The success of these teacher-hiring systems depends on whether the information used to screen candidates accurately predicts teacher quality. However, research has shown that many of the observable characteristics of those training to become teachers fail to predict their future effectiveness in educating students. Therefore, schools may struggle to establish selection criteria to discern the potential of teacher candidates, enhance teacher quality, and ultimately improve students’ learning outcomes. Hiring systems that heavily weight specific indicators—such as licensing requirements, educational attainment, or performance on standardized exams—may be counterproductive, especially when they lead to decisions that ignore or downplay other dimensions that are more predictive of teacher quality, such as skills acquired through experience on the job. Therefore, teacher-hiring policies aimed at improving incoming teacher quality could negatively affect students.
Our research examines the effectiveness of a nationwide reform implemented in 2005 in Colombia that sought to enhance teacher quality and improve student learning in public schools. The reform raised teacher salaries and introduced a centralized, merit-based teacher-hiring system that tied hiring decisions to candidates’ performance on a national standardized exam evaluating subject knowledge and teaching aptitude. The policy replaced many experienced public school teachers without a long-term civil service contract with novice teachers who performed well on the exam. This centralized hiring policy replaced a decentralized one that had been criticized for being subject to political influence. Within 10 years, nearly half of all public school teachers had been hired under the new regulation.
Our research uses data on teachers and students from the Colombian Ministry of National Education spanning 1998–2019. The teacher data allowed us to measure how the reform changed the characteristics of incoming teachers, such as prior test scores, education levels, age, gender, and experience. The student data report the performance of students on high school exit exams and their subsequent college outcomes. The reform did not apply to private schools, so our research compares students in public and private schools before and after the reform to evaluate its effects.
Teachers hired after the reform appeared to be more qualified, as they scored significantly higher on standardized exams taken before their hiring than those hired under the previous system. If these standardized exam scores—meant to measure cognitive ability—perfectly predict teacher effectiveness, then hiring teachers with higher scores may promote student learning.
However, the reform also required the government to replace many teachers who had long-term civil service contracts and several years of teaching experience in public schools. Within two years, public schools had replaced nearly 40,000 such teachers hired before the policy change, which was 13 percent of all teachers. In addition, after the first two years of the reform, districts continued to replace more than 4,000 such teachers annually, constituting 37 percent of all teachers who left the profession each year. As a result, while the newly hired teachers were presumably more effective, they were significantly less experienced than those employed before the reform. Indeed, four years after the reform began, the share of teachers with fewer than five years of experience rose from 10 percent to 30 percent.
The merit-based teacher-hiring system decreased public school students’ scores on high school exit exams and their college enrollment and graduation rates. In the 15 years following the reform, students’ average scores fell by 8.2 percent of a standard deviation. This figure is roughly equivalent to the decrease in test scores documented by other research following a one-standard-deviation decrease in teacher quality. The decrease in students’ performance is largely driven by negative impacts on mathematics and English scores. However, scores decreased on all exams, including reading, natural sciences, and social sciences. The new teacher-hiring system also reduced college enrollment by 3.3 percentage points, equivalent to reducing the likelihood that a public school student will attend college by 21 percent. Additionally, the reform decreased the likelihood of graduating college by 10 percent among students who had attended public schools after the reform began.
The decrease in exam scores and college attendance appears to have been driven by an increase in students’ exposure to teachers with less experience teaching in public schools. These negative effects were largest for students in public schools with a high fraction of novice teachers. Additionally, the effects were more pronounced among schools with high teacher turnover.
Our findings highlight the risks of using hiring policies that heavily rely on specific measures that seek to predict the quality of candidates’ teaching ability. The reform led to hiring new teachers with better measures of cognitive skills, but these measures did not perfectly correlate with improved student learning. The focus on performance in teacher aptitude exams overshadowed other important criteria, such as teaching experience, which research has shown is crucial for student achievement. Thus, our findings should caution policymakers about the unintended consequences of such reforms, especially for school systems with limited resources.
