U.S. High School Graduation Rates: Patterns and Explanations

Richard J Murnane
2013 Journal of Economic Literature  
I survey the evidence on patterns in U.S. high school graduation rates over the period 1970-2010 and report the results of new research conducted to fill in holes in the evidence. I begin by pointing out the strengths and limitations of existing data sources. I then describe six striking patterns in graduation rates. They include stagnation over the last three decades of the twentieth century, significant race-, income-, and gender-based gaps, and significant increases in graduation rates over
more » ... he first decade of the twenty-first century, especially among blacks and Hispanics. I then describe the models economists use to explain the decisions of individuals to invest in schooling, and examine the extent to which the parameters of the models explain recent patterns in graduation rates. I find that increases in the nonmonetary costs of completing high school and the increasing availability of the GED credential help to explain stagnation in the face of substantial gaps between the wages of high school graduates and school dropouts. I point out that there are several hypotheses, but to date, very little evidence to explain the increases in high school graduation rates over the first decade of the twenty-first century. I conclude by reviewing the evidence on effective strategies to increase high school graduation rates, and explaining why the causal evidence is quite modest. Between 1970 and 2000, the high school graduation in the United States stagnated. In contrast, the secondary school graduation rate in many other OECD countries increased markedly during this period. A consequence is that in 2000, the high school graduation rate in the United States ranked thirteenth among nineteen OECD countries (OECD, 2012). Until quite recently, it appeared that the stagnation of the U.S. high school graduation rate had continued into the twenty-first century. However, evidence from two independent sources shows that the graduation rate increased substantially between 2000 and 2010. This increase prevented the United States from losing further ground relative to other OECD countries in preparing a skilled workforce. But graduation rates in other OECD countries also increased during that decade. As a result, the U.S. high school graduation rate in 2010 was still below the OECD average. 3 Given the importance of educational attainment and skills of workers to labor market earnings for individuals and economic growth for countries (Goldin and Katz, 2008; Hanushek and Woessman, 2008), it is important to understand the reasons for the three-decade stagnation and for the recent increase in the U.S. high school graduation rate. As I explain in detail below, during the last three decades of the twentieth century, increased requirements for a high school diploma counteracted substantial economic returns to the credential. More difficult to explain are the recent increase, and the striking gender gap, in the graduation rate. The paper addresses the following four questions: 1. What were the patterns in U.S. high school graduation rates during the last 40 years? 3 The OECD (2012) estimates that the U.S. high school graduation rate increased from 70 percent in 1970 to 77 percent in 2010. The OECD bases its estimates on the Common Core of Data, which is described in Section 2. In 2010, the average upper secondary school graduation rate among OECD countries was 84 percent. In that year, the U.S. high school graduation rate ranked twentieth out of twenty-four OECD countries (OECD, 2012, Table A2 .3, p.55). However, this comparison may be somewhat deceptive in that, as reported in Table A2 .2 of the OECD publication (2012), the high school graduation rate of the population under age 25 in the United States (77 percent) was equal to the OECD average high school graduation rate for people under age 25. The explanation is that while a significant number of people complete secondary education after age 24 in some OECD countries, this is not the case in the United States.
doi:10.1257/jel.51.2.370 fatcat:jmm5z4a4xvaxxfzrvt3z7vxrnm