The summer issue of The Journal of Retirement includes two articles on different aspects of Social Security reform, a timely article on Medicaid and the financing of long-term care, an article on boosting the savings of employees with 401(k) plans, two articles on investing, and a comprehensive overview of state and local pensions and other post-employment benefits.
The finances of Social Security have been unbalanced for some time. The last major reform to bring them into line was in 1983. Since then, there have been many reform proposals, but few have included measures that would affect the pensions drawn by retirees. In “Generational Equity and Social Security Financing Reform,” Sylvester Schieber makes the case for placing at least part of the burden of a financial adjustment on retirees, thereby alleviating the burden placed on younger Americans. He begins by summarizing the proposals of the Bipartisan Policy Center’s recent Commission on Retirement Income Security and Personal Saving. The Commission’s proposals are not a carbon copy of any other recent major proposal, but its basic structure is similar to that of other proposed reforms. Its major elements include a progressive revision of the benefit formula and a minimum benefit, indexation of the normal retirement age, a cap on the spousal benefit, an increase in the payroll tax rate and the maximum wage or salary subject to tax, a switch to a chained consumer price index (CPI), and the elimination of the favorable taxation of the Social Security benefit. Only the last two affect retired Americans.
The changes to revenues come into play sooner than the changes to benefits, with the result that older workers approaching retirement are not much affected, while young workers bear the brunt of the adjustment. Single individuals and couples in the higher brackets, who have benefited from the spousal allowance, may experience a substantial reduction in net benefits. In general, apart from low-income workers, the current and future generations of workers pay for this reform and reforms like it.
Various arguments have been advanced for holding harmless the current generation of retirees, as well as workers approaching retirement. In particular, it is often argued that retirees, having worked and contributed under the existing set of rules, are entitled to retire under those same rules. In effect, changing the terms of their bargain with Social Security would be unfair, because they have a right to expect a certain treatment. A second argument maintains that retirees have little capacity to return to the labor force. The author takes issue with both arguments. He points out that the current generation of retirees has benefited from the failure of congressional action to rectify the deteriorating finances of Social Security. Had action been taken, their net return from participation in Social Security would have been lower. Whatever the truth of the second argument, it probably applies to older workers as well as retirees. Schieber notes that there is solid evidence that official figures understate the income of older Americans. This point is addressed in an article he wrote with Billy Jean Miller in the winter 2014 issue of The Journal of Retirement.
The author concludes that older Americans, at least those with higher incomes, should bear some of the cost of a reform of Social Security. As he puts it, “those who have benefited from delay (i.e., delay of reform) should help pay those added costs if it does not threaten their retirement income security.” He also notes recent proposals to limit the cost-of-living adjustment (COLA) for those with incomes above some stipulated amount. Finally, a failure to require older Americans to bear part of the financing could make younger generations view Social Security as stacked against them.
Like Schieber, John Turner, in “Social Security Policy Procrastination: A Behavioral Economics Response,” addresses a key aspect of Social Security reform—why Congress has dragged its feet for so long and how insights from behavioral economics might be applied to jump-start reform. Behavioral economics has emphasized the counterproductive role of certain individual behaviors—notably, inertia. Individuals often put off dealing with a problem (e.g., saving for retirement) even when it is clear that they would be better off taking the bull by the horns. Behavioral economics emphasizes the potentially valuable role of default settings, among other measures, in these circumstances. Congress undoubtedly procrastinates. Although Congress is not an individual, the insights of behavioral economics can shed some light on its behavior. Delaying reform—procrastinating—increases costs and prolongs the uncertainty of Americans saving for retirement, but it may also reduce the risk of electoral loss. Inertia—doing nothing—could also change the mix of revenues and expenditures in favor of revenues, making delay more attractive to the party favoring revenue measures.
The relationship between Congress and the electorate can also be viewed as an agency problem, whereby Congress is not effectively carrying out the electorate’s wishes. At the same time, there is little public pressure for reform. Reform of Social Security may have been hindered by the 75-year projection period used by U.S. Social Security Administration (SSA) actuaries. A shorter period might allow measures that would be less drastic, even if a subsequent adjustment would be necessary.
