Temperature Effects on Productivity and Factor Reallocation: Evidence from a Half Million Chinese Manufacturing Plant
(with Peng Zhang, Kyle Meng, and Junjie Zhang)
NBER Working Paper No. 23991

This paper uses detailed production data from a half million Chinese manufacturing plants over 1998-2007 to estimate the effects of temperature on firm-level total factor productivity (TFP), factor inputs, and output. We detect an inverted U-shaped relationship between temperature and TFP and show that it primarily drives the temperature-output effect. Both labor- and capital- intensive firms exhibit sensitivity to high temperatures. By mid 21st century, if no additional adaptation were to occur, we project that climate change will reduce Chinese manufacturing output annually by 12%, equivalent to a loss of $39.5 billion in 2007 dollars. This implies substantial local and global economic consequences as the Chinese manufacturing sector produces 32% of national GDP and supplies 12% of global exports.

Link to paper

Maybe Next Month? Temperature Shocks, Climate Change, and Dynamic Adjustments in Birth Rates
(with Alan Barreca and Melanie Guldi)
NBER Working Paper No. 21681

Dynamic adjustments could be a useful strategy for mitigating the costs of acute environmental shocks when timing is not a strictly binding constraint. To investigate whether such adjustments could apply to fertility, we estimate the effects of temperature shocks on birth rates in the United States between 1931 and 2010. Our innovative approach allows for presumably random variation in the distribution of daily temperatures to affect birth rates up to 24 months into the future. We find that additional days above 80 °F cause a large decline in birth rates approximately 8 to 10 months later. The initial decline is followed by a partial rebound in births over the next few months implying that populations can mitigate the fertility cost of temperature shocks by shifting conception month. This dynamic adjustment helps explain the observed decline in birth rates during the spring and subsequent increase during the summer. The lack of a full rebound suggests that increased temperatures due to climate change may reduce population growth rates in the coming century. As an added cost, climate change will shift even more births to the summer months when third trimester exposure to dangerously high temperatures increases. Based on our analysis of historical changes in the temperature-fertility relationship, we conclude air conditioning could be used to substantially offset the fertility costs of climate change.

Link to paper

Media Coverage: CNN

Media Coverage: The Globe and Mail

Media Coverage: Stephen Colbert

A Forward Looking Ricardian Approach: Do Land Markets Capitalize Climate Change Forecasts?

(with Christopher Severen and Christopher Costello)
NBER Working Paper No. 22413


The hedonic pricing method is one of the fundamental approaches used to estimate the economic value of attributes that affect the market price of an asset. In environmental economics, such methods are routinely used to derive the economic valuation of environmental attributes such as air pollution and water quality. For example, the Ricardian approach is based on a hedonic regression of land values on historical climate variables. Forecasts of future climate can then be employed to estimate the future costs of climate change. This extensively-applied approach contains an important implicit assumption that current land markets ignore current climate forecasts. While this assumption was defensible decades ago (when this literature first emerged), it is reasonable to hypothesize that information on climate change is so pervasive today that markets may already price in expectations of future climate change. We show how to account for this with a straightforward empirical correction (called the Forward-Looking Ricardian Approach) that can be implemented with readily available data. We apply this empirically to agricultural land markets in the United States and find evidence that these markets already are accounting for climate change forecasts. Failing to account for this would lead a researcher to understate climate change damages by 36% to 66%.


Link to paper