Abstracts of Selected Recent NBER Working Papers

Abstracts of Selected Recent NBER Working Papers

NBER Working Paper 12762
Beth J. Soldo, Olivia S. Mitchell, Rania Tfaily, John F. McCabe
Cross-Cohort Differences in Health on the Verge of Retirement

Baby Boomers have left a unique imprint on US culture and society in the last 60 years, and it might be anticipated that they will also put their own stamp on retirement, the last phase of the life cycle. Yet because Boomers have not all fully retired, we cannot yet judge how they will fare as retirees. Instead, we focus on how this group compares with prior groups on the verge of retirement, that is, at ages 51-56. Accordingly, this chapter evaluates the stock of health which Early Boomers bring to retirement and compare these to the circumstances of two prior cohorts at the same point in their life cycles. Using three sets of responses from the Health and Retire-ment Study, we find some interesting patterns. Overall, the raw evidence indicates that Boomers on the verge of retirement are in poorer health their counterparts 12 years ago. Using a summary health index designed for this study, we find that those born 1948 to 1953 share health risks with the War Baby cohort. This suggests that most of the health decline instead began before the late 1940's. A more complex set of health conclusions emerges from the specific self-reported health measures. Boomers indicate they have relatively more difficulty with a range of everyday physical tasks, but they also report having more pain, more chronic conditions, more drinking and psychiatric problems, than their HRS earlier counterparts. This trend portends poorly for the future health of Boomers as they age and incur increasing costs associated with health care and medications. Using our health index, only those at the 75th per-centile or higher are likely to be characterized as having good or better health.

NBER Working Paper 12764
Dhaval Dave, Robert Kaestner
Health Insurance and Ex Ante Moral Hazard: Evidence from Medicare

Basic economic theory suggests that health insurance coverage may cause a reduction in prevention activities, but empirical studies have yet to provide evidence to support this prediction. However, in other insurance contexts that involve adverse health events, evi-dence of ex ante moral hazard is more consistent. In this paper, we extend the analysis of the effect of health insurance on health be-haviors by allowing for the possibility that health insurance has a direct (ex ante moral hazard) and indirect effect on health behaviors. The indirect effect works through changes in health promotion information and the probability of illness that may be a byproduct of insurance-induced greater contact with medical professionals. We identify these two effects and in doing so identify the pure ex ante moral hazard effect. This study exploits the plausibly exogenous variation in health insurance as a result of obtaining Medicare cover-age at age 65. We find limited evidence that obtaining health insurance reduces prevention and increases unhealthy behaviors among elderly persons. There is more robust evidence that physician counseling is successful in changing health behaviors.

NBER Working Paper 12803
Allan Begosh, John Goldsmith, Ed Hass, Randall W. Lutter, Clark Nardinelli, John A. Vernon
Black Box Warnings and Drug Safety: Examining the Determinants and Timing of FDA Warning Labels

Comparing the safety of prescription drugs over time is difficult due to the paucity of reliable quantitative measures of drug safety. Both the academic literature and popular press have focused on drug withdrawals as a proxy for breakdowns in the drug safety sys-tem. This metric, however, is problematic because withdrawals are rare events, and they may be influenced by factors beyond a drug's safety profile. In the current paper, we propose a new measure: the incidence and timing of Black Box Warnings (BBWs). BBWs are warnings placed on prescription drug labels when a drug is determined to carry a significant risk of a serious or life-threatening adverse event. Using a unique data set, one that includes all new molecular entities (NMEs) submitted to the FDA be-tween May 1981 and February 2006, and subsequently approved and marketed, we analyze the timing and incidence of BBWs. Our analyses also use data on several drug characteristics likely to affect the probability a new drug will receive a BBW. We draw several conclusions from our analyses. For example, drugs receiving priority FDA review are more likely to have BBWs at the time of approval than NMEs receiving standard review. We also find that early prescription volume and orphan drug status are associated with an in-creased likelihood of receiving a BBW. We do not, however, find a significant difference in the rate of BBWs across time cohorts. A comparison of NMEs approved before and after the 1992 Prescription Drug User Fee Act (PDUFA), which authorized the payment of user fees from drug manufacturers to the FDA in an effort to expedite new drug application (NDAs) review times, did not reveal a sta-tistically significant difference in the rate of BBWs. Critics of PDUFA maintain that reduced FDA-approval times under PDUFA have compromised drug safety. We do not find empirical support for this contention.

NBER Working Paper 12805
Alan J. Auerbach, Ronald Lee
Notional Defined Contribution Pension Systems in a Stochastic Context: Design and Stability

Around the world, Pay-As-You-Go (PAYGO) public pension programs face serious long-term fiscal problems due primarily to actual and projected population aging, and most appear unsustainable as currently structured. Some have proposed the replacement of such plans with systems of fully funded private or personal Defined Contribution (DC) accounts, but the difficulties of transition to funded systems have limited their implementation. Recently, a new variety of public pension program known as "Notional Defined Contribution" or "Non-financial Defined Contribution" (NDC) has been created, with the objectives of addressing the fiscal instability of traditional plans and mimicking the characteristics of funded DC plans while retaining PAYGO finance. Using different versions of the system recently adopted in Sweden, calibrated to US demographic and economic parameters, we evaluate the success of the NDC approach in achieving fiscal stability in a stochastic context. (In a companion paper, we will consider other aspects of the performance of NDC plans in comparison to traditional PAYGO pensions.) We find that the basic NDC scheme is effective at preventing excessive debt accumulation, but does little to prevent significant asset accumulation along many trajectories and on average. With adjustment, however, the NDC approach can be made more stable.

