Climate Change and Farm Use of Weather Information

More rapid than normal global climate change as represented by rising temperatures and more erratic and severe weather events have heightened the interest in how farmers use weather information. The greenhouse influence through driving climate change will likely be affecting agricultural efforts for some years to come. It behooves us to pay attention to this phenomenon, and especially put effort into understanding how farmers will respond to information about climate generally and forecasts in particular. This research is being funded by the U.S. Department of Commerce, National Oceanic and Atmospheric Administration. To address this issue farmers were surveyed in three major agroecological zones representing 1) a western Corn Belt, rainfed corn-soybean regime, 2) a central Great Plains irrigated corn-soybean regime, and 3) a central Great Plains irrigated continuous corn regime. Each of these zones is represented in three counties in eastern Nebraska. To better reflect farmers' weather related needs and issues, focus groups were held in each county to engage farmers in helping the researchers to design the survey instrument. The fact we used focus groups added an intriguing flavor to the study. Between 15-20 farmers in each zone were involved. These farmers were paid $25 for a 2-hour session that culminated in a provided lunch. Each session was also videotaped, providing the opportunity for all members of the research team to see the results of the event (in that only 3-4 members of the larger multidisciplinary research team were allowed to attend an event, due to concerns over affecting farmer responses). The focus was on the farmers' reactions to a series of questions prepared beforehand by the research team, all of which stirred lively dialogue on how farmers actually use weather information. The result was a substantially improved questionnaire. We also sent drafts back to farmer participants for final reviews, and subsequent adjustments were then made reflecting the way they used the words and understood the survey questions. The qualitative data from the focus group events ultimately influenced not only the way questions were asked but also how the modeling is done, and, especially, how the model results are interpreted. A total of 2211 questionnaires were sent, in two separate mailings. There was also a follow-up reminder card. Farmers were offered a payment of $25 to return the questionnaire. A total of 28% took the offer and the overall return rate was 33%, with 698 usable questionnaires in the econometric analysis. A distinctive aspect of this study is the fact that the research team involves active participation not only by agricultural economists but also by a psychologist and a social psychologist, as well as two meteorologists, and an agronomist (who is also a GIS specialist). The result is a nontraditional behavioral economics approach that is sensitive to the climate and agronomic realities faced by farmers in these zones. This approach has a unique two-fold feature; first, it puts special attention on underlying motives, and second, asks whether there may be a complex expression of both private (self) and public (other, community) interest in how forecasts influence farm level decisions. Yet, the modeling still reflects standard derived demand theory and the general expectancy-value or subjective utility perspective, i.e. that farmers have beliefs about fact events and values relating to the outcomes from those events, and that the demand for weather information is derived from the value (profit, sense of well-being, risk-reducing value) it produces for them. The beliefs represent probability statements about outcomes and the values represent the utility or profit related transformations of meaning about the farmer perceptions of the outcomes. The econometric analysis uses proxy measures of the expectancy-value as independent measures, along with such variables as financial capability of the farmer as represented in farm sales, to explain in a Tobit kind of framework 1) the probability of applying weather forecast information since it influences farm level decision(s), and 2) the extent to which this forecast information is influencing these decision(s). The set of four Tobit models in Table 1 test the influence of recent past and current experience (RPE), short (STF) and long-term forecasts (LTF) on 1) agronomic (e.g. selecting the crop type, spraying), 2) insurance, and 3) marketing decisions, within recent past experience/short-term and long-term forecasts. To test the models, we created four indices represented in balance (joint and nonseparable ratio of public (other) to private (self) interests); attitude as a construct of personal belief and value system, influence of social norms, household and community members, county extension, etc; farmers' need for internal control over crop production; and farm sales representing financial limitations. Preliminary analyses suggest that all the farm decisions are influenced by weather forecast information at a different intensity (Table 1). The probability of that influence increases with balance, as the farmer puts more effort into pursuing the self over the community interests. Influence of others and social norms intensify the use of weather information in the decisions as well. Those who want more control over the farm are likely to be more influenced by weather forecasts. Finally, influence of weather forecasts becomes greater as gross farm income (sales) increases. Other intriguing interpretations are suggested by the changes in the size of the parameter elasticities and marginal effects3, e.g. the control parameter is substantively smaller in the insurance decision, which suggests farmers see insurance as offsetting the need for more control over their decisions. As another example, the balance in private and public interests is less significant and less a factor in the very personal, private marketing decision in contrast to "how one farms" (which is likely more sensitive to community scrutiny) in the agronomic decisions. The larger paper explores these refinements in greater detail. Table 1. Intensity of Weather Forecast Influence on Farm Decisions. Variables Agronomy (Cur. Rec. Past Exp. & Short term forecasts) decisions Agronomy (Long Term Forecasts) decisions E1 E 2 ME 1 ME 2 E 1 E 2 ME 1 ME 2 Balance -.37b -.37b -1.109b -.012b -.23a -.24a -.637a -.0202a Attitude .62c .62c .495c .0052c .71c .72c .560c .0178c Norms .12c .12c .153c .0016c .09b .09b .102b .0032b PBC .15c .15c .157c .0017c .19c .19c .178c .0056c Farm Sale .07b .07b .086b .0009b .02 .02 .024 .0008 Easting .06b .06b -2.2E-6b -2.3E-8b -.01 -.01 3.7E-7 1.2E-8 Insurance decisions Marketing decisions Balance -.12 -.13 -.247 -.033 -.29a -.29a -.779a -.040a Attitude .93c .98c .674c .089c .58c .59c .486c .025c PBC .07 .07 .048 .006 .13b .13b .119b .006b Farm Sale .18c .19c .164c .022c .20c .20c .238c .012c Note: Dependent variable is the degree of influence of climate and weather information and forecasts. a p<0.10, b p<0.05, c p<0.001. 3 E1 is the elasticity at the mean that represents the percentage change in the probability that the weather and climate forecast and information influences decisions at all, and; E2 is the elasticity at the mean for those who are being influenced, the percentage change in the degree of influence. ME1 is the effect of the expected value for the weather and climate already influenced farmers; ME2 is the effect of the probability of being influenced by climate and weather information (elasticity of influence).