Search

The econometrics of higher education emerged from the development of human capital theory and efforts to estimate rates of return to education in the 1960s and 1970s. This paper surveys the various strands of the literature on the econometrics of higher education that have developed during the last 40 years and indicates how a collection of papers fits into this literature. It also discusses: the estimation of rates of return to higher education; determinants of college enrollment, college graduation, and choice of major; studies of the academic labor market; studies relating to models of university behavior; and studies relating to higher education as an industry and higher education governance. The paper stresses that the central econometric problem faced by higher education researchers who employ individual-level data is that of selection. Similarly, the central econometric problem faced by researchers who use market-level data is the difficulty of disentangling demand and supply shocks. Much of the challenge that empirical higher education researchers face is attempting to satisfactorily resolve these problems. (Contains 174 references.) (SM)

The cost benefits of lifelong learning in the United Kingdom were estimated, based on quantitative evidence. Between 1975-1996, 43 police force areas in England and Wales were studied to determine the effect of wages on crime. It was found that a 10 percent rise in the average pay of those on low pay reduces the overall area property crime rate by between 0.7 and 1.0 percentage points. The benefit was estimated to be between 1.3 and 1.8 billion British pounds per year. Furthermore, the effect on wages of a one point increase in the proportion of the working age area population with a high school diploma or equivalent qualification was predicted to lie between 10 million and 320 million pounds. It was found that cost benefit increases and crime decreases with more education. Parent background, health, and social cohesion were among key factors underlying the relationship between learning, achievement, and crime reduction. Future qualitative studies were recommended to determine more specifically the causal mechanisms and mediating processes that bear on the wider benefits of learning throughout an individual's life. (Seven tables/figures are included. The bibliography lists 53 references. Appended are tables on the relationship between education and training qualifications and wages and a classification of academic and vocational qualifications.) (AJ)

These papers examine research on the impact of class size on student achievement. After an "Introduction," (Richard Rothstein), Part 1, "Understanding the Magnitude and Effect of Class Size on Student Achievement" (Alan B. Krueger), presents a reanalysis of Hanushek's 1997 literature review, criticizing Hanushek's vote-counting method and suggesting it would be better to count each publication as a single study rather than counting separately each estimate within a publication. This method of analyzing the research suggests a strong connection between school resources used to reduce class size and student outcomes. It also discusses the effects of expenditures per student and looks at economic criterion (Lazear's theory of class size, benefits and costs of educational resources, the critical effect size, and caveats). Part 2, "Evidence, Politics, and the Class Size Debate" (Eric A. Hanushek), focuses on: the history of class size reduction; econometric evidence; the Tennessee Class Size Experiment (Project STAR, or the Student/Teacher Achievement Ratio Study); and policy calculations. The author claims that his own method of counting each estimate as a separate study is the most valid method and that effects of class size reduction will be small and expensive. An appendix discusses the econometric data. (Papers contain references.) (SM)

This paper uses direct measures of literacy skill levels provided by the International Adult Literacy Survey to estimate the return to literacy skills. Using a very simple human capital earnings equation and standard ordinary least squares regression, it tested estimates of the return to literacy skills for their robustness to alternative scalings of literacy attainment. The paper emphasizes the importance of alternative possible scalings of literacy scores, because literacy scores are inherently ordinal, not cardinal, numbers. The paper explores the sensitivity of econometric results to a range of monotonic transformations of literacy scores and concludes that, particularly for males, much of the economic return yielded by education is due to literacy skills--perhaps as much as 40-45 percent. Other conclusions are that, while education has a higher overall rate of return for females than for males, not as much of that return is explained by literacy; literacy may also account for a larger proportion of the impact of education on earnings among those with higher literacy skills; and literacy does affect earnings over quite a wide range of metrics. Alternative possible scalings of literacy scores also imply that the rankings of Canadian provinces in average literacy vary with the scaling adopted, sometimes quite dramatically. Appendixes contain 18 endnotes and 29 references. (YLB)

