Responsibility for opinions expressed in signed articles rests solely with their authors, and publication does not constitute an endorsement by the ILO.
This article is also available in French, in Revue internationale du Travail 165 (2), and Spanish, in Revista Internacional del Trabajo 145 (2).
1. Introduction
The International Monetary Fund (IMF) is a central actor in global governance, influencing developing countries through conditional lending programmes that require policy reforms in exchange for loans. Labour market reforms have long been a part of IMF recommendations and have been shown to negatively affect workers’ incomes and rights (Anner and Caraway 2010; Caraway 2006; Reinsberg et al. 2019; Vreeland 2002). The IMF, however, claims that its reforms increase labour market flexibility, enabling traditionally disadvantaged groups, such as women and young people, to enter the workforce. The IMF Human Resources Director, Margaret Kelly, for example, argues that there is a direct causal relationship between labour flexibility and women’s labour force participation and suggests that:
Access to paid jobs is critical for the advancement of women, in both developing and developed countries. The Fund supports policies that promote flexibility in labor markets, so as to increase employment opportunities, as well as measures such as training that increase productivity and thereby wages. (IMF 2000)
This advice is echoed in Article IV Consultations of the IMF1 with Member States. In Greece in 2009, for instance, IMF staff suggested that “Greece [could] also facilitate more part-time work to boost participation of youths and women in the labor force, and ease employment protection legislation” (IMF 2009, 30). Similarly, in Chile in 2006, “Staff urge[d] the authorities to take bold action […] to help address the still high unemployment rate – especially among the young – and improve labor force participation” by introducing greater flexibility in the labour market, for example, by increasing working hours and the duration of fixed-term contracts and by avoiding a rapid increase in the minimum wage (IMF 2006, 23). Two researchers from the IMF’s Research Department argue that strong labour market institutions reserve high-pay, high-security and high-protection jobs for “insider” groups and make it extremely hard for women and young people to enter the labour market (Duval and Loungani 2019). The IMF suggests that, by offering greater flexibility, its labour market reforms encourage greater participation of traditionally excluded groups, such as women and young people.
Scholars have previously looked at the impact of IMF programmes on women’s economic and political rights (Detraz and Peksen 2016) and on the gendered labour force participation gap under and outside IMF programmes (Kern, Reinsberg and Lee 2024). However, the gendered impact of labour market reforms on women’s labour force participation remains unstudied. Since the IMF’s claim of helping women enter the labour force specifically rests on its labour market flexibilization measures, it is important to study whether IMF-sponsored labour market reforms positively contribute to the labour force participation of women.
This article shows that IMF-sponsored labour market reforms, contrary to the IMF’s claim, are associated with an overall reduction in the labour force participation of women after controlling for economic crises and several covariates of women’s labour force participation. Delving deeper into the negative impact, it shows that the adverse impact is mainly driven by wage decline under IMF-sponsored labour market reforms. Lowering firing costs, however, may temporarily increase female labour force participation by creating higher job turnover.2
I obtain these results using a compound instrumental variable analysis. After accounting for non-random selection into IMF programmes and labour conditions and controlling for relevant covariates, the analysis supports the theory that that IMF-sponsored labour market reforms are associated with lower labour force participation rates for women. Robustness checks provide further solid evidence for this theory. I supplement this with causal mediation analysis and show that labour market reform drives down wages and that this decline, in turn, lowers women’s labour force participation. I also show that lower firing costs might introduce more movement in the market and encourage women’s labour market participation in the short term.
The next section in this article provides an overview of studies on the impact of IMF programmes on labour markets in general and on women’s labour force participation in particular. It proposes a theory as to why IMF-sponsored labour market reforms might reduce women’s labour force participation. The third section explains the research design, and the fourth section provides empirical support for the proposed theory. The final section concludes with some policy recommendations and suggests areas for further research.
2. The IMF and labour market reform
The IMF is arguably the most powerful international organization in terms of its impact on borrowing countries (Stone 2004), as it can launch extensive economic reforms with far-reaching consequences for the economies and politics of those countries. Scholars have demonstrated that programmes can stall economic growth (Dreher 2006; Hajro and Joyce 2009; Przeworski and Vreeland 2000; Vreeland 2003) and fuel corruption (Reinsberg, Kentikelenis and Stubbs 2021). Programmes can also exacerbate inequalities and poverty (Forster et al. 2019; Lang 2021; Oberdabernig 2013). They can also redistribute income away from labour groups (Garuda 2000; Vreeland 2002), undermine their collective bargaining and freedom of association rights (Caraway 2006; Reinsberg et al. 2019) and place a disproportionate burden of economic adjustment on labour groups compared to financial interest groups (Metinsoy 2025).
Because of their far-reaching economic consequences and redistribution of resources, IMF programmes can also create significant political instability (Abouharb and Cingranelli 2007). They can aggrieve immobile labour groups and trigger large-scale protests, strikes and riots (Reinsberg, Stubbs, and Bujnoch 2023; Metinsoy 2024). In addition to political instability and protests, they can provoke repression from the government and consequently undermine the protection of human rights in borrowing countries (Abouharb and Cingranelli 2009; Franklin 1997; Mukherjee and Yadav 2024; Pion-Berlin 1983).
Evidence of the gendered consequences of IMF programmes has recently started to accumulate. Detraz and Peksen (2016) argue that IMF programmes undermine state capacity to protect women’s economic and political rights by curtailing government spending and administrative capacity. Furthermore, as women are more likely to be employed in the public sector (Iversen and Rosenbluth 2010; Steiber and Haas 2012), privatizations under IMF programmes disproportionately hurt women (Detraz and Peksen 2016). More recently, Reinsberg et al. (2024) have shown that the negative impact on women’s employment in the civil service is reduced if there is at least one female minister in the cabinet, underlining the importance of representation.
Mathers (2020) argues that austerity and social spending cuts under international financial institutions’ programmes place the burden of welfare state downsizing on women. When investment in social services, such as in health and education, declines, schools become less safe for girls, and women and girls are the first to lose access to healthcare (Marphatia 2010). This occurs because spending cuts can reduce school supervision and infrastructure and increase the cost of healthcare, while gender norms often lead households to prioritize men’s and boys’ education and health. In a similar vein, Kern, Reinsberg and Lee (2024) suggest that austerity and privatizations under IMF programmes not only widen the gender labour force participation gap, but also trigger an increase in the number of suicides among women and have other adverse effects on women’s health. Ortiz and Cummins (2013, 2021) report that following the 2008 financial crisis, IMF-imposed budget cuts negatively affected children and women, often through retrenching of social assistance and reductions in public healthcare services. Kentikelenis and Stubbs (2023) further show that IMF “social spending floors” often function as ceilings or are poorly implemented, failing to prevent declines in social budgets during adjustment programmes, once again hurting the most vulnerable groups in society. Finally, Buenaventura and Miranda (2017) discuss how VAT increases under IMF programmes hurt women more than men, since women are more likely to spend their already smaller incomes on basic household necessities and have less access to land and assets. Elson (1991) suggests that the “male bias” in the design of structural adjustment programmes lies at the core of this disproportionate negative impact on women.
One might argue that IMF-sponsored labour market reforms, by bringing greater flexibility to the market, can still be beneficial for women and offset the negative impact of privatizations, austerity measures and social spending cuts. This study, building on previous studies, investigates whether IMF-sponsored labour market reforms, as the IMF claims, can increase women’s labour force participation.
Women often face more barriers when entering the labour market compared to men, and face unequal conditions once employed (Shen 2021). It is well documented that women earn less than men overall, are more likely to be discriminated against by employers, spend more time doing unpaid care work and domestic work and often work for free in family businesses (Brettell 2012; Buenaventura and Miranda 2017; Iversen and Rosenbluth 2010; Lutz and Palenga-Mollenbeck 2012; Schafer and Gottschall 2015; Schmitz and Spiess 2022; Steiber and Haas 2012; Waylen 2022). They are traditionally disadvantaged in their labour force participation (Petit 2007).
