# Rustling Sheepskin in the Lone Star State

Schooling means many things, and every educational experience is different. But economists look at diverse educational experiences and see them all as creating human capital: a costly investment in skills from which we also expect to see a return. Some students, like Bertie Gladwin, enjoy school for its own sake and show little interest in economic returns. But many more probably see their schooling as stressful, tiring, and expensive. In addition to tuition costs, time spent in school could have been spent working. Many college students spend relatively little on tuition, but all full-time students pay an opportunity cost. This notion—that a large part of the costs of acquiring an education comes in the form of forgone earnings—leads us to expect each year of additional schooling to generate about the same economic return, whether it’s the tenth, twelfth, or twentieth year at the books. The simple human capital view of schooling embodies this idea.

Of course, people who have not had the benefit of economics training probably don’t think about education like this. Most measure their educational attainment in terms of degrees instead of years. Few job applicants describe themselves as having completed “17 years of schooling.” Rather, applicants list the schools from which they graduated and the dates of degrees received. To an economist, however, degrees are just pieces of paper that should have little or no real value. Master Stevefu is a case in point: though he spent many years in college, attending Susquehanna University in central Pennsylvania (among other fine institutions) he has yet to earn his bachelor’s degree. Reflecting this dismissive view of the value of certification, economists refer to the hypothesis that degrees matter as “sheepskin effects,” after the material on which diplomas were originally inscribed.

The search for sheepskin effects led Masters Damon Clark and Paco Martorell to a clever fuzzy RD research design.— They exploit the fact that in Texas, as in many other states, receipt of a high school diploma is conditional on satisfactory completion of an exit exam in addition to state-required coursework. Students first take this exam in tenth or eleventh grade, with retests scheduled periodically for those who fail. A last-chance exit exam for those who have failed previously is administered at the end of twelfth grade. In truth this isn’t the last chance for a Texas senior to earn a diploma; it’s possible to try again later. Still, for many who take it, the last-chance exam is decisive.

The decisive nature of the last-chance exit exam for many Texas high school seniors is documented in Figure 6.3. which plots the probability of diploma receipt against last – chance exam scores, centered at the passing threshold. The figure, which plots averages conditional on each score value along with fitted values from a fourth-order polynomial estimated separately on either side of the passing cutoff, shows diploma award rates close to.5 for students who miss the cutoff. For those whose scores clear the cutoff, however, diploma award rates jump above 90%. This change is discontinuous and unambiguous: Figure 6.3 documents a fuzzy RD first stage of nearly.5 for the effects of exit exam passage on diploma receipt.

Many of those who earn a diploma go on to college, in which case their earnings stay low until this additional schooling is also completed. It’s therefore important to look far enough down the road for any sheepskin effect in earnings to emerge. Clark and Martorell used data from the Texas unemployment insurance system, which records longitudinal information on the earnings of most workers in the state, to follow the earnings of those taking the last-chance exam for up to 11 years.

Earnings data for a period ranging from 7-11 years after students sat for their last – chance exit exam show no evidence of sheepskin effects. This can be seen in Figure 6.4. which plots average annual earnings against exam scores in a format paralleling that of Figure 6.3 (earnings here are in dollars and not in logs, and the averages include zeros for people who aren’t working). Figure 6.4 is a picture of the reduced form in a fuzzy RD design that uses a dummy for passing the exit exam as an instrumental variable for the effect of diploma receipt on earnings. As always, when the reduced form is zero—in this case, no jump appears in Figure 6.4—we know that the corresponding 2SFS estimate is zero as well.

FIGURE 6.3 Fast-chance exam scores and Texas sheepskin Notes: Last-chance exam scores are normalized relative to passing thresholds. Dots show average diploma receipt conditional on each score value. The solid lines are fitted values from a fourth-order polynomial, estimated separately on either side of the passing cutoff (indicated by the vertical dashed line). |

FIGURE 6.4 The effect of last-chance exam scores on earnings Test score relative to c utoff Notes: Last-chance exam scores are normalized relative to passing thresholds. Dots show average earnings conditional on each score value, including zeros for nonworkers. The solid lines are fitted values from a fourth-order polynomial, estimated separately on either side of the passing cutoff (indicated by the vertical dashed line). |

The 2SFS estimates generated by dividing the first-stage and reduced-form discontinuities seen in Figures 6.3 and 6A show a diploma effect of $52 (with a standard error of about $630). This amounts to less than half a percent of average earnings, which are about $13,000. These are small effects indeed, weighing against the sheepskin hypothesis. On the other hand, the associated confidence intervals also include earnings effects of nearly 10%.

Farge standard errors leave us with the possibility of some sheepskin effects, so the search for evidence on this point will surely continue. Masters know the search for econometric truth never ends, and that what is good today will be bettered tomorrow. Our students teach us this.

master steyefu: Time for you to leave, Grasshopper. You must continue your journey alone. Remember, when you follow the ’metrics path, anything is possible.

master joshway: Anything is possible, Grasshopper. Even so, always take the measure of the evidence.

## Leave a reply