Content area
Abstract
The Indian Gaming Regulatory Act provides a statutory framework for tribal gaming and stipulates that net gaming revenues must be used by tribes to foster economic development. After it was passed in 1988, tribal gaming revenues increased dramatically as did the number of tribal casinos. Using a difference-in-difference methodology with 1990 and 2000 Census data, I examine how the introductions of casinos impact the well-being of on reservation American Indians.
In my first two chapters, I use aggregate reservation data. In Chapter 1, I use least squares. Casinos increase per capita income by about 8% and reduce family and child poverty rates by 4 to 5 percentage points. I find the largest casinos are associated with the largest increases in income and reductions in poverty. However, least squares will only be unbiased if the decision to open a casino is uncorrelated with unobservable heterogeneity in reservation per capita income or poverty. In Chapter 2, I control for this potential endogeneity of the casino choice variables by including instrumental variables and find income effects increase 2 to 4% but lose statistical significance. Poverty effects increase in magnitude to 11 to 14 percentage points for family poverty rates and 16 to 20 percentage points for child poverty rates. Hence, these results may mean tribes with fewer outside opportunities open casinos and least squares biases effects downwards. However, after dropping at least one instrument, there are problems with weak identification, income effects decline to almost zero, and poverty impacts are no longer significant but remain large in magnitude.
In Chapter 3, I use 1990 and 2000 IPUMS-USA micro-data to examine impacts on household per capita total income, earned income, state and federal assistance income, and poverty according to head of household skill and market size. Casinos impacts are largest on low-skill heads but significance is dependent on the model used. If impacts vary by market size, income effects differ on the measure used and the sex of the head of household. On the other hand, with the exception of female headed families' near poverty rates, poverty rates decline more in larger markets.