The IMF suggests that its labour market reforms in borrowing countries remove the privilege of “insiders” (usually men in their prime working age) occupying well-protected and well-paid jobs and facilitate the access of “outsiders”, such as women, to the labour market (Duval and Loungani 2019). This is reflected as policy advice in Article IV Consultations and as conditionality in lending programmes. Indeed, the literature suggests that the dualization of the labour market in industrialized countries may result in a small group of privileged workers, while “outsiders” are relegated to low-security, low-pay and low-protection jobs (Hooijer and Picot 2015; Rueda 2014). Weakening the stronghold of “insiders” in the labour market might result in better outcomes for marginalized groups, such as women, young people and migrants (Duval and Loungani 2019). This arguably results in so-called “lousy jobs” for all, but still provides previously marginalized groups with greater access to available jobs (ILO 1999; Iversen and Rosenbluth 2010).
However, it is also likely that IMF-initiated labour market reforms do not result in greater labour force participation of women as a result of wage reductions. Reforms are often designed to lower unit labour costs and improve exports to bridge the current account deficit (i.e. the gap between exports and imports) (IMF 2009; IMF 2013). In other words, reform policies are often designed to lower wages (Vreeland 2002). This leads to a reduction in the minimum wage, as seen in Brazil in 2013, Chile in 2014 and Hungary in 2005, to name a few examples (Duval and Loungani 2019). Conditionality can also lead to the decentralization of collective bargaining institutions and encourage firm- or individual-level bargaining, which reduces wages by reducing the bargaining power of workers, as seen in Portugal in 2011 and Greece in 2010 (Koukiadaki and Kretsos 2012; IMF 2013). It can directly lower government wages, as seen in Albania in 1994 and Algeria in 1994 (Kentikelenis, Stubbs and King 2016). It can also indirectly lower wages by mandating the lay-off of civil servants and increasing competition in the labour market (Kentikelenis, Stubbs, and King 2016; Rickard and Caraway 2019).
When wages decline in the labour market, women and girls are likely to participate less in the labour force. Classic economic theory would stipulate that supply is reduced at lower price levels, but there seems to be a clear gender gap. Patriarchal societies seclude and isolate women, discouraging them from working outside the home. When wages increase, there are more incentives to relax this patriarchal control (Evans 2022), but when wages decrease, there are fewer incentives to forego this control, and women and girls might withdraw from the labour force. Wage decline might also discourage women from participating in the labour force if childcare costs exceed their wages. However, the negative impact of wages holds both for women who are of childbearing age and those who are not, as women beyond their 40s also seem to withdraw from the labour force when wages decline. Finally, women themselves might decide not to participate in the labour force, especially if it results in a “double burden” of work both inside and outside the home.
Declining firing costs, on the other hand, might have a positive impact on women’s labour force participation by increasing job turnover in the labour market in the short term. In a labour market where there is greater movement in and out of jobs, women and girls might find it easier to secure employment, all else being equal. However, while women may benefit from easier access to jobs in the short term, they may suffer from worse working conditions in the longer term and feel trapped in “lousy jobs” (Goos and Manning 2007). With lowered firing costs, women might never accumulate the tenure required to access decent jobs. Combined with lower wages, this may lead to “in-work poverty” (Gammarano 2019). We know from existing research that permanent employees earn more than their counterparts with temporary contracts (ILO 2015). Women transitioning from one job to another without the necessary social security might join the ranks of the “working poor” (Gammarano 2019).
3. Research design: Labour market outcomes for women and girls
Does IMF-sponsored labour market reform improve women’s labour force participation by facilitating their entry into the labour market, as the IMF claims? In order to investigate this question, I have gathered a dataset of 109 countries spanning the years between 1990 and 2014, focusing on labour conditions of the IMF. It is easier for the IMF to enforce conditionality than to enforce its policy advice in Article IV consultations, as such consultations are voluntary and are not tied to the disbursement of loans. The IMF’s impact on labour market reform, however, can clearly be observed through conditionality. The analysis begins in 1990, when the IMF became more active in labour market issues (see figure 1 depicting the relative frequency of labour conditions within the total conditions between 1980 and 2014), and ends in 2014 to include the soul-searching years at the IMF following the global financial crisis (Clift 2018). Some of the IMF’s ideas on (in)equality shifted during this period, with greater emphasis being placed on “country ownership” (Konstantinidis and Reinsberg 2023) and reducing inequality. It is too soon, however, to analyse whether this change has led to better gender outcomes in its programmes. The unit of analysis in this study is country-year, in line with standard practice in the field (Dreher 2006; Vreeland 2003).
Figure 1. Relative frequency of labour conditions in IMF programmes, 1980–2014
Source: Kentikelenis, Stubbs and King (2016) IMF conditionality dataset.
To measure the impact of IMF labour market reforms on women’s labour force participation, I implement a compound instrumental variable analysis over three simultaneous equations. Scholars have long warned that self-selection into IMF programmes is not random and that there are systematic commonalities among countries that borrow from the IMF, such as ongoing economic crises and political and institutional weaknesses, which might bias the results (Abouharb and Cingranelli 2007, 2009; Dang and Stone 2021; Stone 2008; Vreeland 2003). Similarly, developing countries borrow from the IMF more frequently than developed countries, and in developing countries, women’s labour force participation is traditionally lower. It is now common practice in the field to account for selection into programmes to address these concerns.
More recently, scholars such as Stubbs et al. (2020) and Vreeland (2007) have warned that selection into different branches of conditionality may also be non-random. Countries may negotiate or opt out of certain branches of conditionality due to the influence of powerful groups in domestic politics (Caraway, Rickard and Anner 2012; Nooruddin and Simmons 2006), for example, or request specific conditions in order to shift the blame for unpopular reforms onto the IMF and thereby bypass domestic opposition (Vreeland 2007). Conditionality and political and economic outcomes may therefore be endogenous. After accounting for these endogenous processes, we can then analyse how conditionality influences women’s labour force participation.
I follow Stubbs et al.’s (2020) compound instrumental variable design over a system of three simultaneous equations using maximum likelihood estimation. The modelling accounts for the possibility of endogenous selection into IMF programmes and labour conditions based on factors that would affect women’s labour force participation. For example, a strictly regulated market in domestic politics might lead a government to sign an IMF agreement to bypass labour opposition, prompt a higher number of labour conditions from the IMF and reduce the labour force participation of women due to the strong dualization of the labour market between secure and insecure jobs.
The first equation in the analysis looks at selection into IMF programmes. Here, the variable IMF participation is coded “1” if a country had an IMF programme for at least five months in a given year and “0” otherwise. Data are taken from Dreher, Sturm and Vreeland (2015). Based on a recent innovation in the literature, I instrumentalize programme participation via the interaction of average programme participation during the period of analysis (i.e. between 1990 and 2014) and the IMF’s budgetary constraints (Lang 2021; Nelson and Wallace 2017; Stubbs et al. 2020). As suggested earlier, the interaction of an endogenous variable (i.e. IMF programme participation) with an exogenous variable (e.g. the IMF’s budgetary constraints) can be interpreted as exogenous (Lang 2021; Nelson and Wallace 2017; Stubbs et al. 2020). It follows from the argument that the IMF’s resources are independent of any one borrower’s domestic political and economic processes (i.e. they are excludable) (Lang 2021). However, it predicts IMF programme participation. When the IMF is constrained in terms of its available budget, it may not conclude a conditional lending programme with the requesting country. The IMF’s budgetary constraint is measured by the ratio of its liquid resources, such as Special Drawing Rights contributions and the usable sum of currencies, to its liabilities, which are outstanding payments to borrowers and the IMF borrowing from its Members (Lang 2021; Nelson and Wallace 2017; Stubbs et al. 2020). Data on the IMF’s budgetary constraints are taken from Lang (2021). The IMF participation instrument has a statistically strong predictive capacity for participation in IMF programmes (p < 0.0001, n = 4,610) (see appendix 5 depicting the marginal impact of the instrument predicting IMF programme participation). It is lagged for one year in the analysis in order to account for the delay in the impact on women’s labour force participation.
I proxy labour market reform under an IMF programme via labour conditions assigned by the IMF. Labour conditions in the dataset include labour market flexibilization measures, such as making changes to collective bargaining institutions and reducing their coverage by supporting firm-level or individual-level bargaining; making firing and hiring easier by relaxing employment protection legislation and reducing severance payments and notice periods; reducing minimum wage; mandating public sector lay-offs; reducing the duration of and replacement ratio for unemployment benefits; and increasing the duration of maximum fixed-term contracts and the number of hours that can be worked on part-time contracts. Data on IMF conditionality are taken from Kentikelenis, Stubbs and King (2016) . Appendix 11 documents major categories of labour market reform under IMF programmes, building on raw data from Kentikelenis, Stubbs and King (2016), and provides examples for each category.
