«Abstract A recent education reform in Germany reduced the duration of academic high school education by one year but left the curriculum, and total ...»
The Eﬀect of Compressing High School Education
on University Performance
Michael D¨rsam∗ Verena Lauber†
This Version: January 2016
Preliminary Version. Do not cite without permission.
A recent education reform in Germany reduced the duration of academic high school
education by one year but left the curriculum, and total class time unchanged. We
use a unique data set of university students to investigate the eﬀects of this reduction
in years of schooling on academic achievements at the tertiary level. By exploiting variation in the implementation of the reform across school types over time, we isolate the reform eﬀect from cohort, state, and school type eﬀects. Our results suggest that the reform lowers the opportunity costs of schooling and facilitates an earlier labor market entry as we ﬁnd no detrimental eﬀects while students are one year younger on average.
JEL-Code: I21, H52, C21 Keywords: School Duration, Academic Achievement, Diﬀerence-in-Diﬀerences, Germany ∗ University of Konstanz, Department of Economics, Box D133, 78457 Konstanz, Germany.
Email: firstname.lastname@example.org † University of Heidelberg, Bergheimer Str. 20, 69115 Heidelberg, Germany.
Email: email@example.com 1 Introduction The optimal design of the schooling system is a fundamental issue for economic policies.
Concerning the optimal length of schooling, policy makers face a diﬃcult trade-oﬀ. On the one hand, instruction time has been shown to be positively related to academic achievement (see, e.g., Bellei, 2009; W¨ßmann, 2003), while more years of schooling have been shown to o yield sizable monetary and non-monetary beneﬁts (Card, 1999; Lochner, 2011). On the other hand, the entry into the labor force is delayed and the duration of gainful employment reduced with an increasing length of schooling. Hanushek and Woessmann (2008) furthermore show that cognitive skills rather than mere school attainment determine economic well-being, and that the quality of school institutions is decisive. This raises the question of whether given school resources are used most eﬃciently, and whether the distribution of instruction time, and hence the length of schooling, is optimal. So far, most of the research designs are not able to investigate this question, as most school reforms simultaneously aﬀect instruction time, curriculum covered, and school duration (Patall et al., 2010). A high school reform, recently implemented in Germany, provides a setting to investigate this issue.
Since the early 2000s, most German states have reduced the duration of academic high school education from nine to eight years, but left the curriculum and the number of instruction hours up to the time of graduation unchanged. As a consequence, the number of instruction hours per day, and hence the learning intensity increased. This is the most fundamental reform of the German education system in the last decades and presumably the most controversially debated one. The main concern is that the higher learning intensity may have negative consequences on children’s development, in particular on their learning and human capital accumulation (see Lehn, 2010). Thus, opponents of the reform assume that the ratio of time spent in school to time for recreation was closer to optimal under the old system. However, evidence supporting this claim is missing.
The present paper helps to ﬁll this gap by providing ﬁrst evidence for longer-term eﬀects of this reduction in the duration of high school education. Using a unique data set of undergraduate students, we analyze the eﬀect of the reform on cognitive skills. The validity of our results is ensured by the implementation method of the reform: The German states introduced the eight-year system only in academic high schools, whereas it was not introduced in other high school types. This allows us to disentangle the eﬀect of the reform from cohort, state, and school-type eﬀects by estimating a diﬀerence-in-diﬀerences (DiD) model. Another major advantage of our data is that we observe treatment and control students within the same exams. Consequently, their performance is highly comparable.
The data we exploit was collected by the registrar’s oﬃce of the University of Konstanz, located in the state of Baden-Wuerttemberg. Most students in Germany choose a university close to their home town. Thus, the majority of the students in our sample graduated from high school in Baden-Wuerttemberg, representing our treatment state.1 BadenAs a robustness exercise, we compare our ﬁndings with the treatment eﬀect obtained for Bavarian G8 Wuerttemberg introduced the reform in academic high schools in the school year 2004/05.
The ﬁrst students with an eight-year lasting high school degree (G8 students) then enrolled at the University of Konstanz in fall 2012. At the same time, vocational high school students with a nine-year lasting high school degree (G9 students) enrolled. We exploit this variation across cohorts and between high school types. The main advantage of using variation within a state is that the estimates are not aﬀected by unmeasured state-level policies, as long as all schools are equally aﬀected (Hanushek et al., 1996).
