Privatizing Disability Insurance

The research article by Arthur Seibold, Sebastian Seitz and Sebastian Siegloch has been published in ECONOMETRICA!

We asked Arthur Seibold to share their key insights.

Discover why voluntary private coverage serves as an incomplete substitute for public coverage and what this implies for distributional concerns and social equity in justifying a universal public mandate.

Why have many governments chosen to involve private providers in the delivery of disability insurance as part of their public insurance systems?

Disability insurance (DI) is a social program that helps people who can’t work because of health problems. Across developed countries, the number of people receiving these benefits has been rising quickly. In fact, the administrative data we use show that about one in four German workers will claim public disability insurance at some point in their lives. This rapid growth has put a lot of pressure on governments to cut costs while still protecting vulnerable workers. One way the governments try to do this is by bringing in private insurers, shifting some of the burden to the private market and filling the “gap” left by reductions in public benefits.

And in your research, you are investigating whether this strategy works?

Yes, we’re looking at whether expanding the role of private disability insurance is a realistic and efficient option for governments under fiscal pressure. To do this, we study a 2001 reform in Germany that replaced part of the public disability insurance system with a voluntary private market. Using this reform as a case study, we ask whether private insurance can effectively substitute for public coverage, whether the market works efficiently, and whether the outcomes are fair for everyone.

So what do you find? To what extent does private disability insurance actually address these issues?

Our main finding is that private disability insurance markets cannot fully make up for cuts in public coverage. Even 15 years after the reform, only 26% of workers had purchased private disability insurance. But from a purely economic efficiency standpoint, the plan works better than expected.

So how can the market look efficient if so few people actually buy the insurance?

A major concern with private insurance is adverse selection, where only high-risk people buy insurance, causing the market to fail. However, we found no evidence of adverse selection in this market. The private market successfully covers those who value the insurance most.

And what about fairness and equity?

This is where the strategy may "fail" depending on a government's goals. We document that those who do are typically higher-income, better educated, and face lower disability risk. As a result, many workers who need insurance the most are left without adequate coverage.

How do governments balance the economic efficiency of voluntary private markets with the social equity goals of public mandates when reforming disability insurance?

Our research suggests that if a society cares about redistribution, a public DI mandate remains the best tool, but it must be designed correctly—specifically using income-based contributions. This design effectively redistributes risk and wealth from low-risk, high-income groups to high-risk, low-income groups.

Assessing the potential of private DI in complementing public DI is challenging since complete data on private insurance as well as suitable variation in public DI coverage are needed. How did you overcome this challenge?

To capture all the relevant information, we rely on a new dataset of private disability insurance contracts from a top-10 insurer in Germany, combined with administrative data on all public DI claims.
To study changes in public disability insurance, we take advantage of a reform introduced in 2001 that eliminated part of Germany’s public DI system. We use this exogenous variation in a difference-in-differences framework to identify causal effects. This approach compares how one group responds to a policy change relative to a similar group that was not affected. In our case, the 2001 reform abolished public own-occupation disability insurance for workers born in 1961 or later, who make up the treatment group. Workers born before 1961 were not affected and serve as the control group. By comparing outcomes before and after the reform across these two groups, we can isolate the impact of the policy change.

What implications do your findings have regarding policy?

The results imply that market failures, like adverse selection, are not the primary reason for a government mandate in this context. Instead, the justification for a public system is distributional. Because private insurers charge higher premiums to high-risk workers, these workers often go uninsured. A public system is necessary to ensure they are protected through cross-subsidization. This means that workers in safer and higher-paying jobs contribute more to the system, and part of those contributions helps finance coverage for workers in riskier, often lower-paid occupations. In this way, the system will ensure that those who are most likely to experience disability are still able to afford insurance.

Read the article: Arthur Seibold, Sebastian Seitz and Sebastian Siegloch: Privatizing Disability Insurance In: ECONOMETRICA 2025