This report describes efforts by Wheaton College in Norton, Massachusetts, to address the problem of inadequate faculty salaries by developing and implementing a "profit-sharing" plan that adjusts the faculty salary pool in relation to changes in the institution's financial well-being. It notes that faculty-administration relations at the small liberal arts college had soured in the early 1990s as a result of faculty salary issues. The plan posits that as the college's resources increase or decrease relative to the mean of nine comparison institutions, Wheaton faculty salaries will change by the same margin. The "floor" of the plan stipulates that faculty salaries will increase at the rate of the consumer price index, while the "ceiling" stipulates that if Wheaton faculty salaries become equal to the mean salaries at the comparison institutions, they will not exceed that mean until Wheaton's resource base increases to within 20 percent of the average resources of the comparison institutions. The plan was successfully implemented and created a climate in which faculty began to consider themselves stakeholders in the institution's financial well being, leading to their increased participation in recruitment, retention, and fund raising. (MDM)

This study explores the nature of professional expertise in economics. Two experiments were conducted to examine the existence of problem categories in economics as a basis for problem representation. Findings are that experts use categories that reflect major economics principles, and novices sort by literal cues contained in the problems. In the second trial, experts did not produce stable sortings whereas novices produced a stable sorting of categories. Subjects included 6 experts, 6 intermediate fourth-year university students in economics, and 6 novices in the first year of college with one semester in economics. Subjects were asked to sort 18 economics problems into groups based on similarities of solution. After sorting the problems, subjects were asked to write down the problem-numbers and give a description of the category. A second study was to determine the stability of sorting procedures among experts after a lapse of six weeks. No differences were found in the average number of categories produced per sort between different groups. (EH)

This study attempts to investigate development of expertise in the area of economics. Results suggest that development of expertise in economics has a stage-like development similar to the domain of medicine. Studies in medicine show that expertise development is not simply a process of extending knowledge in existing semantic networks. The study involved 7 experts from the economics department, 8 fourth-year university students, and 10 novices at the end of the first year of the economics programs. Subjects were given a text for a limited period of time and then asked to write down as much as they could remember after the time lapsed. The cases examined were considered to be at intermediate difficulty. Recall and case analysis protocols were scored against the original case text propositions. Data suggest a stage of development among novices, intermediates, and experts. (EH)

This study used the frontier cost function framework to examine cost efficiency in public higher education. The frontier cost function estimates the minimum predicted cost for producing a given amount of output. Data from the annual Almanac issues of the "Chronicle of Higher Education" were used to calculate state level enrollments at two-year and four-year public institutions, total educational and general expenditures, research expenditures, state appropriations for higher education, and average tuition rates for the 1989-90 through 1992-93 academic years. It was found that, overall, inefficiency does exist in public higher education, in that states have costs that are about 20 percent above the estimated frontier. It also found that states with higher two-year school enrollments had higher efficiency rankings than states with lower two-year enrollments, and that larger university systems tended to be more efficient than smaller systems. A significant relationship was also found between the share of total educational expenditures provided by state appropriations and the degree of efficiency, in that states with the smallest and largest shares were more efficient than states in the middle group. (Contains 31 references.) (MDM)

Intergovernmental grants are a significant part of both federal and state budgets. This report examines the federal grant-in-aid system from the perspective of fiscal impact. It focuses on the extent to which the grant system succeeds in two objectives: (1) encouraging states to use federal dollars to supplement rather than replace their own spending on nationally important activities; and (2) targeting grant funding to states with relatively greater programmatic needs and fewer fiscal resources. The first major finding is that every additional federal grant dollar results in less than a dollar of total additional spending on the aided activity: the estimates of substitution clustered around 60 cents of every federal dollar. Therefore, part of the fiscal impact of these transfers is to free up a portion of state funds for other state programs or tax relief. In addition, federal aid is not targeted to offset fiscal imbalances among states with different programmatic needs and fiscal resources. The second major finding is that grant design influences fiscal substitution. Design features work by restricting the use of funds to specify purposes, requiring recipients to contribute their own funds to obtain grant funds, and not restricting federal matching of state funds. Finally, the study found that grants lack features to discourage substitution. Four figures and 11 tables are included. Appendices contain information on the role of grants in the federal system, the research design, three grant design features related to substitution, statistical analyses, challenges in measuring grant-targeting factors, and a list of major contributors to the report. (Contains a 24-item list of related GAO products and 51 references.) (LMI)

