Abstract

In this paper, we investigate whether and how climate change inhibits or induces the transformation of typical agrarian economies in Africa. Structural transformation is an important phenomenon shaping the economies of many developing countries, as it generates efficiency and productivity gains by reallocating labor from the less productive to the more productive sectors of an economy. We examine how medium-term change in temperature and precipitation affects employment shares in agriculture, manufacturing, and the service sectors. In doing so, we also investigate heterogeneities across gender, age, and educational attainment. This is an important transition because it can facilitate economic growth and poverty reduction. To explore potential mechanisms, we examine the impact of medium-term changes in temperature on agricultural productivity, labor force participation, and rural–urban migration. Beyond examining the impact of temperatures on employment shares in agriculture and nonagricultural sectors, we also evaluate impacts on labor force participation of men and women.

To address the above questions, we rely on unique census data from 30 national censuses covering 12 African countries between 1974 and 2014, collated by the Integrated Public Use Microdata Series (IPUMS). These data enable us to study medium- and long-term changes in the reallocation of labor across the three broad sectors of economies: agriculture, manufacturing, and services. We merge these data with well-gridded earth observation precipitation and temperature data from the Terrestrial Air Temperature and Precipitation Gridded monthly time series (Willmott and Matsuura, 2001). Using these data, we observe monthly temperature and precipitation over a long period (1900–2017), with precision (0.5° × 0.5°). We also use the Normalized Difference Vegetation Index (NDVI) dataset from the National Oceanic and Atmospheric Administration (NOAA), which enables us to capture vegetation cover of each district in our database from 1981 to 2014. Combining these datasets enables us to study medium- and long-term responses to climate change, which are likely to differ from short-term responses and adaptation strategies.

Exploiting plausibly exogenous temporal variations in (average) decadal temperature, we note several findings. First, we find that a medium-term (decadal) increase in temperature inhibits structural transformation by increasing the share of labor in agriculture and reducing the corresponding shares in manufacturing and service sectors. A 1◦C increase in decadal temperature increases employment in agriculture by about 6 percentage points. We also document heterogeneities in the impact of climate change. We show that climate change leads to a gendered delay in structural transformation, with the delay triggered by climate change being more pronounced among women. A 1◦C increase in decadal temperature increases the share of women’s employment in agriculture by about 9 percentage points while delaying the transition to the service sector by about 7 percentage points, whereas the corresponding effects for men remain about half. Similarly, less-educated individuals are more likely to face additional constraints (because of climate change) to reallocate labor into nonagricultural sectors compared with educated individuals. These findings offer important insights on the vulnerability of women and less-educated individuals to climate change. Furthermore, they suggest that climate change can trigger gendered structural transformation in Africa and may aggravate existing inequalities in societies.

We also explore potential mechanisms through which climate change may delay structural transformation. We show that climate change leads to a reduction in agricultural yield, which ultimately affects demand for nonagricultural goods and services and the ensuing demand for labor in nonagricultural sectors. This is consistent with previous evidence showing that an increase in temperature is associated with a reduction in agricultural yield and an increase in the use of pesticides. Our results remain robust across a wide range of robustness checks. Moreover, our results are not driven by any specific country or a small share of the sample. Our results remain consistent even when excluding the warmest or coldest districts or when randomly dropping 20 percent of the districts from the sample. To our knowledge, our study provides fresh evidence showing the effects of climate change on the much-needed structural transformation in Africa.

Our second and probably more unique contribution is uncovering whether climate change aggravates existing inequalities by inhibiting or facilitating labor reallocation among different population groups, for which empirical evidence remains scant. We demonstrate that in the presence of climate change women are more likely to face a larger delay in their reallocation of labor from agriculture to nonagricultural sectors, implying that climate change can trigger gendered structural transformation. We also show that less-educated households are less likely to engage in nonagricultural activities in areas experiencing climate change. As agricultural productivity declines limited nonagricultural job opportunities push the most vulnerable groups (women and less-educated individuals) further into agricultural employment. Our final contribution comes through our effort to identify new mechanisms through which climate change can delay structural transformation, including an increase in labor force participation, and a reduction in agricultural productivity.