Turner sets out two proposals to prevent a further worsening of Social Security’s finances. Both provide that Social Security will be declared to be in a critical and declining status if the date of trust fund insolvency projected by SSA actuaries drops to less than 15 years. In that event, Congress is expected to propose reforms that will reestablish solvency for a period of at least 20 years. If Congress takes no action within one year, a pre-designed package of measures will be automatically activated. Overriding that activation will require a supermajority.
The author proposes two packages as candidates for automatic activation in the event of congressional inaction. The first includes both revenue and expenditure measures, with an emphasis on safeguarding older retirees. The second package includes only revenue measures. Neither could be regarded as optimal, but they do address the basic problem of insolvency and may inspire Congress to enact a better-designed program. They also reduce the uncertainty Americans face in planning for retirement and, by advancing the date of reform, allow the baby boomer generation to bear part of the adjustment.
The proposal includes certain features inspired by the behavioral approach to economic policy. Recognition of the need for reform (the reform’s salience) is heightened, while the salience of its costs is reduced by facilitating more frequent but smaller adjustments. The default setting could inspire congressional action but will improve Social Security’s finances even if it doesn’t. Avoiding the default solution is made more difficult by the 60-vote requirement for an override.
Medicaid finances nearly half of the nation’s expenditure on long-term services and supports (LTSS). To become eligible for coverage, older Americans must satisfy an income as well as asset test. Both tests vary from state to state, which makes generalization as to their stringency rather difficult. However, the common impression is that these tests are strict enough that only the poor qualify. In “Improving the System of Financing Long-Term Services and Supports for Older Americans,” Mark Warshawsky and Ross Marchand take issue with this view. Their study focuses mainly on the asset test. The limit that states impose on what are known as countable assets are usually no more than $2,000 ($3,000 for a couple), although states exclude the value of a primary residence up to a value that falls in the range of $560,000 to $840,000. The treatment of financial assets—and notably, the balances of retirement accounts—can vary greatly from state to state, however. The authors find that some states count all of these assets; a few ignore them. For the country as a whole, they estimate that 71% of retirement balances were countable in 2011. This figure may, in fact, be an overstatement, because most states apply exemptions for older Americans with access to well-funded accounts.
The cost of Medicaid is increased by the very limited recoveries that states make from the estates of deceased beneficiaries. The authors estimate that in 2002–2011, recoveries never exceeded 0.5% of total LTSS expenditures. It is generally accepted that Medicaid substitutes in some degree for private long-term care (LTC) insurance, to the extent that LTC insurance is typically acquired only by the better off. Since it is possible to qualify for Medicaid-financed LTSS without being truly poor, even middle-income households may have little incentive to pay the substantial premiums of private insurance. Increasing recovery rates and tightening asset limits would increase the cost of Medicaid-financed LTSS, further reducing the net cost to states of the Medicaid LTSS program, and would increase the relative attractiveness of private LTC insurance.
The authors also present evidence that the median net worth of Americans aged 65–67 was about $139,000 in 2012 and the mean value was about $427,000. They argue that a substantial share of the older population has assets that can be devoted to defraying their retirement expenses, including LTSS expenditure should that be necessary. They remark on the widespread holdings of assets even in the first and second quintiles of the distribution of income.
It is generally accepted that most American workers need to save more for their retirement. To this end, employers offering 401(k) and other retirement plans have been encouraged to apply several strategies inspired by behavioral economics. Autoenrollment, whereby employees are automatically covered by a plan unless they explicitly opt out, and autoescalation, whereby the deferral rate is again automatically increased periodically unless an employee explicitly overrules the increase, have both been successful, particularly autoenrollment. Nonetheless, pension coverage of the work force has remained at about 50% for years. Even those workers who do contribute to a plan may be at risk for a financially straitened retirement, because the typical deferral rate is no more than 6%. Depending on the source, either 46% or 77% of plans with autoenrollment have default deferral rates of 3% or less.
