Are Natural Hazards Temporally Random?
Philip E. Graves
University of Colorado at Boulder - Department of Economics
Anne E. Bresnock
California State Polytechnic University, Pomona
January 1, 1985
Applied Geography, Vol. 5, No. 1, pp. 5-12, 1985
The paper employs statistical hypothesis tests to explore the question of whether natural hazards (hail and tornadoes being considered here) are or are not intertemporally random. The answer to this question, at least for these two hazards, is surprising and has important policy implications: hazards appear to be more likely in year t if an event was experienced in year t- 1: hence, apparent ‘Gambler’s Fallacy’ behavior on the part of farmers and others may indeed be rational. Should this result hold for a full range of climate-related hazard types, as we suspect, hundreds of millions of dollars of crop loss and other damages may be preventable (e.g., rotating to frost-resistant crops if past events suggest greater than usual likelihood of early or late frost}. The degree of positive serial correlation is also seen to vary a great deal regionally. with individuals in some states being unable to benefit from knowledge of prior hazard occurrence.
Number of Pages in PDF File: 8
Keywords: Natural hazards, hail, tornadoes, intertemporal correlation of natural hazards
JEL Classification: H12, H43, Q10, Q19, Q28, Q38, Q54, Q58
Date posted: September 19, 2010