Images: Freeflagicons
The rapid expansion of China’s international trade that began in the early 90s had important effects on national trade expansions worldwide. In BSE Working Paper 1230, “Trade-induced Local Labor market Shocks and Asymmetrical Labor Income Risk,” Tomás R. Martínez and Ursula Mello analyze the consequences and causes of the so-called “China shock” in the labor income risk faced by Brazilian workers.
Why is the shape of the distribution important?
The first step in the analysis rationalizes the impact of trade shocks on individual income risk using a theoretical model. Whenever an economic trade shock occurs, it impacts the individual income risk by changing the distribution of the unobserved income innovations. Intuitively, greater import competition increases the dispersion of permanent and/or transitory income shocks, making the labor market riskier. The econometrician can infer these changes from variations in the empirical distribution of income. By comparing short and long-run variations in the shape of the distribution (i.e., its moments), they can distinguish whether the trade shock had more impact on the transitory or the persistent idiosyncratic risk.
Data and empirical strategy
Brazil provides a particularly interesting case study for trade-induced local labor market shocks due to the large number of local markets with different comparative advantages. The core of the analysis relies on rich employer-employee matched data spanning 1995-2015 covering more than 500 local markets. The data is carefully filtered to mitigate possible biases arising from migration or retirement decisions, among others.
China’s trade shock had heterogeneous impacts across sectors in the Brazilian economy. Whereas both the agriculture and extractive sector benefited from it, the manufacturing sector experienced a negative import shock. This last change constitutes the main research question of the paper: What has been the impact of negative import-competition shocks in the formal sector?
To provide an answer, they first proxy trade shocks and distribution changes by measuring import penetration and changes in the moments of the empirical distribution, respectively. In the next step, they rely on census data to control for local labor market characteristics and thus, to focus on within sector variation. Lastly, as there might be unobserved factors that cause endogeneity bias (i.e., national subsidies to subsectors), an instrumental variable is built using China’s trade flow with the rest of the world (excluding Brazil).
Empirical results
The answer to the previous question is affirmative: trade produces significant changes in the dispersion of the distribution of idiosyncratic income growth. Firstly, import penetration has a significant effect on both the transitory and persistent risk (i.e., a significant increase in the variance), although the latter plays a more prominent role. The next natural question to answer is then: What are the possible drivers of this change?
To understand the type of economic shocks and choices behind a change in yearly labor income, it is important to decompose any change between wages and hours worked. For instance, while a promotion and a health shock can both lead to a change in labor income, the former might be reflected in a higher wage, whereas the latter might lead to lower working hours. The authors found that the increase in volatility of hours worked annually explained a large portion of the changes in the variance.
The second important channel is through labor reallocation. Does the displacement of workers explain the changes in the distribution of income growth? Job and industry switchers (i.e., individuals that have changed employees or industry within a certain period) are the ones most affected by the trade shocks. This result is consistent with a large increase in the variance of yearly hours worked as job and industry transitions provoke higher volatility.
Armed with the estimated coefficients from the previous analysis, the authors calibrate a partial equilibrium life-cycle model to conduct welfare analysis. According to their estimates, a newborn worker is willing to give up 4.4% of lifetime consumption to avoid the riskier labor market created after the China shock. Importantly, omitting the impact of higher moments (i.e., not accounting for nonnormality) reduces this estimate to half its size.
Conclusions
The paper examines the link between trade shocks and asymmetrical labor income risk taking advantage of the China shock and the heterogeneity in the Brazilian local markets. The main insight is to highlight the importance of accounting for higher-order moments of the distribution of income changes when analyzing the causal link between trade and risk in empirical research. Furthermore, the authors can disentangle the main mechanisms behind the high volatility and conduct welfare analysis.
This is the first paper to use regional variation in labor markets to quantify the impact of penetration in income risks, and the analysis sheds light on possible future research questions such as whether this risk spills over across regions.
