We thank Joshua Intriligator and Jennifer Silverstein for capable research assistance.

We thank Joshua Intriligator and Jennifer Silverstein for capable research assistance.

"The cost efficiency of 520 New York school districts is measured using data envelopment analysis and stochastic frontier regression. Mean inefficiency is 14% using either method, and the rank order correlation coefficient of inefficiency between the two methods is 86. The three largest school districts (Buffalo, Rochester, and Syracuse) are among the most efficient. Student test scores are output variables, and socioeconomic control variables are included to adjust for the harshness of the educational environment in which districts operate. (JEL" C29, C61, 121) Copyright 1999 Western Economic Association International.

No abstract

Purpose -This paper aims to estimate empirically the effect on the voluntary turnover (quit) rate of employees when a large public corporation already judged as an outstanding employer is also ranked as being socially responsible by an external review organization.Design/methodology/approach -The paper employs a cross-section regression of the turnover rate of 84 of Fortune magazine's ''100 Best Employers'' against measures of corporate social responsibility (CSR) and several other control variables such as annual wages, ethnic and gender composition of the labor force that economic theory and prior studies have identified as explaining firm labor turnover.Findings -Adoption of business policies that cause the firm to be rated as socially responsible reduce the annual quit rate by 3 percent to 3.5 percent, which amounts to a 25-30 percent reduction, as compared to non-CSR public corporations or a larger comparison set including privately held and not-for-profit firms.Research limitations/implications -The wider universe of public corporations may not realize comparable turnover benefits from CSR as these ''best employers'' because these firms might be especially vulnerable or sensitive to corporate image when hiring workers.Practical implications -The model estimated permits calculation of the annual rise in average wages that would be required to reduce turnover by the same amount as CSR, a sum of approximately $3,700 per year or about 9 percent of the mean wage, with a lower bound estimate of about $1,000. This suggests that these firms can significantly reduce labor costs by investing in worker-friendly employment policies, which account for half of the entire measured CSR impact.Originality/value -This is believed to be the first effort to quantify rigorously the effect of CSR on the employment side of firm performance. Prior labor studies have looked at hypothetical employment scenarios involving students.