Are Pensions Keeping the Best Teachers? Data Shows No Effect on Effort or Retention

In one of the most comprehensive studies of public school teachers and pension systems to date, economists Michael D. Bates and Andrew C. Johnston challenge the foundational idea that pensions improve teacher performance and help retain top educators. Using rich administrative data provided by the North Carolina Education Research Data Center (NCERDC), the researchers explore whether retirement benefits near the end of a teacher's career serve as meaningful incentives to boost effort or ensure that the most effective educators stay in the profession. The results, released in a July 2025 NBER working paper, are both surprising and potentially transformative: pensions, they argue, do neither.

Defined-benefit pensions have long been defended as crucial tools in workforce management. The prevailing theory holds that such plans, which reward years of service and penalize early exit, can increase effort through backloaded compensation and create a “selective retention” effect, holding onto more effective teachers who value long-term financial security. Yet when Bates and Johnston analyzed teacher behavior before and after retirement eligibility thresholds in North Carolina, they found no support for these predictions.

Sharp Drop in Compensation, No Drop in Output

In North Carolina, most teachers become eligible for full pension benefits after 30 years of service or at age 60 with at least 25 years of experience. Crossing that eligibility threshold triggers a steep drop in yearly pension accrual, amounting to a 56 percent reduction in effective compensation. Standard economic models, including those based on gift exchange, morale, and deferred compensation, suggest this change should reduce effort. But the study finds no such decline in teacher productivity.

Using value-added metrics, based on student outcomes in math, reading, and behavioral development, the researchers examined teacher output year by year around the eligibility threshold. Contrary to the theoretical expectations, productivity remains steady. Even teacher absences, often considered a sign of effort or disengagement, do not increase. They slightly decline post-eligibility, suggesting that teachers do not slack off after securing their retirement benefits.

This empirical contradiction strikes at the heart of long-standing personnel economics theories, suggesting that the link between pension incentives and worker effort may be far weaker in practice than assumed.

Are Pensions Keeping the Best Teachers? Not Quite

The second hypothesis under scrutiny is whether pensions help retain the highest-performing educators. If better teachers value pensions more, one would expect them to retire at lower rates before hitting the eligibility mark, and possibly retire faster once the financial incentive ends. To test this, the study groups teachers by their prior performance (based on “value-added” scores) and examines their attrition patterns across the pension threshold.

The findings show that attrition spikes once teachers become eligible for retirement, rising from around 2 percent annually to nearly 20 percent, but this pattern is remarkably consistent across high-, medium-, and low-performing teachers. In other words, pensions do not selectively retain talent. The impact of reaching retirement eligibility is uniform regardless of a teacher’s past effectiveness. Whether looking at academic scores or student behavioral outcomes, high-value teachers are no more or less likely to be retained by the pension structure than their peers.

This undermines the common claim that pensions serve as strategic tools to hold onto the best and brightest educators. Instead, pensions appear to influence all teachers similarly, without shaping the quality composition of the workforce near retirement.

A Theoretical Model Without the Dismissal Threat

To further unpack why these pension incentives might fail to elicit effort, the authors construct a refined theoretical model. Traditional deferred compensation models rely on the threat of dismissal, and workers exert effort to avoid losing access to generous retirement benefits. But in the public sector, and particularly in public education, the risk of dismissal is minimal. The study cites dismissal rates of just 1 percent among public school teachers aged 50 and above.

Bates and Johnston offer an alternative model of deferred compensation that doesn’t rely on dismissal risk. They argue that pensions might still motivate effort by fostering reputational concerns and encouraging longer-term relationships between employees and employers. However, this mechanism, too, appears to fall flat in the data. Even when reputational incentives and relationship longevity are factored in, no boost in effort is observed when pension generosity is highest. This suggests that late-career teachers may be largely unresponsive to pension-based incentives, whether financial or psychological.

Rethinking Pensions as a Workforce Strategy

The broader implications of the study are far-reaching. With many states facing massive pension liabilities, this research provides crucial evidence that the returns on these retirement investments, at least in terms of boosting teacher performance or shaping workforce quality, may be overstated. Policymakers who have relied on pensions to manage the late-career workforce might need to reconsider the design and purpose of these plans.

To be clear, the study does not argue that pensions have no value. They may still serve other purposes, such as promoting retirement security or attracting people to the teaching profession early in their careers. However, as tools for increasing effort or ensuring the best teachers stay in the classroom longer, their impact appears negligible.

Ultimately, the findings call for a shift in how educational institutions think about workforce motivation and retention. If the goal is to drive teacher effectiveness and manage career trajectories, alternative strategies, such as performance-based pay or professional development investments, may prove more effective than traditional pension schemes. As the paper makes clear, the assumptions that have long guided pension policy in education deserve a second look.