Labour market reform conditions include performance criteria, prior actions and benchmarks. In addition to a simple count of labour conditions (i.e. the sum of all labour market-related performance criteria, prior actions and benchmarks), I examine a weighted measure of labour conditions to assess the strictness of labour market reform in the programme (i.e. weighted labour conditions). In doing this, I follow the previous literature and assign a higher weight to prior actions and performance criteria, the fulfilment of which is indispensable for receiving IMF loans (Caraway, Rickard, and Anner 2012). Following Kentikelenis, Stubbs and King (2016), I assign double weight to these conditions. I assign a lower weight to benchmarks, since failure to meet them does not result in loan suspension, and add them as a simple count to the burden-of-adjustment measure (Kentikelenis, Stubbs and King 2016).3
To predict labour conditions, following Stubbs et al. (2020), once again, I examine the interaction of the average number of labour conditions for a country over the period of analysis and the IMF’s budget constraints. This is based on the argument that endogenous processes would be likely to generate similar lines of conditionality among repeat borrowers. Furthermore, the number of conditions would decrease when demand for IMF resources was lower (i.e. when liquidity was higher) and increase when IMF resources were scarce (i.e. when liquidity was lower) (Dreher and Vaubel 2004; Stubbs et al. 2020). This is because the IMF would be cautious about the use of its resources and apply more stringent controls when its resources were stretched (Dreher and Vaubel 2004). Once again, the interaction of an endogenous variable and an exogenous one would yield exogenous results under mild assumptions.4 The variable is lagged for one year, based on the assumption that the impact of labour market reforms would not be felt immediately. The instrument is statistically strongly associated with labour conditions in the dataset, even though the substantive significance is lower than the predictive capacity of the programme participation instrument. For robustness checks, I also run the analysis without the instruments, using ordinary least squares with fixed effects, as well as a treatment-effects model with country and year fixed effects (the results of which are reported in appendix 4).
Figure 1 shows that the share of labour conditions within the total number of conditions in programmes has changed over the years. While it gradually increased until 2005, it began declining as of 2006. This can be attributed to the establishment of the Independent Evaluation Office in 2001 and the strong criticism of IMF programmes during the Asian financial crisis in the late 1990s. The impact of this criticism seems to somewhat sluggishly lead to actual outcomes, reducing labour conditions in IMF programmes. To control for this potentially IMF-induced impact and other unobserved shocks common to all recipient countries, I add year fixed effects to the analysis. I also include country fixed effects to control for country-specific and time-invariant factors that may determine selection into IMF programmes, labour conditionality and women’s labour force participation.
Finally, in order to analyse the impact of labour market reform on women’s labour force participation, I look at several indicators. First, the measure of women’s labour force participation captures the proportion of girls and women who are over the age of 15 and either in employment or actively looking for work. Second, the indicator of girls and young women’s labour force participation captures the proportion of girls over the age of 15 and young women under the age of 24 who are currently in employment for pay or profit and engaged in the production of goods and services in the economy. Third, the sample is further disaggregated using indicators of labour force participation for women aged 20–44 and 45–65. Data for all indicators are taken from the ILO Key Indicators of the Labour Market dataset. I also look at several covariates of women’s labour force participation. All control variables are lagged for one year in the analysis to isolate their independent impact.
Earlier studies found that a shrinking economy and economic crisis negatively affect women’s labour force participation (Detraz and Peksen 2016; Erten and Metzger 2019; Kern, Reinsberg and Lee 2024; Mathers 2020). To isolate the impact of labour market reform that might coincide with output decline under IMF programmes, I control for gross domestic product (GDP) per capita and GDP. Both are log-transformed. Data are taken from the World Bank World Development Indicators (WDI) dataset.
Scholars have also identified that the fertility rate is a strong predictor of women’s labour force participation (Erten and Metzger 2019; Iversen and Rosenbluth 2010). It captures the average number of children that would be born to a woman of childbearing age based on country-specific and age-specific fertility rates for a given year. Data are taken from the WDI dataset.
Education might also affect women’s labour market participation. I measure this impact as a ratio of average years of total schooling for women over the age of 15 (considered the starting age for working) to that of the average years of total schooling for the entire population. Higher numbers indicate a higher level of education for women. Data are taken from the WDI dataset. Unfortunately, data availability for this measure is very poor, resulting in the exclusion of more than half of the sample in the analysis. To retain as many observations as possible and ensure robustness, I run the main model without this variable and add it for robustness checks. The results are reported in appendix 7.
Urbanization might foster women’s participation in the labour force by weakening the impact of traditional values on women’s employment and by providing greater opportunities, especially in the service sector and in export-oriented manufacturing centres (Erten and Metzger 2019). Alternatively, urbanization might reduce women’s economic activities, as women are more active in agriculture in rural settings. To capture the potential impact either way, I look at the total percentage of the population living in urban areas. Data are taken from the WDI dataset.
Formal gender equality rules and regulations might impact women’s employment and labour force participation prospects. Therefore, I include the World Bank’s Women, Business and the Law (WBL) index in the analysis. The index provides a measure of gender equality in terms of formal laws and economic opportunities that are comparable across countries. It is measured on an index of 100, with higher scores indicating greater gender equality. The WBL score for the Organisation for Economic Co-operation and Development (OECD) countries was 95.2 in 2020, whereas the score for the Middle East and North Africa region was 53 for the same year. Data are taken from the WDI dataset (summary statistics for all variables in the analysis are reported in appendix 2).
Finally, one could argue that IMF-sponsored labour market reforms do not cause a decrease in women’s labour force participation, but the lingering effects of pre-IMF labour market institutions in a borrowing country that might protect “insiders” and marginalize “outsiders”. (The IMF suggests that strong regulation in the labour market lowers the labour force participation of women and other marginalized groups.) To account for this argument, I add a regulated labour market variable in the analysis (the details of which appear in appendix 1). The variable accounts for how developed the labour market institutions were during the pre-IMF period. It is lagged an extra year vis-à-vis the IMF participation variable (i.e. it is lagged two years when IMF participation is lagged one year) to ensure that it is not affected by the IMF-sponsored labour market reform. Higher scores indicate a more regulated labour market. As a robustness check, I also control for two-year lagged firing costs instead of the composite regulated labour market variable, as one could argue that the main regulatory variable that influences women’s participation is that of pre-existing firing costs. The firing costs variable measures legally mandated compensation for redundancy. Higher numbers indicate greater legal compensation for redundancy. The data are taken from the Centre for Business Research Labour Regulation Index. The results are reported in appendix 9.
In addition, I examine two ways in which labour market reforms under IMF programmes might affect women’s labour force participation. First, IMF-induced changes to firing costs might affect women’s labour force participation by increasing job turnover in the labour market. Second, building on the existing literature, I examine whether declining wages might reduce the incentives for women to participate in the labour force (Winkler 2022), especially in patriarchal societies. Appendix 8 suggests that there is a positive association between wages and women’s labour force participation, supporting previous literature on this topic. The wage variable is measured as annual wage rates or earnings per worker employed. It is indexed and ranges from 0.016 to 220. It is based on the national sources, as reported to the IMF. Data are taken from the IMF International Financial Statistics database.
4. Empirical findings
4.1 Gendered impact of IMF labour market reform
The results of the econometric analysis are presented in table 1. The table shows that IMF-sponsored labour market reform not only does not have a positive impact on women’s labour force participation after controlling for relevant variables, but in fact has a negative impact. One additional labour condition is associated with a 0.082 percentage point decrease in women’s labour force participation, all else being equal (significant at the 5 per cent level). Given an average participation rate of 50.24 per cent in the sample, this effect corresponds to a 0.16 per cent decline in relative terms. Since the average labour force participation of women is even lower in developing countries, the negative impact is higher. For a country with a 30 per cent female labour force participation rate, the presence of five labour conditionality measures under IMF programmes is associated with a 0.41 percentage point decline in women’s labour force participation, equivalent to a 1.37 per cent relative decrease. In a country with 40 million women, for example, approximately 164,000 fewer women would participate in the labour force as a result of those five labour conditions.
