Across countries, women and men allocate time differently between market work, domestic services, and care work. In this paper, we document the gender division of work, drawing on a new harmonized dataset that provides us with high-quality time use data for 50 countries spanning the global income distribution. A striking feature of the data is the wide dispersion across countries at similar income levels. We use these data to motivate a macroeconomic model of household time use in which country-level allocations are shaped by wages and a set of “wedges” that resemble productivity, preferences, and disutilities. Taking the model to country-level observations, we find that a wedge related to the disutility of market work for women plays a crucial role in generating the observed dispersion of outcomes, particularly for middle-income countries. Variation in the division of non-market work is principally shaped by a wedge indicating greater disutility for men, which is especially large in some low- and middle-income countries.

We revisit the role of human capital and skill-biased technology in explaining the cross-country variation in GDP per worker. We propose a general-equilibrium accounting model in which workers of different human-capital groups (education and age) sort across broad occupational categories. The occupational assignment is determined by the comparative advantage of workers as well as productivity and distortion parameters that are specific to each human capital and occupation pair. We map the model to a harmonized micro dataset of 43 countries spanning the entire development spectrum that allows us to measure wages by human capital and occupation. The calibration reveals that cross-country productivity gaps are increasing in the complexity of the occupation. Also, conditional on occupation, the cross-country productivity gaps vary more strongly for low-skill than high-skill workers. The process of development is hence biased toward white-collar occupations but not biased toward higher-skilled workers. We then use the model to perform several counterfactuals. We find that the composition of human capital explains 14% of the non-agricultural GDP per worker gap between the richest and poorest countries. Occupational distortions are more pronounced in poor countries, but have a minor quantitative effect on aggregate output.

Across countries, the organisation of production differs widely, and firms are substantially smaller in low-income countries. How are these differences related to a country’s skill endowment? In this paper, we study the interplay between skill endowment and firm size along the development path. Empirically, we measure the skill intensity of employment by firm size for 57 countries and document four facts. First, we show that the share of employment in large firms is about four times as high in high-income countries as in low-income countries. Second, we find that across countries, employees of large firms are more skilled than those of small firms. Third, whereas small firms in rich countries are almost as skill intensive as large firms, small firms employ much fewer skilled workers in poor countries. Fourth, whereas small firms are almost as skill intensive as large f irms when the skill premium is low, they are much less skill-intensive when the skill premium is high. This evidence suggests that small firms can easily substitute low-skill for high skill workers when high-skill workers are scarce and expensive, but large firms are less flexible. As a result, the low skill endowment of low-income countries limits the size of firms in these countries. We then use a span-of-control model with worker skill heterogeneity and two technologies (large and small scale) to analyze the effect of skill endowments on the firm size distribution and economic development. Calibrated to the US and varying only skill endowments, our model closely replicates skill intensity by firm size across countries. It implies that, to a large extent, it is the lower skill endowments of poor countries that lie behind the four facts we document. Our findings also imply that greater skill levels benefit production not only directly, but also through a shift to more large scale production units. Finally, we illustrate how our work relates to the literature on misallocation.

We exploit the quasi-random allocation of asylum seekers across Swiss cantons and the high frequency of national referenda to identify the causal effect of immigration on political outcomes in receiving countries. We find that the arrival of asylum seekers causes voters to increase their support for right-wing and conservative policies. However, this effect is driven by episodes of unusually high inflows of asylum seekers. Moreover, we find that for votes on immigration and refugee policy, the arrival of more asylum seekers shifts voters towards policies endorsed by conservative and centre-right parties but not towards positions backed by the rightmost anti-immigration party. In contrast, the shift towards the rightmost stances is sizeable in votes related to the welfare state, international integration, and the rights of minorities.