ISR|SRC|Michigan Retirement Research Center|UofM
Mailing List | News | Contact Us
Follow onTwitter
Main photo Promoting research on retirement and Social Security policy

Volume 17, Issue 1 - March 2017

Key findings from 2016 projects touch on work transitions and more

Social Security and Retirement

Occupational Transitions at Older Ages: What Moves are People Making? by Amanda Sonnega, Brooke Helppie McFall, and Robert J. WillisWP 2016-352

This paper uses data from the Health and Retirement Study (HRS) to examine occupational transitions at the detailed level toward the end of working life.

Among workers who transition between occupations, the most common transitions are between those that are closely related. However, even within closely related occupations, there are no large pipelines between any two.

By age 62, 57 percent of workers are no longer in the labor force, 26 percent are still in their “career” occupation, and 17 percent have changed from their career occupation to another occupation. Beginning at age 66, however, the percentages in different occupations, which may be bridge employment or unretirement, are very similar to the percentages remaining in career occupations.

Occupational changes later in life tend to be accompanied by decreases in hourly earnings, suggesting that if workers are seeking flexible or part-time bridge employment, it may come at a cost.

Do State Laws Protecting Older Workers from Discrimination Laws Reduce Age Discrimination in Hiring? Experimental (and Nonexperimental) Evidence by David Neumark, Ian Burn, Patrick Button, and Nanneh ChehrasWP 2016-349

Age discrimination may make it difficult for policymakers to increase employment of older workers.

Experimental data that directly measures age discrimination does not provide clear evidence that stronger laws protecting older workers from discrimination reduce measured hiring discrimination against older workers.

Nonexperimental data indicate that the relative hiring of older workers is higher in states with stronger protections against disability discrimination. If this relationship is causal, it could imply that even though these stronger laws do not reduce measured discrimination against older workers, they could increase employment of older workers, perhaps by changing who looks for work.

Older People’s Willingness to Delay Social Security Claiming by Raimond H. Maurer and Olivia S. MitchellWP 2016-346

In our HRS module, we show that many older Americans would be willing to delay claiming their Social Security benefits if they were offered an actuarially fair lump sum to do so.

Overall, half of the respondents say they would delay claiming if no work requirement were in place under the status quo, and only slightly fewer with a work requirement.

If no work is required, the average amount needed to induce delayed claiming is about $60,400, while when part-time work is required, the average is $66,700. This implies a low utility value of leisure foregone of only $6,300, or less than 20 percent of average household income.

Delayed claiming would have a positive effect on
their retirement security, in that their Social Security income streams rise per year of delay. Indeed benefits, claimed at age 70 are more than 75 percent higher than at age 62.

The Interaction between Consumption and Health in Retirement by John Karl Scholz and Ananth Seshadri WP 2016-344

We study the interaction between consumption and health in retirement. Our main contribution is the estimation of a consumption Euler equation taking health into consideration.

We find that health shocks play an important role in slowing down the decline of consumption with age in retirement. Without health shocks, retirees will run down their wealth at a much higher speed.

We also find that including health into the utility function provides interesting interactions between health and consumption and could help explain the heterogeneous consumption-age profiles related to health.

Finally, we find that health investments, such as physical exercise, have a significant effect on the evolutions of both health and consumption in retirement.

The findings in this paper suggest that public programs, such as Medicare and Medicaid, play an important role in shaping the consumption and health behavior of the retirees, as well as the retirement and saving decisions of working people.

Wealth and Retirement Income

Consumption Smoothing During the Financial Crisis: The Effect of Unemployment on Household Spending by Michael Hurd and Susann RohwedderWP 2016-353

We found that at unemployment, total monthly spending declined rapidly from the $3,560 it averaged during employment to about $2,980, a decline of about 17 percent.

Spending remained at that level until about 30 weeks of unemployment when it declined further to about $2,500, or around 70 percent of initial spending.

Spending on big-ticket items and other, infrequently purchased items decreased more rapidly than high-frequency expenditures such as those on food.

Income decreased much more rapidly than spending to about 37 percent of pre-unemployment levels where it remained unless the duration of the unemployment spell was very long. An implication is that some spending while unemployed was financed out of savings or accumulation of debt.

At re-employment, increased rapidly, much more rapidly than spending.

An overall conclusion is that households were able to maintain spending following unemployment at a fairly high percentage of pre-unemployment spending provided the duration of unemployment was moderate.

