Files
Abstract
Conventional wisdom argues that tourist expenditures and recreation activities generate
demands for traded goods and services, and create jobs and income for local residents in
counties endowed with rich natural amenities. However, more recent studies have suggested
that regions with high levels of amenities can experience lower wages and higher unemployment
because amenities are capitalized into wages and rents in a manner that can hinder economic
growth. Attempting to estimate the impact of tourism and retirement activities on the
local economy, a few studies have performed multiple least squares regression analysis to discount
activities generated by both local residents and nonresidents who travel for purposes
other than tourism. However, the least square regression provides nothing more than an estimate
of the average of the response (dependent) variable conditioned on the covariates (independent
variables). In almost all regression settings with the exception of the rather naive constant-
error-variance setup, the upper and lower quantiles (percentiles) often depend on the covariates
quite differently from the mean or the median response. Investigating quantiles other
than the mean or median using quantile regression analysis, we have found interesting dependency
effects that cannot be discovered otherwise. The results of this analysis provide crucial
information to policy makers while discussing public policy effectiveness in natural resource
management and community development. If policy is to rely on the structural shift
that is taking place in rural America, we need a better understanding on how amenities, quality
of life attributes, and tourism affect regional economic performance.