The global cost of gender inequality in labour market opportunities

This article first appeared in VoxDev

Despite important progress, persistent gender gaps in labour market outcomes remain. Globally, women have lower labour force participation and earnings, engage more often in unpaid or informal work, and spend more time on family care and household work. While a neoclassical approach may view these disparities as efficient labour allocations stemming from men and women’s different comparative advantages, an alternative view is that they reflect gender-specific frictions in labour markets.

Gender gaps and economic growth

If gender gaps result from distortions that limit women’s access to good jobs, they can constrain growth by underutilising women’s skills and misallocating talent across the labour market. Hsieh et al. (2019), for example, document that such forms of misallocation are important and that improvements in allocative efficiency accounted for a sizable share of US economic growth in the post-war period. Analysing cross-country data, Chiplunkar and Kleineberg (2024) show that frictions are important for explaining gender gaps even when other determinants of gender-specific labour demand are considered. Similar work on India (Chiplunkar and Goldberg 2024) focuses on female entrepreneurship, underscoring the need to address both demand- and supply-side distortions.

In recent work (Goldberg et al. 2025), we propose a new Global Gender Distortions Index (GGDI) to better understand these dynamics. The GGDI is grounded in economic theory and combines multiple aspects of gender-based misallocation into a single metric – aggregate productivity losses – to address a simple question: How much higher would economic activity be if women had the same opportunities as men?

A new framework for quantifying the cost of gender inequality in labour market opportunities

Using harmonised cross-country survey data on labour market indicators from a sample of countries comprising around two-thirds of the global population, we document gender gaps in employment, labour income, and educational attainment across time and countries at different income levels.

To construct the GGDI we build on Hsieh et al. (2019), using data on labour income and employment shares across activities to infer distortions faced by women in the labour market and quantify their impact on aggregate productivity. Our model assumes similar intrinsic preferences across genders, but posits that norms and institutions constrain employment choices. Women can face both labour demand distortions (e.g. hiring discrimination) and labour supply distortions (e.g. norms that discourage women’s employment outside the home). We treat gender differences in job-specific human capital as exogenous and measure it directly using the microdata.

Importantly, we distinguish between wage work, self-employment, unpaid work, and home production – reflecting the prevalence of informal work arrangements in developing countries. As such, unlike Hsieh et al. (2019), our model distinguishes between market-based GDP and overall welfare-relevant economic activity, as home production generates output but is not included in GDP.

Five insights on how gender gaps shape productivity

1. Global gender gaps persist in economic participation, types of work, and pay

Globally, we find that men consistently participate in wage work and self-employment at higher rates than women. In low-income countries, 37% of men but only 16% of women work for a wage, while 42% of men but only 28% of women are self-employed. These gaps narrow in high-income countries, suggesting greater labour market opportunities for women.

By extension, we find that women are more likely to engage in unpaid work (for example, as contributing family workers) – 15% in low-income countries, compared to 5% of men. Wage gaps also remain large. Women earn 35 percentage points (p.p.) less in low-income countries, 20 p.p. less in middle-income countries, and 30 p.p. less in high-income countries.

2. Gender gaps in labour markets remain large in many countries, despite gains in educational outcomes

We find that gender-based distortions in the labour market remain large in many countries, even where education gaps have narrowed or reversed. In low-income countries, while self-employed women have 20 p.p. fewer years of schooling than men, and women engaged in unpaid work have 43 p.p. less, there is no educational gender gap among wage workers. In middle-income countries, the schooling gap among self-employed women disappears while women working for a wage have 9 p.p. more years of schooling compared to men in wage work. In high-income countries, on average women are more educated but still engage in paid work less than men – suggesting persistent barriers to female labour force participation (FLFP) despite educational gains.

3. Gender distortions tend to decrease with economic development, but this is not guaranteed

While gender-based misallocation has declined in some countries, we find that others have seen little change despite growth. Across 51 countries, gender distortions are negatively correlated with economic development. However, the cross-country variation in distortions at similar income levels is large and growth does not automatically deliver gender equality – a result consistent with Agte et al. (2024). As Figure 1 shows, gender-based misallocation has fallen in Brazil, Chile, Mexico, Korea, and the US in recent decades. Yet India saw little net change, despite rapid growth.

Figure 1: The GGDI in different countries

Figure 2 shows the GGDI’s negative correlation with GDP per capita as well as large cross-country variation. Eliminating distortions could raise GDP per capita by 24% in Egypt but just 5% in Peru, which is slightly richer. This highlights how country-specific factors shape gender-based misallocation and influence the potential gains from the reduction of these distortions. Figure 2 also groups countries by region to underscore this heterogeneity.

Figure 2: The GGDI and economic development

4. Policies to reduce gender distortions in labour markets can increase FLFP and economic output, particularly on the labour demand side

Reducing gender distortions would draw more women into different job types in the labour market, raising productivity through more efficient talent allocation. Figure 3 shows that eliminating distortions substantially increases FLFP. We find that misallocation is largely driven by distortions in labour demand: in most countries, demand-side distortions (blue) are more important in explaining women’s lower participation rates than supply-side distortions (pink). Because labour demand distortions may be easier to address with policies than entrenched supply distortions like norms, this suggests that policymakers should focus on frictions limiting labour demand. Importantly, we find that demand and supply distortions are complementary: policies targeting demand-side distortions can capture much of the welfare benefit of reducing gender-based misallocation altogether.

Figure 3: Distortions and FLFP

5. GDP gains from reducing gender-based distortions would likely exceed our GGDI estimates

We also compare the effects of eliminating demand and supply distortions when measured by the GGDI versus GDP. Figure 4 shows that estimated GDP gains are uniformly higher (since official GDP accounts do not assign any value to non-market work like home production as we do) – suggesting that our GGDI estimates understate the gains policymakers would see in official GDP statistics. Removing distortions allows women to shift into the types of market work that are currently measured by GDP.

Figure 4: The GGDI and changes in GDP vs. total output

Implications for policy and research

Grounded in economic theory and comparable across countries and over time, the GGDI offers a clear unified metric to quantify the role of both demand- and supply-side distortions in the labour market. By using widely available data on labour income and job types, our methodology is easy to implement and provides flexibility for studying a range of labour market dimensions. As such, we view the GGDI as a complementary tool to other measures of gender inequality and we hope future researchers will make use of it. As part of our paper, we provide a full set of easy-to-use codes to compute the GGDI for other datasets that contain the required information on employment shares, earnings, and human capital. For policy actors, such enhanced understanding can open the door to targeted investigation and intervention on the driving forces behind these costly distortions.

Acknowledgements: We gratefully thank Greg Larson and the Yale Economic Growth Center (EGC) for their assistance on this VoxDev article. We thank the Development Policy and Finance team at the Gates foundation for financial support through EGC’s Gender and Growth Gaps project. Gottlieb gratefully acknowledges financial support from Structural Transformation and Economic Growth (STEG) LRG Grant 943.

Authors’ note: The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the World Bank and its affiliated organizations, nor those of the Executive Directors of the World Bank, nor the governments they represent.