Demand shocks and fluctuations in output and productivity

Hyojung Lee, Purdue University

Abstract

An important series of papers (Koren and Tenreyro 2007&2013; Kraay and Ventura 2007; Krishna and Levchenko 2012) have shown that output growth rates fluctuate more in developing countries. Leading explanations focus on supply side explanations. In my dissertation, I provide a novel explanation focused on general equilibrium linkages between sectors and the critical importance of meeting subsistence food needs in the face of declining agricultural productivity. Unlike the previous literature, I provide evidence using a specific and identifiable shock that hits both poor and rich countries. Further, I highlight the importance of income effects in consumption, rather than differences in institutions or industrial structure that gives rise to differences in equilibrium output responses. To develop the idea, I build a two-sector static general equilibrium model featuring Stone-Geary preferences with a subsistence requirement for food. Under a closed economy, a negative shock to agricultural productivity, such as drought, causes food prices to rise. The expenditure on the subsistence requirement for food then rises, and there will be less income leftover for manufacturing, which leads poor households to shift consumption away from manufactures. On the production side, in order to meet the subsistence requirement in the face of the decreasing agricultural productivity, capital and labor resources have to move away from manufacturing and into agriculture, further reducing manufacturing output. Perversely, the economy shifts resources toward the sector with declining productivity, sharply curtailing total factor productivity and total output in the economy. However, this phenomenon is reversed under the small open economy. When there is a decrease in agricultural productivity, manufacturing sector looks relatively more productive, and resources move toward manufacturing. Using this framework, I show that there is a significant positive link between agricultural productivity and manufacturing output, and manufacturing output growth rates fluctuate more in developing countries. This effect is larger the closer is income to the subsistence level. Calibration exercises show that the growth rate of industrial output fluctuates significantly more in poor countries in response to changes in crop yields, a proxy for agricultural productivity. In addition, I test the predictions empirically. I utilize annual manufacturing data and instrument for crop yields using year-to-year changes in rainfall. The results show that a 1% decrease in crop yields induced by shortages in rainfall decreases manufacturing output by 0.38%, capital investment by 1.56%, and employment by 0.20% across 44 developing countries. Overall, crop yield variation (instrumented by rainfall shocks) can explain about 28% of industrial output growth fluctuations in developing countries. In addition, I show that an adverse shock to agricultural productivity can sharply curtail aggregate productivity and output in low-income countries. This pattern is most pronounced when the country is closed to trade, as resources reallocate toward agriculture when its productivity is declining in order to meet the subsistence needs. The opportunity cost of such resource reallocations is quite high if we compare with the small open economy case where resources can be reallocated toward relatively more productive sectors. For example, I find that such opportunity cost can be about 6% of the aggregate output in Ethiopia. Therefore, the trade liberalization can help lessen output fluctuations especially in developing countries by weakening the strong role of subsistence requirements in the consumer demand.

Degree

Ph.D.

Advisors

Hummels, Purdue University.

Subject Area

Commerce-Business

COinS