Table 1. Women’s labour force participation and the IMF’s labour market reform
| Variables | (1) | (2) | (3) |
| Women’s labour force participation (% of female population) | IMF participation | Labour conditionality | |
| Lagged IMF participation | 1.560 (1.441) |
||
| Lagged labour conditions |
–0.0820** (0.0390) |
||
| Lagged GDP per capita (log) | –0.106* (0.0561) |
0.000307 (0.0306) |
0.00717 (0.0233) |
| Lagged GDP (log) | –2.390 (1.893) |
–1.509** (0.625) |
–1.409*** (0.484) |
| Lagged fertility rate | –0.0872 (0.700) |
–0.259 (0.238) |
–0.191 (0.173) |
| Lagged (2 yr) regulated labour | –0.428 (0.291) |
0.0310 (0.138) |
–0.0583 (0.103) |
| Lagged urban population | 0.0177 (0.110) |
0.0205 (0.0369) |
–0.0119 (0.0186) |
| Lagged WBL | 0.0766* (0.0393) |
–0.00284 (0.0117) |
0.0160 (0.0107) |
| Lagged IMF prog. instrument | –0.755 (0.500) |
||
| Labour conditions instrument | –0.973*** (0.315) |
||
| Constant | 72.00 (46.51) |
39.56*** (15.01) |
35.52*** (11.85) |
| Observations | 2 510 | 2 510 | 2 510 |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
-
Notes: Compound instrumental variable analysis over three simultaneous equations; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
One could perhaps argue that the positive impacts of labour market reforms are felt only in the long term and that one year might not suffice to feel such impacts. In order to capture longer term effects, I have also lagged the labour conditions variable for two or more years in the analysis. It is striking that the statistically significant negative impact of an additional labour condition persists for 13 years before losing its significance.
Based on the recent, convincing argument in the literature that coefficients of control variables are highly dependent on other co-variates and should not be interpreted unless all co-variates of those control variables are specified (Hünermund and Louw 2023), I do not interpret the control variables. In other words, the impact of the regulated labour market on women’s labour force participation should not be interpreted, as regulated labour market variables have other unspecified co-variates, and we cannot determine the isolated impact of the regulated labour market unless all of these co-variates are specified. For the same reason, I do not interpret column 2 (selection into IMF participation) or column 3 (selection into labour conditions), as these models are not specified to explain which countries enter IMF agreements or accept labour conditions. Rather, I control for the influence of variables that affect women’s labour force participation, such as the regulated labour market or the WBL index, in the process of self-selection into the IMF and labour conditions. This serves as a safeguard in case countries with higher WBL index scores are more or less likely to enter IMF agreements or agree to the IMF’s labour conditions.
Delving deeper into the causal mechanism of why IMF-sponsored labour market reforms harm women’s labour force participation, I conduct a causal mediation analysis with a structural equation model and check the causal pathway using two mediators – wage levels (measured using the lagged wage index) and firing costs – and estimate the direct effects of labour market reforms and IMF programme participation. The model includes country and year fixed effects, allowing me to isolate within-country variation between 1990 and 2014 and account for time-specific shocks and unobserved heterogeneity. The results are displayed in table 2.
Table 2 demonstrates that both mediators are strongly associated with women’s labour force participation. A higher wage index is positively and significantly associated with greater labour force participation of women, with a one-point increase in the wage index linked to a 0.06 percentage point rise in the labour force participation of women (p < 0.01). In contrast, higher firing costs are associated with significantly lower labour force participation: a one-point increase in the firing cost index corresponds to an 11.44 percentage point decrease in women’s labour force participation (p < 0.01). These results suggest that wages and firing costs strongly mediate women’s labour market inclusion under IMF-sponsored reforms.
Table 2. Causal mediation analysis on labour market reform, declining wages, firing costs and IMF programme participation
| Variables | (1) | (2) | (3) |
| Women’s labour force participation | Lagged wages | Lagged firing costs | |
| Lagged wages | 0.0602*** (0.0125) |
||
| Lagged firing costs | –11.44*** (0.918) |
||
| Lagged IMF participation | –1.194 (0.929) |
||
| Lagged labour conditions | –0.162 (0.168) |
–2.694*** (0.505) |
–0.0220*** (0.00681) |
| Constant | 50.55*** (1.088) |
77.11*** (1.188) |
0.422*** (0.0160) |
| Observations | 611 | 611 | 611 |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
-
Notes: Structural equation model with country and year fixed effects; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Furthermore, table 2 shows that IMF labour market reforms significantly reduce both the wage index (–2.694, p < 0.01) and firing costs (–0.022, p < 0.01). The direct effect of reforms on women’s labour force participation is negative (–0.162) but not statistically significant once the mediators are included. This suggests that IMF labour market reforms affect women’s labour force participation primarily through their effect on firing costs and wages, rather than directly. The estimated coefficient for IMF programme participation remains negative (–1.194) but statistically insignificant, suggesting that participation in an IMF programme alone does not independently affect women’s labour force participation after controlling for the mediators and the reform.
4.2 Robustness checks
I check the robustness of the findings with alternative model specifications, alternative dependent variables, additional causal mediation analysis, a placebo test, as well as a sensitivity test using multiple imputation analysis.
First, I run an ordinary least squares model with panel fixed effects, as well as a treatment effects model with selection into IMF programmes based on variables such as GDP, GDP per capita, current account balance (as a percentage of GDP) and recidivism (repeated participation in IMF programmes). The same covariates for employment and labour force participation of women apply (the results of which are reported in appendix 4). These additional models confirm that IMF labour conditions reduce women’s labour force participation. In fact, statistical significance and substantive impact increase.
Second, to isolate the effects of labour market reforms from other potentially adverse policy measures, I examine the impact of privatization and social policy conditionality. Both types of conditions are weighted in the same way as labour conditions – that is, greater weight is assigned to performance criteria and prior actions than to benchmarks. The data are taken from Kentikelenis and Stubbs (2023). The results (see appendix 6) show that neither privatization nor social policy conditions significantly predict a decline in women’s labour force participation, which strengthens the interpretation that labour market reforms drive the decline.
Third, checking for the placebo effect, I replace labour conditions with fiscal conditions in IMF programmes. Fiscal conditions do not predict lower labour force participation of women. Fourth, one could argue that labour conditions might encourage the employment of girls and young women between the ages of 15 and 24, more so than women in older age groups, through part-time and temporary work. Although IMF-sponsored labour market reforms seem to have a slightly positive impact for girls and young women in this age group, this impact does not reach statistical significance (see appendix 3).
One could also argue that men’s labour force participation is also negatively affected by IMF-sponsored labour market reforms and thus that such reforms do not have a gendered impact. However, while reforms seem to negatively influence men’s labour force participation, their impact is not statistically significant (p = 0.437). Furthermore, the substantive effect is much smaller than on women: one additional condition reduces men’s labour force participation by 0.03 percentage points, whereas one additional condition reduces women’s labour force participation by more than 0.08 percentage points (see appendix 10). It is also worth noting that the IMF claims to increase the labour force entry of marginalized groups, such as women. Even if men were negatively affected by reforms in a similar way, the evidence would still show that the IMF’s claim of helping marginalized groups through labour market flexibilization measures does not hold in the case of women.
Finally, to address the issue of missingness, particularly in employment-related indicators, I check the robustness of the findings by conducting a multiple imputation analysis. The results (see appendix 12) show that IMF-sponsored labour market reforms negatively influence women’s labour force participation in borrowing countries, even when this alternative estimation method is used. An additional labour condition is associated with a 0.111 percentage point decrease in women’s labour force participation, on average, holding other factors constant and accounting for country and year fixed effects.
There is, however, a certain commonality among countries lacking data in the sample, such as having significantly higher GDP levels and stronger legal rights for women, as measured by the WBL index. The results are therefore most representative of middle- and high-income countries and should be interpreted with caution when generalizing about lower-income or more fragile States. Future studies could draw on more comprehensive data to more accurately reflect the experiences of lower-income or institutionally weaker States.
5. Conclusion
This article has examined the IMF’s claim that its labour market reforms encourage the entry of traditionally excluded groups, such as women and young people, into the labour force through its flexibilization measures. I have argued that, contrary to this claim, IMF-sponsored labour market reforms are associated with lower labour force participation of women overall. In particular, if reforms lower wages, they negatively affect women’s labour force participation. This may reflect patriarchal constraints, whereby male relatives restrict women’s work outside the home when wages are low (Evans 2022), or women’s own views that lower wages are not worth the “double burden” of working both inside and outside the home (Hochschild and Machung 2012). Lower firing costs can, on the other hand, temporarily increase women’s access to employment. This study contributes to the existing scholarship by looking at the gendered consequences of IMF-sponsored labour market reforms and identifying the mechanisms, i.e. lower wages and firing costs, that affect women’s participation.