The existing evidence on the short-run eﬀects of the reform does not generate clear expectations about the reform’s long-run eﬀects. On the one hand, ﬁfteen-year-olds have been shown to beneﬁt from the new system in terms of PISA performance (Homuth, 2012; Andrietti, 2015). High school graduation rates have furthermore been shown to be unaﬀected, while students are on average almost one year younger at the time of graduation(Huebener and Marcus, 2015). On the other hand, grade repetition rates in grade ten increased under the new system (Huebener and Marcus, 2015), and math scores in the ﬁnal high school exam decreased, at least in the ﬁrst G8 cohort in Saxony-Anhalt (B¨ttner and Thomsen, 2013).
u Even if the latter result applies to all German states and all aﬀected cohorts, the reform may still positively aﬀect university performance if it improves at the same time students’ non-cognitive skills, such as the ability to cope with stress.2 In fact, we ﬁnd that students of the ﬁrst G8 cohort who graduated from academic high schools together with students of the last G9 cohort (in the following also referred to as double cohort students) performed similar to students of the control group. Considering the second G8 cohort, we ﬁnd signiﬁcant positive eﬀects on the average grade obtained, the likelihood to fail an exam, as well as on the likelihood to obtain a top grade. These positive eﬀects stem particularly from the female students, while we ﬁnd no signiﬁcant eﬀects for the male students. Robustness checks support these ﬁndings. Given the one-year younger student body, our results suggest that the reform lowers the opportunity costs of education and facilitates an earlier labor market entry.
The paper proceeds as follows: In section 2, we present background information on the German education system, and especially the German high school reform. Section 3 gives an overview of the related literature. Section 4 describes our data. In section 5, we present our identiﬁcation strategy, provide graphical support for its validity, and discuss potential threats. Section 6 presents our ﬁndings, and section 7 concludes.
convention) set the years of schooling required to earn the Abitur (high school diploma) to 13 years. Following this convention, students in most West German states spent four years in primary school and nine years in secondary school until receiving the high school diploma (see K¨hn et al., 2013).3 In contrast, the German Democratic Republic pursued a policy u resulting in 12 years of schooling, including eight years of high school. After the German reuniﬁcation in 1990, the requirement for receiving the high school diploma amounted to 13 years of schooling in almost all German states. Only Saxony and Thuringia, two of the former East German states, held on to 12 years. In 1997, ”The Standing Conference of the Ministers of Education and Cultural Aﬀairs of the L¨nder in the Federal Republic of Gera many” (KMK), representing the most important interface of the German states within the national education policy, decided that the number of instruction hours up to the high school diploma must be identical across states, comprising 265 so-called ”Jahreswochenstunden”.
The number of school years, however, was allowed to vary. These criteria set in 1997 apply to both academic and vocational high schools, two types of high school where students can earn the high school diploma. G8 programs, however, were only implemented in academic high schools.
Between 2001 and 2007, 13 out of 16 federal states implemented the eight-year high school program. The only state still planning to keep the old program is Rhineland-Palatinate.4 At the time of implementation, students were in ﬁfth grade in most states.5 As a consequence of the reform, the learning intensity increased for G8 students: While G9 students have on average only a bit more than 29 instruction hours per week, G8 students spend on average about 33 hours per week in school. This increase is even more pronounced in grades seven to ten, because the number of weekly instruction hours was basically left unchanged in grades ﬁve and six.
The policy makers’ primary goal was to reduce the labor market entry age. Before the
reform, the age of German high school graduates was high by international comparison:
While the average graduation age in the OECD is at about 18 years, German high school graduates were aged between 19 and 20 (see OECD, 2014). By reducing the graduation age, policy makers aimed to counter the eﬀects of demographic change and to ensure sustainability of the social security systems by increasing the number of workers available to the labor market. At the same time, the quality of the high school diploma should not be reduced by the reform. Therefore, the curriculum and the number of instruction hours were remained unchanged, thus increasing the international competitiveness of German students (see Klemm, 2008).
After primary schooling, students are tracked into one of three secondary school tracks. The basic and intermediate tracks include schooling up to grade 9 and 10, respectively, usually followed by vocational training.
Today some states discuss returning to G9 because of public concerns.
With the following exceptions: In Saxony-Anhalt (ST) and Mecklenburg-Western Pomerania (MV) the reform was implemented in grades ﬁve to nine. In Bavaria (BY) and Lower Saxony (NI) it was implemented in grade ﬁve and six.
3 Related Literature There are only a few studies that identify the direct eﬀect of a variation in the years of schooling on educational outcomes. Most research dealing with school time variations estimate a mixed eﬀect of diﬀerences in the amount and the distribution of instruction time, and the curriculum covered. 6 Krashinsky (2014) and Morin (2013) explore a policy change in Canada similar to the German G8 reform that reduced high school by one year. However, in contrast to the G8 reform, only part of the curriculum remained unchanged, and the choice of the program was not random. Based on an instrumental variable strategy, Krashinsky (2014) ﬁnd signiﬁcant negative eﬀects of the shortened high school education on later educational outcomes. Morin (2013) explores the fact that the mathematics curriculum was shortened from ﬁve to four years while the length of the biology curriculum for the same students remained unchanged.
His results point to a small positive eﬀect of an extra year of high school mathematics on university performance.
A relatively clear setting for the analysis of the impact of learning intensity on academic achievement provides the introduction of a four-day-school week. In this case, a given amount of instruction time is distributed over fewer days. This increases the learning intensity during the school days, but also increases the number of recreation days. In this setting, Anderson and Walker (2015) identify negative eﬀects for math and reading skills. Eren and Millimet (2007) explore the eﬀects of the length of the school year, the number of class periods per day, and the average length per class period in the United States. They ﬁnd that shorter class periods are associated with higher test scores.