How Home Equity Extraction and Reverse Mortgages Affect the Credit Outcomes of Senior Households by Stephanie Moulton, Donald Haurin, Samuel Dodini, and Maximilian D. Schmeiser
WP 2016-351

Senior homeowners can extract home equity through different borrowing channels, and the channel through which they borrow equity is associated with different credit trajectories both before and after extraction.

Senior homeowners extracting equity through a home equity line of credit (HELOC) tend to have the strongest credit profiles of all types of borrowers prior to extraction that remain strong post extraction (e.g., high credit scores, low rates of delinquency). By contrast, senior homeowners extracting equity through a reverse mortgage had lower credit scores at the time of origination and were more likely to have had a credit shock prior to getting their loan than other borrowers and nonborrowers. The credit shock is observed as a drop in credit score of 25 or more points and an increase in credit card balances within the two years prior to loan origination.

While there is some decline in nonhousing debt across all borrowing channels in the three years after extracting from home equity, the decline is the greatest among seniors extracting home equity through a reverse mortgage, and particularly for credit card debt.

Cohort Changes in Social Security Benefits and Pension Wealth by Chichun Fang, Charles Brown, and David Weir WP 2016-350

The near-elderly population in the most recent Health and Retirement Survey (HRS) cohort is not as financially prepared for retirement as their peers in earlier cohorts. In relative terms, retirement wealth does not grow as fast as lifetime earnings; in absolute terms, retirement wealth declines among men and minorities since the turn of the century.

There has been a stable shift from defined benefit (DB) plans to nonannuitized retirement wealth such as defined contributions (DC) or individual retirement accounts (IRA). The increase in DC and IRA was not enough to make up the loss in DB, which was reflected in the decline in retirement wealth. As a result, the proportion of Social Security benefits in total retirement wealth was increasing, especially among minorities.

The HRS used different earnings projection methods to produce Social Security and pension wealth estimates. The earnings projection in Social Security wealth estimation requires administrative data linkage, but performs reasonably well for those who stay in the same job in 2010. The earnings projection in pension wealth does not require data linkage and is easier to implement, but pension wealth estimates in 2004 were affected by larger errors in earnings projection for the same group of respondents.

Time Discounting and Economic Decision-making among the Elderly by David Huffman, Raimond H. Maurer, and Olivia S. Mitchell WP 2016-347

We find that implied Internal Rates of Return (IRR) used by our older respondents to discount future payments is 0.54.

IRRs rise with age so that a 15-year increase in age from 70 to 85 would be associated with about a one standard deviation higher IRR.

Whites and the better-educated have lower IRRs, while people with serious health conditions implying reduced life expectancy have 11 to 30 percent higher IRRs. Also IRRs are 35 percent higher than average for individuals diagnosed with a cognitive condition (dementia or Alzheimer’s).

Net wealth is significantly lower for the least patient individuals. Additionally, the impatient are much less likely to engage in healthy behaviors and make little provision for end-of-life challenges.

Program interactions

Long-term Individual and Population Consequences of Early-life Access to Health Insurance by Etienne Gaudette, Gwyn Pauley, and Julie Zissimopoulos
WP 2016-355

Gaining access to health insurance in childhood had long-term consequences for adult well-being. Compared to adult health and economic outcomes for individuals without childhood health insurance, gaining health insurance in childhood resulted in:

A decline in the prevalence at age 65 of most major diseases, excluding cancer;

An increase in life expectancy of 11 months and 16 additional months lived free of disability;

No change in lifetime total medical spending as increases in spending due to additional life years were offset by lower health care spending due to improved health;

A decline in Medicare and Medicaid outlays and expenditures on lifetime Disability Insurance;

An increase in lifetime earnings of approximately
8 percent.

The Effect of the Affordable Care Act on the Labor Supply, Savings, and Social Security of Older Americans by Eric French, Hans-Martin von Gaudecker, and John Bailey JonesWP 2016-354

We construct a retirement model that includes health insurance, uncertain medical costs, a savings decision, a non-negativity constraint on assets, and a government-provided consumption floor. We model the ACA as a change in government insurance provisions rather than the provision of insurance where none existed before.

We present evidence that those who cannot keep their employer-provided health insurance when they leave their job tend to remain on their job until age 65.

Those who can maintain their insurance after they leave their job tend to exit the labor market earlier. This provides evidence that access to health insurance reduces labor supply.