Future studies should carefully examine the long-term consequences of lowering firing costs, as a less regulated market might discourage women’s labour force participation in the long run by driving wages down. Future research should also further explore the mechanisms linking lower wages to women’s labour force participation, distinguishing between patriarchal control and women’s voluntary decisions. Qualitative approaches, such as interviews and focus groups, could shed light on these issues.
Finally, there has been a recent effort by the IMF to mainstream gender in its lending programmes, surveillance and capacity development. Future studies should analyse the impact of its Gender Team, established in 2022, on the gendered impact of IMF programmes in general and labour market reforms in particular. Under the supervision of this team, in the future, the IMF might design labour market reforms to tackle gender inequalities in the labour market that might stem from patriarchy and the double burden.
Notes
- The IMF holds regular consultations on the macroeconomic indicators of Member States in its surveillance and monitoring capacity. ⮭
- The long-term impact of lowered firing costs, however, should be approached with caution. Please refer to the discussion of the potential consequences of long-term employment in temporary jobs in section 2. ⮭
- Weighted labour market reform measure for country j in year t can be written as follows: weighted labour market reform j, t = [(prior actions × 2) + (performance criteria × 2) + benchmarks]. ⮭
- For further information, see Stubbs et al. (2020). ⮭
- The coefficient on lambda in model 3 is –0.624 (p < 0.01), which is negative and statistically significant. This suggests that countries with a higher predicted probability of entering an IMF programme tend to have lower labour force participation of women than we would expect if selection into IMF programmes were random. In other words, the negative lambda indicates that selection into IMF programmes is non-random and related to unobserved factors that also negatively affect women’s labour force participation. ⮭
Acknowledgments
I would like to thank Anna Minasyan and Bernhard Reinsberg, the participants of the Academic Council on the United Nations System workshop held in London in June 2022 (especially Rod Abouharb and Inken von Borzyskowski), the participants of the International Political Economy Colloquium at the University of Groningen (Donya Ahmadi, Malcolm Campbell-Verduyn, Gregory Fuller, Herman Hoen, Lukas Linsi and Caitlin Ryan), as well as the participants of the European and Global Governance Colloquium at Erasmus University Rotterdam (Geske Dijkstra, Clara Egger, Markus Haverland, Michal Onderco, Stefano Scibilia, Pieter Tuytens and Asya Zhelyazkova), for their excellent comments and suggestions on earlier drafts of this article.
Competing interests
The author declares that they have no competing interests.
References
Abouharb, M. Rodwan, and David Cingranelli. 2007. Human Rights and Structural Adjustment. Cambridge: Cambridge University Press.
Abouharb, M. Rodwan, and David Cingranelli. 2009. “IMF Programs and Human Rights, 1981–2003”. The Review of International Organizations 4 (1): 47–72. http://doi.org/10.1007/s11558-008-9050-5.
Anner, Mark, and Teri Caraway. 2010. “International Institutions and Workers’ Rights: Between Labor Standards and Market Flexibility”. Studies in Comparative International Development 45 (2): 151–169. http://doi.org/10.1007/s12116-010-9064-x.
Brettell, Caroline B. 2012. “Gender, Family, and Migration”. In The Oxford Handbook of the Politics of International Migration, edited by Marc R. Rosenblum and Daniel J. Tichenor, 478–508. New York, NY: Oxford University Press.
Buenaventura, Mae, and Claire Miranda. 2017. “The IMF and Gender Equality: The Gender Dimensions of the IMF’s Key Fiscal Policy Advice on Resource Mobilisation in Developing Countries”. London: Bretton Woods Project.
Caraway, Teri L. 2006. “Freedom of Association: Battering Ram or Trojan Horse?” Review of International Political Economy 13 (2): 210–232. http://doi.org/10.1080/09692290600625462.
Caraway, Teri L., Stephanie J. Rickard, and Mark S. Anner. 2012. “International Negotiations and Domestic Politics: The Case of IMF Labor Market Conditionality”. International Organization 66 (1): 27–61. http://doi.org/10.1017/S0020818311000348.
Clift, Ben. 2018. “Ideational Change at the IMF after the Crash” In The IMF and the Politics of Austerity in the Wake of the Global Financial Crisis, edited by Ben Clift, 27–55. Oxford: Oxford University Press.
Dang, Trung A., and Randall W. Stone. 2021. “Multinational Banks and IMF Conditionality”. International Studies Quarterly 65 (2): 375–386. http://doi.org/10.1093/isq/sqab010.
Detraz, Nicole, and Dursun Peksen. 2016. “The Effect of IMF Programs on Women’s Economic and Political Rights”. International Interactions 42 (1): 81–105. http://doi.org/10.1080/03050629.2015.1056343.
Dreher, Axel. 2006. “IMF and Economic Growth: The Effects of Programs, Loans, and Compliance with Conditionality”. World Development 34 (5): 769–788. http://doi.org/10.1016/j.worlddev.2005.11.002.
Dreher, Axel, Jan-Egbert Sturm, and James Raymond Vreeland. 2015. “Politics and IMF Conditionality”. Journal of Conflict Resolution 59 (1): 120–148. http://doi.org/10.1177/0022002713499723.
Dreher, Axel, and Roland Vaubel. 2004. “Do IMF and IBRD Cause Moral Hazard and Political Business Cycles? Evidence from Panel Data”. Open Economies Review 15 (1): 5–22. http://doi.org/10.1023/B:OPEN.0000009422.66952.4b.
Duval, Romain, and Prakash Loungani. 2019. “Designing Labor Market Institutions in Emerging Market and Developing Economies: Evidence and Policy Options”. IMF Staff Discussion Note 19/04. Washington, DC: IMF.
Elson, Diane. 1991. “Male Bias in Macroeconomics: The Case of Structural Adjustment”. In Male Bias in the Development Process, edited by Diane Elson, 164–190. Manchester: Manchester University Press.
Erten, Bilge, and Martina Metzger. 2019. “The Real Exchange Rate, Structural Change, and Female Labor Force Participation”. World Development 117 (May): 296–312. http://doi.org/10.1016/j.worlddev.2019.01.015.
Evans, Alice. 2022. “The Middle East and North Africa’s Patrilineal Trap”. Brookings Institution. 25 January 2022. https://www.brookings.edu/blog/future-development/2022/01/25/the-middle-east-and-north-africas-patrilineal-trap/.
Forster, Timon, Alexander E. Kentikelenis, Bernhard Reinsberg, Thomas H. Stubbs, and Lawrence P. King. 2019. “How Structural Adjustment Programs Affect Inequality: A Disaggregated Analysis of IMF Conditionality, 1980–2014”. Social Science Research 80 (May): 83–113. http://doi.org/10.1016/j.ssresearch.2019.01.001.
Franklin, James. 1997. “IMF Conditionality, Threat Perception, and Political Repression: A Cross-National Analysis”. Comparative Political Studies 30 (5): 576–606. http://doi.org/10.1177/0010414097030005003.
Gammarano, Rosina. 2019. “The Working Poor or How a Job is No Guarantee of Decent Living Conditions: A Study Based on ILO’s Global Estimates of Employment by Economic Class”. Spotlight on Work Statistics, No. 6. Geneva: ILO.
Garuda, Gopal. 2000. “The Distributional Effects of IMF Programs: A Cross-Country Analysis”. World Development 28 (6): 1031–1051. http://doi.org/10.1016/S0305-750X(00)00002-4.
Goos, Maarten, and Alan Manning. 2007. “Lousy and Lovely Jobs: The Rising Polarization of Work in Britain”. The Review of Economics and Statistics 89 (1): 118–133. http://doi.org/10.1162/rest.89.1.118.
Hajro, Zlata, and Joseph P. Joyce. 2009. “A True Test: Do IMF Programs Hurt the Poor?” Applied Economics 41 (3): 295–306. http://doi.org/10.1080/00036840601007229.
Hochschild, Arlie, and Anne Machung. 2012. The Second Shift: Working Families and the Revolution at Home. New York, NY: Penguin Books.
