Sunday, January 15, 2012

Interdisciplinary Environmental Research is Hard: The Case of a MIT Full Prof's Estimates of Wind Damage Caused to Property

Kerry Emanuel is a climate scientist at MIT.  He has the misfortune to be receiving nasty emails related to his research.  I was interested in who he is and what he does so I went to his research webpage  and downloaded this paper titled "Global Warming Effects on U.S. Hurricane Damage".  Professor Emanuel does not believe in technological change. In equation (1) of his paper (see page 264), he writes down a time invariant function of local wind speeds causing property damage.  Apparently, there are no investments or improvements in the building stock that can be made that would reduce the impact of wind speed on our building stock.  This sounds silly to me and I will return to this point below.

Here is the abstract for his "economics paper"

"While many studies of the effects of global warming on hurricanes predict an increase in various metrics of
Atlantic basin-wide activity, it is less clear that this signal will emerge from background noise in measures of
hurricane damage, which depend largely on rare, high-intensity landfalling events and are thus highly volatile
compared to basin-wide storm metrics. Using a recently developed hurricane synthesizer driven by large-scale meteorological variables derived from global climate models, 1000 artificial 100-yr time series of Atlantic hurricanes that make landfall along the U.S. Gulf and East Coasts are generated for four climate models and for current climate conditions as well as for the warmer climate of 100 yr hence under the Intergovernmental Panel on Climate Change (IPCC) emissions scenario A1b. These synthetic hurricanes damage a portfolio of insured property according to an aggregate wind-damage function; damage from flooding is not considered here. Assuming that the hurricane climate changes linearly with time, a 1000-member ensemble of time series of property damage was created. Three of the four climate models used produce increasing damage with time, with the global warming signal emerging on time scales of 40, 113, and 170 yr, respectively. It is pointed out, however, that probabilities of damage increase significantly well before such emergence time scales and it is shown that probability density distributions of aggregate damage become appreciably separated from those of the control climate on time scales as short as 25 yr. For the fourth climate model, damages decrease with time, but the signal is weak."

In the conclusion of the paper, the author issues the caveat; "A number of caveats apply to the present analysis. First, we have held constant the distribution and value of insured property, not accounting for changing demographics or adaptation strategies that might reduce vulnerability to damage."

So, I respect his honesty. He is admitting that he has chosen to focus on the worst case scenario and to ignore the highly likely behavioral response to changing actions by Mother Nature.  The climate scientists need to incorporate adaptation into their models but they assume it away because they do not know how to model it.  That's not productive.  These smart guys need to return to the rational expectations and investment under uncertainty models that economists have developed over the last 30 years and read their Lucas Critique.  As a starting point, they should read Climatopolis!

Permit me to make a prediction.  Due to endogenous innovation the relationship between wind speed and property damage will sharply decline over time.  The good new ideas that engineers and architects will figure out will be spread around the world and the resulting damage caused by windstorms will decline sharply.  How quickly will this occur?  I do not know that but the MIT professor needs to remember that necessity is the mother of invention.   So, our disagreement concerns how far human ingenuity can go to mitigate Mother Nature's wraith.  Some of his MIT students are likely to come up with the solutions that will make his projections look like a sharp upper bound!

UPDATE:  Let's assume that Prof. Emanuel is correct that we never figure out how to produce buildings that can withstand the wind.  He then faces a new extrapolation issue.  As the wind grows more intense in specific geographical areas, and we are aware of this, why do we keep building in those areas?   The answer is we won't.  We will migrate to a safer area and build our cities there.   Implicit is so much "climate change science" is a behavioral extrapolation that we will continue to live under a BAU (business as usual) scenario even when the rules of the game have changed (i.e climate change).  This violates 50 years of modern economics and I don't believe this.  The smart scientists at MIT need to walk down to E-52 and take a course in rational expectations investment theory with Acemoglu or anyone else on the floor.