This paper provides new information about the interrelated issues of teacher turnover (both within and across school districts and inside and outside of teaching) and the importance of nonpecuniary school characteristics, such as race and poverty, using new administrative data on Georgia teachers and the elementary schools in which they teach. Simple descriptive statistics indicate that teachers are more likely to change schools if they begin their teaching careers in schools with lower student test scores, schools with lower income students, or schools that have higher proportions of minority students. A linear probability and a competing risks model of transitions out of first teaching jobs allow us to separate the importance of these highly correlated school characteristics.The estimates from the model imply that teachers are much more likely to exit schools with large proportions of minority students, and that the other univariate statistical relationships associated with student test scores and poverty rates are driven to a large extent by the correlations of these variables with the minority variable. Thus we find that, while the common notion that teachers are more likely to leave high poverty schools is correct, it occurs because teachers are more likely to leave a particular type of poor school -that which has a large proportion of minority students.1

No abstract

This paper examines the incidence of the implicit lottery tax and the distribution of benefits from lottery-funded programs in Georgia. Georgia's lottery is unique in that revenues are earmarked for three educational programs-HOPE College Scholarships, universal pre-kindergarten, and education infrastructure. We estimate separate models of household-level lottery purchases and of household benefits from lottery-funded programs. Our estimates suggest that lower income and non-white households tend to have higher purchases of lottery products while receiving lower benefits, as compared to higher income and white households. Benefits of HOPE Scholarships, in particular, accrue disproportionately to higher income and more educated households.

Within the housing segregation literature major disagreements have developed over two fundamental issues: (1) the role that whites' aversion to racially mixed neighbourhoods plays in causing modern segregation in the US; and (2) the factors that underlie this aversion, including the effects of inter-racial contact on whites' neighbourhood racial preferences and whether these preferences reflect neighbourhood stereotyping as opposed to pure racial prejudice. Extant evidence on these issues is either old or indirect. This paper provides direct evidence on these issues using new data from the Multi-City Study of Urban Inequality. The results suggest that (1) whites' neighbourhood racial preferences play an important role in explaining the racial composition of their neighbourhoods; (2) inter-racial contact in neighbourhoods and workplaces leads to a greater willingness among whites to live with blacks; and (3) although younger and more educated whites express a stronger taste for integration than other whites, the magnitude of these differences leads to only a small increase in the black percentage of the neighbourhood. In addition, the results provide no evidence in support of the hypothesis that whites stereotype black neighbourhoods rather than blacks per se.