Modeling Structural Changes in Market Demand and Supply

Economic events may cause structural changes in markets. To know the effect of the economic event we should analyze the structural changes in the market demand and supply. The purpose of this dissertation is to analyze the effect of selected economic events on market demand and supply using econometric models. Structural changes can be modeled according to the types of changes. For an abrupt and instantaneous break, a dummy variable model can be used. For a smooth and gradual movement, proxy variables which represent the event can be applied, if we know the variables. If we don?t know the appropriate proxy variables, a smooth transition regression model can be employed. The BSE (Bovine Spongiform Encephalopathy) outbreak in the U.S. in 2003 is assumed to make abrupt and instantaneous changes in Korean meat consumption. To analyze the effect on Korean meat consumption, the Korean demands of beef, pork, chicken, and U.S. beef are estimated using an LA/AIDS (Linear Approximate Almost Ideal Demand System) model with the dummy variable specifying the time before and after the BSE. From the results we can confirm that food safety concerns caused by the BSE case changed Korean meat consumption structure. Korean beef and U.S. beef became less elastic, and pork and chicken got more elastic to budget. Korean beef became less price elastic, but pork and U.S. beef got more price elastic. The changes of U.S. natural gas supply caused by technology development and depletion in reserves are analyzed using a smooth transition regression model. From the results, we can confirm that the productivity improvement by technology development is greater than the labor cost increase by depletion, but not greater than the capital cost increase by depletion in mid-2000s. The effects of posting the winning bid in a repeated Vickrey auction are examined using a proxy variable. By applying an unobserved effect Tobit model to the experimental auction done by Corrigan and Rousu (2006) for a candy bar, we can confirm that the changes of bidding behavior are significant, especially when the winning bid is high. By extracting the bid affiliation effects, we showed that true willingness to pay can be estimated.