@article{Jithitikulchai:112670,
      recid = {112670},
      author = {Jithitikulchai, Theepakorn},
      title = {U.S. Alcohol Consumption: Tax Instrumental Variables in  Quadratic Almost Ideal Demand System (QUAIDS)},
      address = {2011-07},
      number = {321-2016-11045},
      series = {Selected Poster},
      pages = {2},
      year = {2011},
      note = {Replaced with revised version of poster 08/30/11.},
      abstract = {This study analyzes the annual alcohol consumption on  prices, taxes, and other economic and demographic  characteristics from 50 states and the District of Columbia  for 1985-2002 to study the U.S. alcohol demand behavior on  beer, spirits, and wine. Young and Bielinska-Kwapisz (2003)  found that the data prices contain measurement error, so  using the state and federal taxes as the instrumental  variables could mitigate the problem. There is the  improvement in estimation to use the set of all taxes  instead of each individual tax. Therefore, we use the  Quadratic Almost Ideal Demand System (QUAIDS) model,  proposed by Bank, Blundell, and Lewbel (1997), employing  the alcohol prices or the tax instrumental variables on  both pooled and clustered datasets to compare with
the  classical linear models. The statistical inference is based  on the bootstrap variance-covariance matrix. After correct  the heteroskedasticity and multicollinearity issues, the  existing linear regressions models are GLS, GMM-IV, FE, RE,  FE-IV, and RE-IV models. The analysis of elasticities  reveals that the QUAIDS model is perform better than all  linear regressions in term of explanation on consumer  behavior. We use local constant and local linear estimators  with the Gaussian kernel and the optimal cross-validation  bandwidth to study the effects of prices and income on  alcohol consumption. We also find the increasing trends of  consumptions in all types of alcohol for each state. The  empirical results indicate that the gross substitution  effects among them could imply policy necessity on the  simultaneous alcohol excise taxes imposed. In term of  estimation, the instrumental tax variables could improve  the QUAIDS model on reflecting the different responsiveness  of price and income for light, medium and heavy consumption  levels.},
      url = {http://ageconsearch.umn.edu/record/112670},
      doi = {https://doi.org/10.22004/ag.econ.112670},
}