Issuers of life-contingent annuities may be making significant overpayments due to gaps in accessible death databases. This can result in the payment of millions of dollars that are not owed, as well as significantly inflated company reserves.

The Declining Reliability of the Death Master File

Historically, issuers of life-contingent annuities have relied on the Death Master File (DMF) maintained by the Social Security Administration (SSA) to verify the survival of their annuitants, rather than relying exclusively on self-reporting by annuitants’ family members and representatives. Although DMF records have never been completely reliable, recent changes in the DMF have compounded the problems associated with monitoring mortality.

Specifically, in 2011, the SSA removed from the public DMF certain protected state records, including certain death certificates, based on privacy concerns and publication restrictions.1 Prior to the record reduction, the historical public DMF contained approximately 89 million records. In 2011, the DMF removed 4.2 million records based on privacy concerns, and then announced that it expected to withhold an additional 1 million records annually for privacy reasons.2 This represents a substantial decrease in publicly available death records. For example, in 2016, only 972,426 deaths were reported to the DMF; yet, in 2000, 2,243,450 deaths were reported to that same database.

The risk of overpaying on a life-contingent annuity is particularly great where the payment recipient is a guardian, conservator, or trustee, because the annuity issuer does not have direct contact with the measuring life, and because minors’ deaths often are unreported. Indeed, in 2000, the DMF reported approximately 17,271 minors’ deaths, which reporting figure fell to approximately 6,092 in 2016.

Supplemental Sources of Information

Other sources of mortality information do exist, but currently those sources are either inaccessible to annuity issuers or themselves incomplete. For example, the National Death Index (NDI), a database of American citizens’ deaths established by the National Center for Health Statistics to help health and medical researchers assess mortality and to inform public policy, is not yet accessible to annuity issuers. Currently, access to the NDI is only available to private entities for medical or clinical research purposes.

Another mortality database is the Electronic Verification of Vital Events system developed by the National Association for Public Health Statistics and Information Systems (NAPHSIS). This electronic verification system allows for searches of all participating states’ death record databases simultaneously. Each of the 57 vital registration jurisdictions in the U.S. (comprising the 50 states, five territories, the District of Columbia, and New York City) has a direct statistical reporting relationship with the National Center for Health Statistics and each is a member of NAPHSIS. Significantly, in March 2017, portions of this database became accessible to insurers for a subscription fee. However, while this system is promising, currently insurers have access to death records for only 31 of the 57 vital registration jurisdictions in the U.S.

Fortunately, third-party vendors can fill some of the voids resulting from the limitations of the DMF, NDI and NAPHSIS. However, annuity issuers should be aware that vendors vary in terms of the sources they use to collect data – they are also subject to restricted access to protected state information and the information they collect is, therefore, incomplete.

The Comprehensive Approach

Given the patchwork nature of existing mortality databases, the prudent course for life-contingent annuity issuers is to use a combination of sources to verify measuring lives’ survival.

Efforts to avoid and recover overpayments can begin with an examination of existing inventories of life-contingent annuities to identify those that are most likely to involve a deceased measuring life (e.g., annuities payable to measuring lives who would now be close to or over 100 years old, or who were severely injured and unlikely to have survived this long). Once potential overpayment situations are identified, a variety of measures can be taken to confirm survival and, ultimately, to recover the overpayments.