NBER Working Paper 12820
M. Kate Bundorf, Melinda Henne, Laurence Baker
Mandated Health Insurance Benefits and the Utilization and Outcomes of Infertility Treatments

During the last two decades, the treatment of infertility has improved dramatically. These treatments, however, are expensive and rarely covered by insurance, leading many states to adopt regulations mandating that health insurers cover them. In this paper, we explore the effects of benefit mandates on the utilization and outcomes of infertility treatments. We find that use of infertility treatments is significantly greater in states adopting comprehensive versions of these mandates. While greater utilization had little impact on the number of deliveries, mandated coverage was associated with a relatively large increase in the probability of a multiple birth. For rela-tively low fertility patients who responded to the expanded insurance coverage, treatment was often unsuccessful and did not result in a live birth. For relatively high fertility patients, in contrast, treatment often led to a multiple, rather than a singleton, birth. We also find evidence that the beneficial effects on the intensive treatment margin that have been proposed in other studies are relatively small. We conclude that, while benefit mandates potentially solve a problem of adverse selection in this market, these benefits must be weighed against the costs of the significant moral hazard in utilization they induce.

NBER Working Paper 12837
Rodrigo R. Soares
On the Determinants of Mortality Reductions in the Developing World

This paper presents and critically discusses a vast array of evidence on the determinants of mortality reductions in developing countries. We argue that increases in life expectancy between 1960 and 2000 were largely independent from improvements in income and nutrition. We then characterize the age and cause of death profile of changes in mortality and ask what can be learned about the de-terminants of these changes from the international evidence and from country-specific studies. Public health infrastructure, immuniza-tion, targeted programs, and the spread of less palpable forms of knowledge all seem to have been important factors. Much of the recent debate has revolved around antagonistic approaches, which are not supported by the evidence discussed here. Finally, the paper suggests that the evolution of health inequality across and within countries is intrinsically related to the process of diffusion of new technologies and to the nature of these new technologies (public or private).

NBER Working Paper 12842
Jeffrey R. Brown, Scott J. Weisbenner
Who Chooses Defined Contribution Plans?

This paper provides new evidence on what types of individuals are most likely to choose a defined contribution (DC) plan over a defined benefit (DB) plan. Making use of administrative data from the State Universities Retirement System (SURS) of Illinois, we study the decisions of nearly 50,000 new employees who make a one-time, irrevocable choice between a traditional DB plan, a portable DB plan, and an entirely self-managed DC plan. Because the SURS-covered earnings of these employees are not covered under the Social Security system, their choices provide insight into the DB vs. DC preferences of individuals with regard to a primary source of their retirement income. We find that a majority of participants fail to make an active decision and are thus defaulted into the traditional DB plan after 6 months. We also find that those individuals who are most likely to be financially sophisticated are most likely to choose the self-managed DC plan, despite the fact that, given plan parameters, the DC plan is inferior to the portable DB plan under reason-able assumptions about future financial market returns. We discuss both rational and behavioral reasons that might explain this find-ing.

NBER Working Paper 12858
Jonathan Gruber, Kosali Simon
Crowd-Out Ten Years Later: Have Recent Public Insurance Expansions Crowded Out Private Health Insurance?

The continued interest in public insurance expansions as a means of covering the uninsured highlights the importance of estimates of "crowd-out", or the extent to which such expansions reduce private insurance coverage. Ten years ago, Cutler and Gruber (1996) suggested that such crowd-out might be quite large, but much subsequent research has questioned this conclusion. We revisit this issue by using improved data and incorporating the research approaches that have led to varying estimates. We focus in particular on the public insurance expansions of the 1996-2002 period. Our results clearly show that crowd-out is significant; the central tendency in our results is a crowd-out rate of about 60%. This finding emerges most strongly when we consider family-level measures of public insurance eligibility. We also find that recent anti-crowd-out provisions in public expansions may have had the opposite effect, lowering take-up by the uninsured faster than they lower crowd-out of private insurance.

NBER Working Paper 12870
Dana Goldman, Tomas Philipson
Integrated Insurance Design in the Presence of Multiple Medical Technologies

The classic theory of moral hazard concerns the insurance of a single good and predicts that co-insurance is larger when the elasticity of demand is higher and when small risks are insured. We extend this analysis to the insurance of multiple goods; for example, the simultaneous insurance of medical services and prescription drugs. We show that when multiple goods are either complements or substitutes--so that a change in co-insurance for one service affects the demand of others--the classic moral hazard results do not hold. For example, the single good model would predict high co-payments for prescription drugs since drug demand is elastic and of modest financial risk. However, a model of multi-good insurance suggests such drug coverage may optimally involve zero or even negative co-insurance when it is a substitute to other services insured such as hospital care or physician services. We summarize some of the empirical evidence in support of markets adopting optimal integrated pricing structures rather than individually optimal pricing structures.


NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us