Hooijer, Gerda, and Georg Picot. 2015. “European Welfare States and Migrant Poverty: The Institutional Determinants of Disadvantage”. Comparative Political Studies 48 (14): 1879–1904. http://doi.org/10.1177/0010414015597508.
Hünermund, Paul, and Beyers Louw. 2023. “On the Nuisance of Control Variables in Causal Regression Analysis”. Organizational Research Methods 28 (1), 138–151. http://doi.org/10.1177/10944281231219274.
ILO. 1999. Decent Work: Report of the Director-General. International Labour Conference, 87th Session. Geneva.
ILO. 2015. “Employment Patterns, Poverty and Income Inequality”. World Employment and Social Outlook 2015: The Changing Nature of Jobs, 39–71. Geneva.
IMF (International Monetary Fund). 2000. “Address by Margaret Kelly to the Ad Hoc Committee of the Whole, Special Session of the General Assembly, Women 2000: Gender Equality, Development and Peace for the Twenty-first Century”. 5 June 2000. https://www.imf.org/en/News/Articles/2015/09/28/04/53/sp060500a.
IMF (International Monetary Fund). 2006. “Chile: 2006 Article IV Consultation – Staff Report; Staff Supplement; Public Information Notice on the Executive Board Discussion; and Statement by the Executive Director for Chile”. IMF Country Report No. 06/335. Washington, DC.
IMF (International Monetary Fund). 2009. “Greece: 2009 Article IV Consultation – Staff Report; Staff Supplement; Public Information Notice on the Executive Board Discussion; and Statement by the Executive Director for Greece”. IMF Country Report No. 09/244. Washington, DC.
IMF (International Monetary Fund). 2013. “Portugal: Letter of Intent, Memorandum of Economic and Financial Policies, and Technical Memorandum of Understanding”. 24 October 2013. Washington, DC.
Iversen, Torben, and Frances Rosenbluth. 2010. Women, Work, and Politics: The Political Economy of Gender Inequality. New Haven, CT: Yale University Press.
Kentikelenis, Alexandros, and Thomas Stubbs. 2023. A Thousand Cuts: Social Protection in the Age of Austerity. New York, NY: Oxford University Press.
Kentikelenis, Alexander E., Thomas H. Stubbs, and Lawrence P. King. 2016. “IMF Conditionality and Development Policy Space, 1985–2014”. Review of International Political Economy 23 (4): 543–582. http://doi.org/10.1080/09692290.2016.1174953.
Kern, Andreas, Bernhard Reinsberg, and Claire Lee. 2024. “The Unintended Consequences of IMF Programs: Women Left behind in the Labor Market”. The Review of International Organizations 20 (4): 741–767. http://doi.org/10.1007/s11558-024-09542-7.
Konstantinidis, Nikitas, and Bernhard Reinsberg. 2023. “Government Ownership of IMF Conditionality Programs: Conceptualization, Measurement, and Validation”. International Studies Quarterly 67 (4): sqad074. http://doi.org/10.1093/isq/sqad074.
Koukiadaki, Aristea, and Lefteris Kretsos. 2012. “Opening Pandora’s Box: The Sovereign Debt Crisis and Labour Market Regulation in Greece”. Industrial Law Journal 41 (3): 276–304. http://doi.org/10.1093/indlaw/dws031.
Lang, Valentin. 2021. “The Economics of the Democratic Deficit: The Effect of IMF Programs on Inequality”. The Review of International Organizations 16 (3): 599–623. http://doi.org/10.1007/s11558-020-09405-x.
Lutz, Helma, and Ewa Palenga-Möllenbeck. 2012. “Care Workers, Care Drain, and Care Chains: Reflections on Care, Migration, and Citizenship”. Social Politics: International Studies in Gender, State & Society 19 (1): 15–37. http://doi.org/10.1093/sp/jxr026.
Marphatia, Akanksha A. 2010. “The Adverse Effects of International Monetary Fund Programs on the Health and Education Workforce”. International Journal of Social Determinants of Health and Health Services 40 (1): 165–178. http://doi.org/10.2190/HS.40.1.j
Mathers, Jennifer G. 2020. “Women, War and Austerity: IFIs and the Construction of Gendered Economic Insecurities in Ukraine”. Review of International Political Economy 27 (6): 1235–1256. http://doi.org/10.1080/09692290.2020.1725903.
Metinsoy, Saliha. 2024. Aggrieved Labor Strikes Back: Inter-Sectoral Labor Mobility, Conditionality, and Unrest under IMF Programs. Cambridge: Cambridge University Press.
Metinsoy, Saliha. 2025. “Who Adjusts? Exchange Rate Regimes and Finance versus Labor under IMF Programs”. The Review of International Organizations 20 (3): 631–668. http://doi.org/10.1007/s11558-024-09540-9.
Mukherjee, Bumba, and Vineeta Yadav. 2024. The IMF, Financial Crisis, and Repression of Human Rights. Cham: Springer.
Nelson, Stephen C., and Geoffrey P. R. Wallace. 2017. “Are IMF Lending Programs Good or Bad for Democracy?” The Review of International Organizations 12 (4): 523–558. http://doi.org/10.1007/s11558-016-9250-3.
Nooruddin, Irfan, and Joel W. Simmons. 2006. “The Politics of Hard Choices: IMF Programs and Government Spending”. International Organization 60 (4): 1001–1033. http://doi.org/10.1017/S0020818306060334.
Oberdabernig, Doris A. 2013. “Revisiting the Effects of IMF Programs on Poverty and Inequality”. World Development 46 (June): 113–142. http://doi.org/10.1016/j.worlddev.2013.01.033.
Ortiz, Isabel, and Matthew Cummins. 2013. “Austerity Measures in Developing Countries: Public Expenditure Trends and the Risks to Children and Women”. Feminist Economics 19 (3): 55–81. http://doi.org/10.1080/13545701.2013.791027.
Ortiz, Isabel, and Matthew Cummins. 2021. “The Austerity Decade 2010-20”. Social Policy and Society 20 (1): 142–157. http://doi.org/10.1017/S1474746420000433.
Petit, Pascale. 2007. “The Effects of Age and Family Constraints on Gender Hiring Discrimination: A Field Experiment in the French Financial Sector”. Labour Economics 14 (3): 371–391. http://doi.org/10.1016/j.labeco.2006.01.006.
Pion-Berlin, David. 1983. “Political Repression and Economic Doctrines: The Case of Argentina”. Comparative Political Studies 16 (1): 37–66. http://doi.org/10.1177/0010414083016001002.
Przeworski, Adam, and James Raymond Vreeland. 2000. “The Effect of IMF Programs on Economic Growth”. Journal of Development Economics 62 (2): 385–421. http://doi.org/10.1016/S0304-3878(00)00090-0.
Reinsberg, Bernhard, Alexander Kentikelenis, and Thomas Stubbs. 2021. “Creating Crony Capitalism: Neoliberal Globalization and the Fueling of Corruption”. Socio-Economic Review 19 (2): 607–634. http://doi.org/10.1093/ser/mwz039.
Reinsberg, Bernhard, Andreas Kern, Mirko Heinzel, and Saliha Metinsoy. 2024. “Women’s Leadership and the Gendered Consequences of Austerity in the Public Sector: Evidence from IMF Programs”. Governance 37 (1): 303–321. http://doi.org/10.1111/gove.12764.
Reinsberg, Bernhard, Thomas Stubbs, and Louis Bujnoch. 2023. “Structural Adjustment, Alienation, and Mass Protest”. Social Science Research 109 (January): 102777. http://doi.org/10.1016/j.ssresearch.2022.102777.
Reinsberg, Bernhard, Thomas Stubbs, Alexander Kentikelenis, and Lawrence King. 2019. “The Political Economy of Labor Market Deregulation during IMF Interventions”. International Interactions 45 (3): 532–559. http://doi.org/10.1080/03050629.2019.1582531.
Rickard, Stephanie J., and Teri L. Caraway. 2019. “International Demands for Austerity: Examining the Impact of the IMF on the Public Sector”. The Review of International Organizations 14 (1): 35–57. http://doi.org/10.1007/s11558-017-9295-y.
Rueda, David. 2014. “Dualization, Crisis and the Welfare State”. Socio-Economic Review 12 (2): 381–407. http://doi.org/10.1093/ser/mwu015.
