The importance of data structure and nonlinearities in estimating climate impacts on outdoor recreation

Abstract

Credible empirical estimation of the economic impacts of climate change is dependent on data structure (e.g., cross sectional, panel) and the functional relationship between weather data and behavioral outcomes. We show here how these modeling decisions lead to significantly different results when estimating the effects of weather and simulating the potential welfare impacts of climate change on outdoor recreation. Using participation data from 1.6 million households in the United States from 2004 to 2009, we estimate the impact of temperature and precipitation on participation decisions for marine shoreline recreational fishing. Results from linear models suggest temperature positively impacts participation and, by implication, climate change is likely to improve welfare associated with outdoor recreation in all regions of our study area. Conversely, nonlinear specifications suggest more days with extreme heat reduce participation and lead to significant declines in welfare under future climate scenarios. Differences in the treatment of how weather enters recreation participation decisions change both the sign and magnitude of welfare effects by nearly $1 billion annually. Differences in data structure, however, only affect the magnitude of welfare impacts but not the sign. Disaggregation of welfare estimates suggests warmer baseline climates are more susceptible to these choices. Our results demonstrate the critical nature of modeling decisions about data structure and the use of weather data to assess the future impacts of climate change, especially with nonmarket goods where value is related to environmental quality such as outdoor recreation.