079 |
Diff (T-C) | 0.0800 | 0.0637 | 0.116 | 0.117 | -0.0363 | -0.0171 |
Diff-in-diff | 0.176** | 0.0753* | 0.203** | 0.0974** | 0.207** | 0.0464* |
(0.0263) | (0.0306) | (0.0267) | (0.0312) | (0.0168) | (0.0207) | |
Covariates | no | yes | no | yes | no | yes |
Year dummy variables | yes | yes | yes | yes | yes | yes |
Observations | 20448 | 11525 | 18595 | 10476 | 73894 | 54982 |
Control | 5771 | 3667 | 5771 | 3667 | 61070 | 48173 |
Intervention | 14677 | 7858 | 12824 | 6809 | 12824 | 6809 |
R-squared | 0.024 | 0.210 | 0.028 | 0.220 | 0.020 | 0.318 |
Means and standard errors were estimated using linear regression. Robustness of Standard Errors: Significance: ** p<0.01; * p<0.05. |
To verify the obtained results, we conducted a test for resistance to the time trend: in the specification 1-6, instead of the year dummy variables, we included a linear time trend and the cross term “intervention group * linear trend”. Difference-in-differences estimates retained their sign, although the differences became more significant. Both variables were significant across all specifications (Table
Estimation by the difference-in-differences method, changes in the logarithms of the average per capita household income in the Republic of Armenia, expressed in drams, depending on the presence or absence of migrants and the direction of migration, tests for resistance to the time trend
Intervention | Households with migrants currently in Russia | Households with migrants currently in Russia ONLY | Households with migrants currently in Russia ONLY | |||
Control | Households with migrants outside Russia ONLY | Households without migrants | ||||
(1) | (2) | (3) | (4) | (5) | (6) | |
Before | 2012-2014 | |||||
Control | 10.031 | 9.816 | 10.033 | 9.875 | 10.283 | 9.816 |
Intervention | 9.985 | 9.946 | 10.011 | 10.061 | 10.011 | 9.844 |
Diff (T-C) | -0.0455 | 0.130 | -0.0195 | 0.187 | -0.272 | 0.0276 |
After | 2015-2017 | |||||
Control | 9.862 | 9.677 | 9.862 | 9.734 | 10.297 | 9.827 |
Intervention | 10.027 | 10.082 | 10.106 | 10.254 | 10.106 | 10.040 |
Diff (T-C) | 0.165 | 0.406 | 0.244 | 0.520 | -0.192 | 0.213 |
Diff-in-diff | 0.211** | 0.275** | 0.264** | 0.333** | 0.0803* | 0.185** |
(0.0473) | (0.0547) | (0.0484) | (0.0560) | (0.0314) | (0.0381) | |
Covariates | no | yes | no | yes | no | yes |
Year dummy variables | instead of year dummy variables, a linear time trend and a cross term “experimental group*linear trend” are included | |||||
Observations | 25629 | 14455 | 23215 | 13116 | 97337 | 73010 |
Control | 7182 | 4574 | 7182 | 4574 | 81304 | 64468 |
Intervention | 18447 | 9881 | 16033 | 8542 | 16033 | 8542 |
R-squared | 0.035 | 0.202 | 0.037 | 0.211 | 0.027 | 0.324 |
Means and standard errors were estimated using linear regression. Robustness of Standard Errors: Significance: ** p<0.01, * p<0.05. |
In placebo tests, we also tested sustainability of the results by replacing the intervention group with a group of the households with migrants who are not in Russia at the time of the Survey (specifications 11-12), as well as replacing the year of the formation of the union (2015 with 2013) (specifications 7-10). The use of the placebo group and placebo period in all specifications with the inclusion of covariates gave results with the opposite sign. However, specifications 7 and 9 (without including covariates) turned out to be significant as well, but in them the difference-in-differences turned out to be positive; in specification 11, the difference remained negative, yet insignificant. The results obtained in all specifications indicate that 2015 was indeed a turning point in changing the indicators of financial well-being of households in the intervention group, as well as the fact that Armenia’s accession to the EAEU had an effect on the households with migrants in Russia only (Table
Estimation by the difference-in-differences method, changes in the logarithms of the average per capita household income in the Republic of Armenia, expressed in drams, depending on the presence or absence of migrants and the direction of migration, placebo tests
Placebo period | Placebo group | |||||