This paper classifies the university economic impacts on a local community into their short- and long-term effects. The study then offers a classification scheme of university communities by the size of their regions and by their influences. Next, the paper suggests ways that economic impacts vary across university types and describes and critiques examples of short- and long-term studies that have been done. The local expenditure impact of a university is immediately noticeable. Knowledge impacts affect the community over a longer term. The extent of the economic impact of a college or university is related to the range of its influence and the character of the community in which it is located. Institutions in college towns have large regional expenditure and small knowledge impacts; metropolitan schools with regional reputations have small expenditure and large knowledge impacts. Economic impact studies previously done include income expenditure studies, American Council of Education studies, economic base studies, input-output studies, econometric studies, and knowledge impact studies. Seven exhibits presenting the data are included. (Contains 86 references.) (CK)

Conducted at a time when increasing numbers of the nation's young children are in child care and when the American public is concerned about children's readiness for school, the Cost, Quality, and Child Outcomes Study provides the first comprehensive econometric and psychometric analysis of child care and children's outcomes. The study was designed to examine the relationships among the costs of child care and the nature and effects of children's child care experiences. Cost and quality data were collected through visits to 50 non-profit and 50 for-profit centers in each of four states: California, Colorado, Connecticut, and North Carolina. Trained data collectors conducted interviews with and distributed questionnaires to center directors, teachers, and parents; they also observed two randomly chosen classrooms in each center. Data were then collected on 826 children from preschool classrooms visited earlier. The study found that while child care varies widely within and between states and sectors of this industry, most child care is mediocre in quality, sufficiently poor to interfere with children's emotional and intellectual development. Market forces constrain the cost of child care and at the same time depress the quality of care provided to children. It costs somewhat more to provide good quality care than to produce poor quality care; however, higher costs are not obviously reflected in parent fees, which are relatively similar in centers of different quality. Based on the findings, the following recommendations were made: (1) launch efforts to educate parents on identifying high quality programs; (2) implement higher state standards; (3) increase investments in child care staff; and (4) assure adequate financing and support of child care. (BGC)

Any interstate comparison that does not take differences in the cost of education into account will give an incorrect impression of the relative levels at which different states support their schools. The lack of cost-adjusted statistics on state expenditures for elementary and secondary education interferes with policy analysis, resource allocation, and decision making in education. An examination of state cost data indicates that there is substantially more interstate variation in per-pupil spending than in the amounts of the main educational resource--instructional staff--that states provide for each pupil. In fact, the main conceptual and technical problems in constructing a cost-of-education index arise from the labor-intensive nature of education and the types of labor on which schools rely. This report evaluates some ways of constructing cost-of-education indices, focusing on existing approaches rather than proposing new ones. Conceptual advantages of econometric supply-demand modeling approach are noted, as are its practical disadvantages. Improvements in databases should make construction of a model-based cost-of-education index much more feasible in the near future. Chapters are: (1) "Introduction: Cost Differentials and Cost-of-Education Indices"; (2) "Conceptual Foundations of Cost-of-Education Indies": (3) "Alternative Index Construction Methods"; (4) "Demonstration and Assessment of Simple Cost Indices and Proxies"; and (5) "Conclusions: Current Capabilities and Promising Options." (Contains 19 tables and 48 references.) (SLD)

This paper analyzes the effects of the 1988-1992 macroeconomic crisis in Peru on the schooling and employment decisions taken by school-aged children in urban areas. It discusses the Peruvian setting during this period and describes the data used in the analysis and the econometric specification. Two basic findings were made: (1) there was no effect of the crisis on attendance rates of school-aged children, however, the fraction of children who were both employed and in school decreased significantly during the crisis; and (2) mean educational attainment increased significantly for children exposed to the crisis than for those unexposed. The paper argues that these findings may be related. Children who are not employed have more time available and, therefore, may put more effort into school. The analysis concludes with a discussion of the implications of these findings for the design of appropriate social protection policies. (Includes 21 notes, six figures, and four tables. Contains a 41-item bibliography.) (Author/BT)

Comparison of vocational training evaluation in the United States, Organization for Economic Cooperation and Development states, and the European Union identified certain advantages to the U.S. model of microeconometric impact evaluation, including better policy and expenditure of public money. (Contains 49 references.) (SK)