Schäfer, Andrea, and Karin Gottschall. 2015. “From Wage Regulation to Wage Gap: How Wage-Setting Institutions and Structures Shape the Gender Wage Gap across Three Industries in 24 European Countries and Germany”. Cambridge Journal of Economics 39 (2): 467–496. http://doi.org/10.1093/cje/bev005.
Schmitz, Sophia, and C. Katharina Spiess. 2022. “The Intergenerational Transmission of Gender Norms—Why and How Adolescent Males with Working Mothers Matter for Female Labour Market Outcomes”. Socio-Economic Review 20 (1): 281–322. http://doi.org/10.1093/ser/mwab023.
Shen, Kailing. 2021. “Gender Discrimination”. IZA Discussion Paper No. 14897. Bonn: Institute of Labor Economics.
Steiber, Nadia, and Barbara Haas. 2012. “Advances in Explaining Women’s Employment Patterns”. Socio-Economic Review 10 (2): 343–367. http://doi.org/10.1093/ser/mwr039.
Stone, Randall W. 2004. “The Political Economy of IMF Lending in Africa”. American Political Science Review 98 (4): 577–591. http://doi.org/10.1017/S000305540404136X.
Stone, Randall W. 2008. “The Scope of IMF Conditionality”. International Organization 62 (4): 589–620. http://doi.org/10.1017/S0020818308080211.
Stubbs, Thomas, Bernhard Reinsberg, Alexander Kentikelenis, and Lawrence King. 2020. “How to Evaluate the Effects of IMF Conditionality: An Extension of Quantitative Approaches and an Empirical Application to Public Education Spending”. The Review of International Organizations 15 (1): 29–73. http://doi.org/10.1007/s11558-018-9332-5.
Vreeland, James Raymond. 2002. “The Effect of IMF Programs on Labor”. World Development 30 (1): 121–139. http://doi.org/10.1016/S0305-750X(01)00101-2.
Vreeland, James Raymond. 2003. The IMF and Economic Development. Cambridge: Cambridge University Press.
Vreeland, James Raymond. 2007. The International Monetary Fund: Politics of Conditional Lending. New York: Routledge.
Waylen, Georgina. 2022. “Gendering Global Economic Governance after the Global Financial Crisis”. Review of International Political Economy 29 (4): 1007–2106. http://doi.org/10.1080/09692290.2021.1888142.
Winkler, Anne E. 2022. “Women’s Labor Force Participation”. IZA World of Labor Article, 289v2.
Appendix 1. Indicators included in the regulated labour market variable
The maximum duration of fixed-term contracts
Overtime premia
Limits to overtime working
Maximum daily working time
Legally mandated notice period
Legally mandated redundancy compensation
Minimum qualifying period of service for the normal case of unjust dismissal
Law imposes substantive constraints on dismissal
Extension of collective agreements
Lockouts (equals 1 if lockouts are not permitted; equals 0 if they are permitted)
Appendix 2. Summary statistics
Summary statistics for table 1
| Variable | Obs | Mean | Std Dev | Min | Max |
| Female labour force pa. | 2 510 | 49.798 | 14.558 | 10.041 | 85.03 |
| IMF participation | 2 508 | .303 | .46 | 0 | 1 |
| Labour conditions | 2 510 | .562 | 2.001 | 0 | 26 |
| GDP pc (logged) | 2 510 | 8.261 | 1.667 | 1.792 | 11.861 |
| GDP (logged) | 2 510 | 25.789 | 1.728 | 21.207 | 30.535 |
| Fertility rate total | 2 510 | 2.885 | 1.549 | 1.078 | 7.208 |
| Regulated labour mar. | 2 447 | 5.259 | 1.407 | 1.21 | 7.8 |
| Urban population | 2 510 | 58.408 | 22.576 | 8.49 | 100 |
| WBL | 2 510 | 67.049 | 18.312 | 22.5 | 100 |
Summary statistics for table 2
| Variable | Obs | Mean | Std Dev | Min | Max |
| Wage index | 611 | 75.737 | 29.356 | .024 | 186.795 |
| Firing costs | 611 | .411 | .39 | 0 | 1 |
Descriptive statistics for appendix 10
| Variable | Obs | Mean | Std Dev | Min | Max |
| Male labour force part. | 2 421 | 73.609 | 8.258 | 48.749 | 96.067 |
Descriptive statistics for appendix 4 treatment effects model
| Variable | Obs | Mean | Std Dev | Min | Max |
| GDP pc (logged) | 2 067 | 7.117e+11 | 1.843e+12 | 1.761e+09 | 1.781e+13 |
| GDP (logged) | 2 067 | 10335.153 | 13681.344 | 15 | 141635 |
| Current acco. (% GDP) | 2 067 | 0 | 0 | 0 | 0 |
| Recidivism | 2 067 | 1.588 | 1.918 | 0 | 5 |
Source: Own calculations.
Appendix 3. Girls’ and young women’s labour force participation
| Variables | (1) | (2) | (3) |
| Girls’ and young women’s (aged 15–24) labour force participation | IMF participation | Labour conditionality | |
| Lagged IMF participation | 1.680 (1.081) |
||
| Lagged labour conditions |
0.0157 (0.0814) |
||
| Lagged IMF instrument | –0.601 (0.437) |
||
| Lagged GDP per capita | 0.133 (0.107) |
0.00134 (0.0310) |
0.00639 (0.0230) |
| Lagged GDP | 2.374 (2.648) |
–1.751*** (0.554) |
–1.402*** (0.483) |
| Lagged fertility rate | –0.259 (1.282) |
–0.217 (0.227) |
–0.184 (0.170) |
| Lagged (2 yr) regulated mark. | 0.377 (0.570) |
0.0294 (0.142) |
–0.0605 (0.103) |
| Lagged urban population | 0.0706 (0.157) |
0.00897 (0.0328) |
–0.0113 (0.0184) |
| Lagged WBL | –0.0558 (0.0574) |
–0.00264 (0.0115) |
0.0159 (0.0107) |
| Lagged labour conditions inst. | –0.916*** (0.287) |
||
| Constant | –29.89 (67.69) |
45.09*** (13.59) |
35.32 (11.84) |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
| Observations | 2 510 | 2 510 | 2 510 |
-
Notes: Instrumental variable analysis over three simultaneous equations; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 4. Robustness checks: OLS with fixed effects and treatment effects model
Table A4.1. OLS regression with panel fixed effects
| Variables | (1) |
| Women’s labour force participation (% female population) | |
| Lagged labour conditions |
–0.0979*** (0.0355) |
| Lagged urban population | 0.0726*** (0.0249) |
| Lagged fertility rate | –0.698*** (0.193) |
| Lagged WBL | 0.128*** (0.0101) |
| Lagged (2 yr) regulated labour market | –0.365*** (0.124) |
| Lagged IMF participation | 0.540*** (0.177) |
| Lagged GDP per capita (log) | –0.0863* (0.0468) |
| Lagged GDP (log) | –0.842*** (0.322) |
| Constant | 63.32*** (7.720) |
| Observations | 2 617 |
| Number of countries | 109 |
| R-squared | 0.121 |
| Country fixed effects | YES |
| Year fixed effects | YES |
-
Notes: OLS regression for panel data with country and year fixed effects; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Table A4.2. Treatment effects model and women’s labour force participation
| Variables | (1) | (2) | (3) |
| Women’s labour force participation (% female population) | IMF participation | Hazard | |
| Lagged labour conditions |
–0.0956*** (0.0350) |
||
| Lagged fertility rate | –0.734*** (0.234) |
||
| Lagged WBL | 0.0498*** (0.0123) |
||
| Lagged regulated labour mar. | –0.422*** (0.143) |
||
| Lagged IMF participation | 1.183*** (0.279) |
||
| Lagged GDP per capita (log) | –1.68e–05 (1.08e–05) |
||
| Lagged GDP (log) | –0* (0) |
||
| Lagged current account (% GDP) | 9.054e+07 (7.796e+07) |
||
| Lagged recidivism | 0.717*** (0.0437) |
||
| Lambda5 | –0.624*** (0.194) |
||
| Constant | 17.61*** (1.890) |
–1.683** (0.669) |
|
| Observations | 2 067 | 2 067 | 2 067 |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
-
Notes: Treatment effects model with country and year fixed effects in both the treatment and outcome equations, using two-step estimation with robust standard errors; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 5. IMF programme participation instrument and labour conditions instrument
Source: Own calculations.