Intervention | Households with migrants currently in Russia | Households with migrants currently in Russia ONLY | Households with migrants currently outside Russia | |||
Control | Households with migrants currently outside Russia ONLY | Households without migrants | ||||
Before | (7) | (8) | (9) | (10) | (11) | (12) |
2012 | 2012-2014 | |||||
Control | 10.089 | 9.874 | 10.089 | 9.934 | 10.403 | 9.819 |
Intervention | 9.984 | 9.981 | 10.007 | 10.088 | 10.210 | 9.702 |
Diff (T-C) | -0.104 | 0.107 | -0.0820 | 0.154 | -0.192 | -0.117 |
After | 2013-2017 | 2015-2017 | ||||
Control | 10.375 | 10.214 | 10.375 | 10.274 | 10.572 | 10.122 |
Intervention | 10.368 | 10.195 | 10.388 | 10.286 | 10.368 | 9.972 |
Diff (T-C) | -0.00654 | -0.0194 | 0.0131 | 0.0116 | -0.204 | -0.150 |
Diff-in-diff | 0.0979** | -0.126** | 0.0951** | -0.143** | -0.0116 | -0.0325* |
(0.0269) | (0.0320) | (0.0276) | (0.0332) | (0.0170) | (0.0189) | |
Covariates | no | yes | no | yes | no | yes |
Year dummy variables | yes | yes | yes | yes | yes | yes |
Observations | 25629 | 12268 | 23215 | 13116 | 90900 | 70381 |
Control | 7182 | 4574 | 7182 | 4574 | 81304 | 64468 |
Intervention | 18447 | 9881 | 16033 | 8542 | 9596 | 5913 |
R-squared | 0.026 | 0.205 | 0.025 | 0.214 | 0.019 | 0.337 |
Means and standard errors were estimated using linear regression. Robustness of Standard Errors: Significance: ** p<0.01; * p<0.1. |
Until 2015, the incomes of the households sending migrants to Russia were on average the same as the incomes of the households sending migrants to other countries, and 6% lower than those of the households without migrants. After Armenia’s accession to the EAEU, the incomes of the households sending migrants to Russia became 6-11% higher compared to the households sending migrants to other countries, and the income gap with the households without migrants decreased to 2%. The methodology of the study suggests that the changes are due to Armenia’s accession to the EAEU in 2015, while this effect failed to spread to other groups of the households.
Since the study considered the period until 2017 only and the effect of creating a single labor market exclusively, we did not touch upon the impact of significant global events on migration flows. Thus, the study neither focused on the impact of the COVID-19 pandemic in 2019 and 2020 and the associated significant restrictions on movements of international migrants nor considered the escalation of the armed conflict in Nagorno-Karabakh (2020-2022). These developments are certainly additional factors that can reduce the effects of the creation of a single labor market on household incomes and are potential areas for future research on this topic.
Since the poorest households mainly send migrants to Russia, the chosen form of migration policy in the EAEU can become one of the tools to reduce poverty in the Republic of Armenia. The confirmed positive impact of the single labor market functioning makes it possible to recommend continuing the process of harmonization of the legislation of the EAEU member states in relation to labor migration, especially against the background of instability arising under the influence of external factors.
Chudinovskikh O, Denisenko M (2014) Migration Between CIS Countries: Working Paper —WP3. URL: https://www.hse.ru/data/2013/04/15/1233209379/Денисенко-докл.pdf
COFACE. Armenia / Economic Studies. URL: https://www.coface.com/Economic-Studies-and-Country-Risks/Armenia (Accessed on: 13.12.2022)
Perezagruzka otnoshenij ES — Armeniya: Moskva ne protiv? [Reset of EU—Armenia relations: Does Moscow mind?] 1.03.2021. URL: https://www.dw.com/ru/es-armenija-perezagruzka-otnoshenij/a-56718077 (Accessed on: 18.02.2022) (in Russian)
World Bank (2019) Armenia International Outmigration: An Exploration on the Effects on Armenian Households’ Welfare. World Bank, Washington, DC. © World Bank. https://openknowledge.worldbank.org/handle/10986/32537. License: CC BY 3.0 IGO
Oksinenko Valeria Gennadievna – 1st category engineer of the Laboratory of Population Economics, Lomonosov Moscow State University, Moscow, 119234, Russia. Email: leraoksinenko@gmail.com