We show differences in both total and out-of-pocket medical spending prior to the enactment of the ACA. We show that average total medical spending is high for all groups: Those with no health insurance do not spend much more out-of-pocket than those who have private insurance.

Those uninsured receive health care through a variety of sources, such as worker’s compensation and default on medical bills, which we refer to as a “consumption floor,” which protects low-income individuals against catastrophic medical spending.

Those who appear to have the highest resources appear to be those who pay the most for health care, consistent with the view that those with low resources are covered by the consumption floor, whereas those with high resources face the most medical expense risk and might have the largest labor supply responses.

Health Reform and Health Insurance Coverage of Early Retirees by Helen Levy, Thomas C. Buchmueller, and Sayeh NikpayWP 2016-345

Between 2013 and 2014, the fraction of early retirees without health insurance declined from 14.7 percent to 11.2 percent, reversing a trend toward increasing uninsurance in recent years.

Gains in coverage among early retirees were driven by increases in both Medicaid and private, nongroup coverage.

Gains in coverage were larger in states that implemented the Affordable Care Act’s Medicaid expansion in January 2014 than in states that did not.

The gains in coverage disproportionately benefited low-income early retirees and therefore reduced the gradient in coverage with respect to income.

There is no evidence of an acceleration of the decline in employer-sponsored coverage for early retirees, either overall or in states that expanded Medicaid.

The Affordable Care Act as Retiree Health Insurance: Implications for Retirement and Social Security Claiming by Alan L. Gustman, Thomas L. Steinmeier, and Nahid Tabatabai WP 2016-343

Using data from the Health and Retirement Study,
we find no evidence that for those with health insurance at work but not in retirement, the Affordable Care Act (ACA) increased retirements over the period 2010 to 2014.

We also find no evidence that the ACA changed retirement expectations or expected age of Social Security benefit claiming of those who, before ACA, had health insurance coverage when working, but not when retired.

An analysis based on a structural retirement model suggests that eventually ACA will increase the probability of retirement by those who initially had health insurance on the job but did not have employer-provided retiree health insurance by about half a percentage point at each year of age.

The structural retirement analysis also suggests that much of the effect of ACA on retirement will be realized within a few years of the change in the law.

The Retirement and Social Security Benefit Claiming of U.S. Military Retirees by David Knapp, Beth Asch, James Hosek, and Michael MattockWP2016-336

Theoretically, under alternative exploratory assumptions, the income effect will tend to dominate at higher levels of the military pension, implying that the military retirement benefit increases retirement.

Empirically, the income effect is also the dominant effect; a higher military retirement benefit increases the likelihood of retiring from the labor force among Medicare-eligible military retirees, conditional on still working full time at age 64.

The difference-in-difference estimate is statistically significant and large, implying a 1 percent increase in the monthly benefit would result in a 5.9 percent greater likelihood of retirement at ages 65 and older.

Expansion of the military pension is also associated with delayed claiming of Social Security benefits, though our estimates are not statistically significant.

Additional research is needed, preferably using administrative military personnel data merged with Social Security data, permitting larger sample sizes for analysis.


The Dynamic Effects of Health on the Employment of Older Workers by Richard Blundell, Jack Britton, Monica Costa Dias, and Eric FrenchWP 2016-348

The dynamic properties of health are well described by the sum of a highly persistent AR(1) component, plus a transitory component.

Transitory health shocks have little impact on employment.

Permanent health shocks have much bigger effects on employment.

Employment is highly persistent. Lagged employment strongly predicts current employment, even after accounting for the persistence in health.

Model estimates suggest a larger impact of health
on employment than what ordinary least squares estimates imply.

Means Testing Social Security: Modeling and Policy Analysis by Rafal Chomik, John Piggott, Alan D. Wood-land, George Kudrna, and Cagri Kumru WP 2016-337

Contrary to conventional wisdom, our analysis
has found that means testing delivers fewer overall distortions and increased welfare compared to a universal pension offering the same maximum
benefit level.

Compared to an OECD average of 9.5 percent of GDP allocated to pension spending, Australia’s annual spend is 2.9 percent of GDP. Hence, the comparison with many OECD countries reveals that Australian pension arrangements are cheaper both at present and based on future projections. The lower cost is largely due to the flat-rate benefit and means testing.

Means testing also increases incentives for self-provision. Reduced or zero public benefits payable to the richest groups of retirees can improve labor force participation for that group. Lower public spending on pensions can lead to higher aggregate labor supply — modeled to be 1.4 percent higher with means testing than with a universal pension.