Appendix 6. Privatization conditions, social policy conditions and women’s labour force participation
| Variables | (1) | (2) |
| Women’s labour force participation (% female population) | Women’s labour force participation (% female population) | |
| Lagged privatization conditions | 0.0451 (0.0417) |
|
| Lagged social policy conditions | –0.0471 (0.0762) |
|
| Lagged urban population | 0.0225 (0.0253) |
0.0237 (0.0253) |
| Lagged fertility rate | –0.0937 (0.206) |
–0.0969 (0.206) |
| Lagged WBL | 0.0738*** (0.0118) |
0.0741*** (0.0118) |
| Lagged (2 yr) regulated labour ma. | –0.389*** (0.123) |
–0.383*** (0.123) |
| Lagged IMF participation | 0.279 (0.172) |
0.336** (0.170) |
| Lagged GDP per capita (log) | –0.109** (0.0464) |
–0.107** (0.0464) |
| Lagged GDP (log) | –2.816*** (0.419) |
–2.875*** (0.419) |
| Constant | 116.2*** (10.65) |
81.91*** (10.28) |
| Observations | 2 617 | 2 617 |
| R-squared | 0.149 | |
| Number of countries | 109 | 109 |
| Country fixed effects | YES | YES |
| Year fixed effects | YES | YES |
-
Notes: OLS regression for panel data with fixed effects; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 7. Labour market reform and women’s labour force participation with the education variable added
| Variables | (1) | (2) | (3) |
| Women’s labour force participation | IMF programme participation | Labour conditionality | |
| Lagged IMF participation | –0.335 (3.367) |
||
| Lagged weighted labour conditions | –0.285** (0.118) |
||
| Lagged IMF participation instrument | –0.181 (1.762) |
||
| Lagged female education (% total) | –0.712 (1.162) |
3.112 (2.365) |
0.327 (0.224) |
| Lagged GDP per capita | –0.169 (0.158) |
0.131 (0.102) |
–0.0879 (0.0798) |
| Lagged GDP | –1.171 (3.140) |
–3.074 (2.154) |
–0.415 (0.634) |
| Lagged fertility rate | 0.600 (0.885) |
0.0614 (0.629) |
–0.269 (0.313) |
| Lagged firing costs | –1.338 (1.546) |
0.390 (1.397) |
–0.490 (0.654) |
| Lagged urban population | –0.00835 (0.137) |
0.0148 (0.120) |
–0.0529 (0.0483) |
| Lagged WBL | 0.0697 (0.0477) |
0.00118 (0.0233) |
0.0389* (0.0214) |
| Lagged labour condition instrument | –0.388 (0.676) |
||
| Constant | 40.53 (77.81) |
73.25 (51.87) |
12.56 (15.47) |
| Observations | 393 | 393 | 393 |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
-
Notes: Robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 8. Positive association between wage increase and women’s labour force participation
| Variables | Women’s labour force participation |
| Lagged labour market wages percentage change |
0.0436*** (0.00531) |
| Lagged GDP per capita (log) | –0.115 (0.0831) |
| Lagged GDP (log) | 3.311*** (0.946) |
| Lagged fertility rate | 0.787 (0.543) |
| Lagged firing costs | –5.683*** (0.864) |
| Lagged urban population | 0.263*** (0.0533) |
| Lagged WBL | 0.150*** (0.0255) |
| Constant | –65.69*** (23.43) |
| Observations | 579 |
| Number of countries | 33 |
| R-squared | 0.323 |
| Country fixed effects | YES |
| Year fixed effects | YES |
-
Notes: OLS regression for panel data with fixed effects; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 9. Impact of IMF labour market reform on women’s labour force participation – model with firing costs
| Variables | (1) | (2) | (3) |
| Women’s labour force participation | Lagged IMF participation | Labour conditions | |
| Lagged IMF participation | 1.540 (1.448) |
||
| Lagged labour conditions |
–0.0826** (0.0396) |
||
| Lagged GDP per capita | –0.107* (0.0581) |
0.00286 (0.0300) |
0.00962 (0.0249) |
| Lagged GDP | –2.235 (1.928) |
–1.505** (0.646) |
–1.439*** (0.472) |
| Lagged fertility rate | –0.00963 (0.699) |
–0.234 (0.243) |
–0.167 (0.173) |
| Lagged firing costs | –2.340** (1.118) |
0.438 (0.643) |
–0.0295 (0.424) |
| Lagged urban population | 0.0117 (0.111) |
0.0198 (0.0380) |
–0.0104 (0.0186) |
| Lagged WBL | 0.0727* (0.0383) |
–0.00266 (0.0117) |
0.0155 (0.0104) |
| Lagged IMF participation instru. | –0.700 (0.509) |
||
| Lagged labour conditions instru. | –0.967*** (0.315) |
||
| Constant | 67.66 (47.62) |
39.16** (15.38) |
35.91*** (11.49) |
| Observations | 2 528 | 2 528 | 2 528 |
-
Notes: Instrumental variable analysis over three simultaneous equations; robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 10. Impact of IMF labour market reform on men’s labour force participation – model without firing costs
| Variables | (1) | (2) | (3) |
| Men’s labour force participation (% male population) | IMF participation | Labour conditionality | |
| Lagged IMF participation | –0.0400 (0.355) |
||
| Lagged labour conditions |
–0.0293 (0.0377) |
||
| Lagged GDP per capita (log) | 0.00302 (0.0292) |
0.000808 (0.0310) |
0.00658 (0.0231) |
| Lagged GDP (log) | 0.614 (1.353) |
–1.744*** (0.550) |
–1.404*** (0.483) |
| Lagged fertility rate | –0.130 (0.565) |
–0.215 (0.226) |
–0.185 (0.171) |
| Lagged (2 yr) regulat. labour | –0.347 (0.259) |
0.0243 (0.143) |
–0.0599 (0.103) |
| Lagged urban population | –0.0485 (0.0773) |
0.00915 (0.0326) |
–0.0115 (0.0184) |
| Lagged WBL | 0.0202 (0.0213) |
–0.00280 (0.0114) |
0.0159 (0.0107) |
| Lagged IMF prog. instrument | –0.620 (0.438) |
||
| Labour conditions instrument | –0.929*** (0.296) |
||
| Constant | 65.26** (33.23) |
44.95*** (13.49) |
35.37*** (11.84) |
| Observations | 2 510 | 2 510 | 2 510 |
| Country fixed effects | YES | YES | YES |
| Year fixed effects | YES | YES | YES |
-
Notes: Robust standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.
Appendix 11. Labour market reform categories
| Labour market reform category | Description and examples |
| Civil service lay-offs | Staff cuts, voluntary/assisted retirements, reduction targets in civil service |
| Wage bill control and pay reform | Wage ceilings, wage freezes, bonus elimination, salary caps, new pay grids, performance evaluations, merit-based pay, pay-for-performance schemes |
| Pension system reform | Changes to the pension system and payments |
| Social security and family allowances | Cuts in social security transfers, subsidies, child/family benefits |
| Hiring controls and freezes | Hiring suspensions, limits to education/health staff recruitment |
| Collective bargaining and tripartite dialogue | Changes to collective bargaining institutions, wage councils, employer-employee negotiations, social pacts |
| Hiring and hiring flexibilization | Fixed-term contracts, market flexibility, changes to hiring and firing regulations |
Appendix 12. Multiple imputation analysis for women’s labour force participation
| Variables | Women’s labour force participation (% female population) |
| Lagged IMF participation | 0.489** (0.177) |
| Lagged labour conditions |
–0.111*** (0.035) |
| Lagged GDP per capita (log) | –0.087* (0.047) |
| Lagged GDP (log) | –0.792** (0.329) |
| Lagged fertility rate | –0.702*** (0.196) |
| Lagged (2 yr) regulated labour market | –0.388*** (0.123) |
| Lagged urban population | 0.074*** (0.026) |
| Lagged WBL | 0.125*** (0.010) |
| Constant | 62.340*** (7.891) |
| Observations | 2 508 |
| Country fixed effects | YES |
| Year fixed effects | YES |
-
Notes: Standard errors in parentheses; *, ** and *** indicate statistical significance at the 10, 5 and 1 per cent levels, respectively.
Source: Own calculations.