What can the past teach us about the economic impacts of climate change?
Modern empirical economics tries to use past experience and large amounts of data to learn about the impacts of policies or events. Many refer to these advances as the “credibility revolution” in economics, a new set of methods that allows economists to generate knowledge based on data rather than theory.
The economic impacts of climate change seem like a fruitful application of these modern methods. Temperature is one of the pieces of data we have been recording for a long time and fluctuations in weather from year to year are basically random. Numerous papers have tried to study the impacts of climate change on GDP, productivity, violence, or test scores. Yet the usefulness of all of this work is limited by a simple fact, we don’t have any prior experience with climate change, so all the empirical work we can do is based on weather.
Weather or climate, it seems like semantics, but the distinction is crucial for understanding how climate change is going to transform our world.
Uncertainty is the driving concept distinguishing climate from weather. The climate is what we expect the weather to be in the future, weather is what actually happens. For example, the fall in New England is relatively cool (that’s climate), but this Wednesday, temperatures in Cambridge, MA, reached 80 degrees (that’s weather).
Why would we care about this distinction? Economic theory can help us understand.
Suppose I’m a farmer planning out what crops to plant in the spring. My decision will be based on what I expect the weather to be in the summer. Ideally, I would have very good weather forecasting service and I would know that the average daily temperature in the summer of 2017 is going to be 85 degrees Fahrenheit. This would allow me to plant the right crop mix and irrigate the correct share of my land to maximize my profits during an 85-degree summer. In reality, I don’t have a perfect weather forecast so I have to use the climate as a guide.
The average daily temperature during the summer on my farm has been 85 degrees. All signs early in the spring say I can expect this summer to be typical, but I can’t be completely certain. There’s some risk of a very hot summer so I may want to hedge by building in some extra irrigation. I would not do this if I was sure that the temperature was going to be 85 degrees.
Our question: How will profits change if the climate increases by one degree?
The way we empirically study how temperature impacts economic outcomes is to look at how our outcome of interest (income) changes with changes in temperature relative to the average temperature for that location. The reason we do this instead of just looking at how profits change with temperature is that hotter places are different from cooler places for reasons that are unrelated to weather. For example, the northeastern US is wealthier than the southeastern US for many reasons besides temperature such as average education levels, local politics, or natural resource endowments. By looking at relative changes in temperature we make sure that we don’t misattribute the economic impact of fixed local characteristics to temperature. Unfortunately, it also means is that we can only measure the impact of weather on the economy, not the impact of climate.
Why is that?
Let’s go back to the farming example. Since I can’t perfectly forecast the weather, my decisions are based on the climate. If I expect a daily average temperature of 85 degrees but the temperature is actually 86 degrees instead, I won’t be able to adjust my crop choices or irrigation plans in response to this change since these choices were made in the spring.
If the shift from 85 to 86 degrees is in the climate, then I can respond, because I will know about this shift in the expected temperature for the summer ahead of time. I will also be responding to a slightly different type of shift. A change in the climate doesn’t mean the temperature will be one degree higher for sure. There’s still some chance of cooler days and really hot days that I have to plan for.
The challenge in using historical data to measure the economic impacts of climate change is that we are stuck with looking at the impact of weather and thus fail to account for economic forces that can lead to large differences in impacts:
- We will know about changes in the climate before they happen, so we may take some actions to ameliorate their impacts.
- We will react to changes in weather differently than we would to changes in climate because when we respond to changes in climate we are making decisions before we know what the weather is going to be.
- A lot of this distinction would disappear if we were able to predict the weather in advance. But it will remain very difficult to predict the weather perfectly far into the future and we will therefore have to respond to decade-long shifts by making large fixed investments based on risks of events in the following decade that may or may not actually happen.
When economists take their empirical estimates of the economic impacts of temperature or rainfall and use these results to predict what the economic cost of climate change will be, we have to remember that these estimates are based on unpredictable changes in the weather and that climate change is predictable (up to a point). These are estimates of the wrong effect.
This type of research still has immense value in helping us understand how temperature can disrupt an economic process. These new empirical methods can help us understand how adaptation can ameliorate the costs of climate change, particularly by studying why some places are more vulnerable than others. Remember, adaptation does not come at zero cost.
Image courtesy of Flickr. Originally published by S&S on October 25, 2016.