A BBC reporter asked me whether Hurricane Sandy reduced the death count relative to what it would have been during a normal period? She was making a nice "crowding out" point. The expectation of a macro shock (the storm) leads to more investment in self protection such as staying inside and not driving. Note the Lucas Critique logic, if people engaged in their usual amount of driving and outdoor activity when a storm hits then Sandy would certainly cause a spike in deaths but when the rules of the game change, we change our behavior.
Now the key unknown here is the daily death count by geographic area. If we had such data, we could see if there was a positive or negative residual during the Hurricane Sandy event. I found this document on the death count by state/year. In New Jersey in 2008, 70,000 people died so that means that on the average day 192 people died while in New York, 149,000 people died or 408 people a day. The total deaths reported in the Northeast for Sandy was 113.
The issue of "crowding out" arises again and again in urban settings. Do mandatory seat belts laws cause fatalities? They will if people drive faster (i.e change their behavior) when mandated to wear them. Do increases in police patrols increase crime? They will if people feel safer and reduce their self pre-cautions such as starting to take midnight walks and jogs. In both cases, government actions to increase safety lead to more "nasty events" because self interested people changed their behavior in response to the event. You can think of Hurricane Sandy as the flip event. In the case of Sandy, risk went up and people responded by hunkering down! Economics matters!
We are not passive victims of national policy and shocks. We play a strategic game with Bernanke, the President and Mother Nature. Long live the game theorists!
Demand for my Youtube environmental lectures appears to be finite. I may need to lower the price of viewing below zero! My son has told me to stop wasting my time but it is my time and I will allocate it based on my own optimization problem.