This week was the Wold Congress of Environment and Resource Economics (WCERE) in Montreal, an event which happens only once every four years. With 800 talks, I couldn't see most of them. But below are titles, abstracts, links to papers and some of my thoughts on those talks that I found most interesting. There's obviously sampling biases and my comments are based only on the talks, not the papers, so errors are not unlikely. In a future post, I'll include links to papers presented by students in the Columbia PhD program, which is why they are omitted below. To download the full working paper, follow the link to the WCERE site and click on the "download PDF" link on the right
INTER-ANNUAL WEATHER VARIATION AND CROP YIELDS
Wolfram Schlenker, Columbia University
Abstract: While the effects of rising mean temperatures on agricultural output have been studied extensively, there is limited discussion of the impact of inter-annual weather variation on crop yields. This paper estimates the link between weather and crop yields separating the influence of (i) mean weather outcomes (i.e., climate) to which a farmer can adapt from (ii) unpredictable year-to-year weather fluctuations to which a farmer can only partially adapt as crops are planted before the weather shock is realized. We find that corn in extreme climates, both hot and cold, are more sensitive to inter-annual weather variation than the ones in moderate climates. Global warming has two effects on corn yields: First, warming will induce farmers in moderate-temperate climates to plant varieties that are less robust to weather fluctuations, while farmers in cool climates will plant varieties that are more robust to weather fluctuations. Second, the elasticity of reductions in expected corn yields with respect to an increase in the standard deviation of weather fluctuations is -0.4. Since most farmers are eligible for subsidized crop insurance, an increase in weather variation also directly translates into added government payments.Comments: One of this paper's nicest aspects is that it explicitly looks for the "micro" structure underlying aggregate production curves. We often assume there is a sequence of production functions for several crops, each of which has a different peak point along some dimension (here, temperature). [For an example of this assumed structure, see the discussion in Deschenes and Greenston (AER, 2007); their working paper is here.] The usual assumption is that farmers who adapt to climate changes produce along the upper envelope of several overlapping production curves and are therefore less vulnerable to long term changes in temperature than they are to short term fluctuations. This paper tries to test this assumption explicitly and finds that it holds up. The other findings of the paper are nice, but to me, this seems like the most important contribution.
Comments: The data assimilation involved with this paper is impressive, so I bet they will produce more work using these tornadoes as instruments. The results are clean and somewhat larger than I would have expected. However, I think the title and interpretation may not be appropriate, since it was not clear that they ever tested for impacts associated with the risk of events occurring, only the events themselves.
Abstract: A critical issue in climate-change economics is the specification of the so-called "damages function" and its interaction with the unknown uncertainty of catastrophic outcomes. This paper asks how much we might be misled by our economic assessment of climate change when we employ a conventional quadratic damages function and/or a thin-tailed probability distribution for extreme temperatures. The paper gives some numerical examples of the indirect value of various GHG concentration targets as insurance against catastrophic climate-change temperatures and damages. These numerical examples suggest that we might be underestimating considerably the welfare losses from uncertainty by using a quadratic damages function and/or a thin-tailed temperature distribution. In these examples, the primary reason for keeping GHG levels down is to insure against high-temperature catastrophic climate risks.
Comments: I think this is the most constructive of Weitzman's string of papers on catastrophic risk, perhaps because the final result (which is actually the absence of a result) doesn't rest on infinite negative losses. I was most struck by his case that the quadratic loss function assumed by Nordhaus et al. is insufficiently flexible to express large aversions to catastrophic events. I think this is an example of a seemingly innocuous, esoteric assumption made twenty years ago coming back to bite us later.
Abstract: We study the consequences of poverty alleviation programs for environmental degradation in Mexico. We exploit the community-level eligibility discontinuity for Oportunidades (a conditional cash transfer program) to identify the impacts of income increases on local deforestation, and use random variation in the initial program phase to explore household responses. We find that additional income significantly increases demand for resource-intensive consumption goods. The corresponding production response increases deforestation but is localized only where communities have poor road infrastructure. The results suggest that better access to markets simply disperses environmental harm; the true impacts are only observable in infrastructure poor areas.
Comments: I really like what this paper is trying to do, the data they are willing to integrate and the importance of the findings. It has all the makings of a great paper. My only concern is whether the absence of deforestation in communities with high road densities is robust. In general, its far harder to demonstrate a "null" result than to show that some sort of non-zero relationship exists, since failing to reject a null-hypothesis doesn't tell you much. I think that in order to make the claim described in the last two sentences of the abstract, they'll have to do substantially more work with the NDVI data they are using and/or look at data on trade or other markets.
Abstract: Recent research indicates that monsoon rainfall became less frequent but more intense in India during the latter half of the Twentieth Century, thus increasing the risk of drought and flood damage to the country’s rice crop. Our statistical analysis of state-level Indian data confirms that drought and extreme rainfall negatively affected rice yield during 1966-2002. Using Monte Carlo simulation, we find that yield would have been 2% higher on average if monsoon characteristics, especially drought frequency, had not changed since 1960. Yield would have received an additional boost of nearly 5% if two other meteorological changes (warmer nights, lower surface radiation) had not occurred. Climate change has evidently already negatively affected India’s hundreds of millions of rice producers and consumers.
Comments: I think their Monte Carlo method will become standard in the next few years. Its a nice way at getting at "climate" in comparison to "weather".
Abstract: We identify two issues with a standard time series approach to reconstruction of past climate fluctuations from paleoclimatic data series, one related to specification of the estimated relationship between climate and the paleoclimatic index, the other to the methodology of estimation. We show that the standard approach provides biased estimates of the reconstructed climate series and underestimates the true variability of historical climate. We demonstrate that inversion of the estimated response function between tree ring growth and climate indicators provides consistent estimates of historical climate. The inversion method results in an overestimation of the variance. We show analytically as well as using Monte Carlo experiments and actual tree ring data, that use of the new specification and reconstruction procedure can be crucial for inferences about the nature of past climate and interpretation of recent climate variations.
Comments: I think this paper is going to have a big impact. And I think the idea of having statisticians from different fields check one another's work is brilliant. I hope this is implemented more in the future and I hope economists invite non-economists to check their math, so the exchange is bidirectional.
Abstract: In this paper we estimate the impacts of climate change on the allocation of time using econometric models that exploit plausibly exogenous variation in daily temperature over time within counties. We find large reductions in U.S. labor supply in industries with high exposure to climate and similarly large decreases in time allocated to outdoor leisure. We also find suggestive evidence of short-run adaptation through temporal substitutions and acclimatization. Given the industrial composition of the US, the net impacts on total employment are likely to be small, but significant changes in leisure time as well as large scale redistributions of income may be consequential. In developing countries, where the industrial base is more typically concentrated in climate-exposed industries and baseline temperatures are already warmer, employment impacts may be considerably larger.
Comments: Clean, simple and intuitive. I think there will be more work in this direction. The only surprise is that in the twenty years since Schelling's armchair assessment, nobody thought of this.