In “Save More Today: Improving Retirement Savings Rates with Carrots, Advice, and Nudges,” David Blanchett reports on and analyzes a detailed empirical study based on both plan data and an online survey of plan members. He presents the findings of this work and suggests a novel approach to raising the savings of plan members. In particular, he finds that simply increasing the default savings rate may be an effective way of increasing savings. Roughly one-half of the workforce accepts the default rate regardless of its level. Even those who do not accept the default rate tend to pick a higher rate at higher default rates. This behavior is evident in the plan data and survey responses. (Default survey rates used by the survey could be higher than those taken from the sample of retirement plans.) Interestingly, the probability of active participation in a plan increased with the default savings rate. Stretching out the employer’s match (e.g., by changing a 50% match for the first three percentage points of deferral to a 30% match for the first five points) would also have a positive effect on savings. These are among the most important findings of this substantial work. The author emphasizes increases in the default savings rate as a particularly effective measure, and one that costs the employer nothing. Autoenrollment and autoescalation remain useful measures too, as does in-plan advice.
Target-date funds (TDFs) are designed to reduce the investor’s exposure to more risky assets as time passes and the putative date for retirement draws nigh. In the case of a fund with just stocks and bonds, the share of the fund in stocks is typically reduced once a year by a predetermined amount. An important feature of the standard TDF is that its glide path is not adjusted to reflect changes in market conditions.
In “Target-Date Funds: A Regime-Based Approach,” Mark Kritzman makes the case for an allocation mechanism that would reflect changes in the market’s volatility. He first explains that the probability that a TDF suffers a loss of any size declines with the investment horizon, which implies that as retirement approaches and the horizon diminishes the risk of loss increases. This increase is the rationale for shifting a TDF’s portfolio away from stocks.
The author then argues that a TDF should also be concerned with changes in market turbulence as well as the investment horizon. These should clearly affect the choice of a fund’s assets. He proposes two different measures of market fragility. The first is designed to capture market turbulence and is based on a measure known as the Mahalanobis distance. This indicator is intended to capture extreme movements in asset prices and unusual correlations. A second indicator, the absorption ratio, captures the extent of comovement of asset prices. Tightly coupled prices mean that negative shocks travel more quickly and widely.
In the most technical section of the article, to which the editor must confess that this summary does not do full justice, the author explains how these two indicators are calibrated and how they are used to divide prior periods into fragile and normal periods. The optimal allocation of a portfolio is then calculated for each period, using a covariance matrix calculated for fragile periods and a normal covariance matrix for normal periods. The portfolio’s allocation is shifted whenever there is a regime change. The performance of this regime-sensitive portfolio is then compared with a modified version of a conventional TDF. The modified version optimizes on the basis of the whole prior period, which is not divided in two. In addition, as each year passes, the covariance matrix is calculated using an additional year of data.
One notable difference between the performances of the two TDFs is that the regime-sensitive fund shifts significantly away from risky assets prior to the start of the global financial crisis. The exposure of the modified conventional fund to risky assets remains relatively high for over two years. The performance of the two funds is simulated over 2003–2016; the annualized return of the regime-sensitive fund is 7.22%, compared with 5.75% for the conventional fund. The regime-sensitive fund outperforms the conventional fund in other respects as well.
It is worth noting that the modified conventional fund the author uses for the basis of comparison with the regime-sensitive fund might have a higher rate of return than the conventional fund, whereby the shares of risky and less-risky assets are determined in a purely mechanical way. That said, the apparent superiority of the regime-sensitive fund comes at the cost of a much more complicated algorithm for determining asset allocation. Implementing it might not pose a huge challenge to TDF providers, but explaining its workings to investors might well do so.
The Australian defined-contribution (DC) system, under which all employers above a certain size are now obliged to withhold 9.5% of their employees’ salaries and transfer it to a DC fund, has been in place for a quarter century. As a result, Australians are accumulating substantial balances in their accounts. How to invest these balances during retirement, and how much retirement capital may be safely drawn down each year, are important questions.
In “Adding Direction to the Consumption Rate in Retirement,” Aaron Minney establishes a useful conceptual framework for addressing these questions. A basic decision that needs to be made is whether retirees will choose to spend down all of their capital in retirement or leave a legacy. The author examines two polar (or nearly polar) opposite cases—the choice to leave no legacy at all or to preserve capital for the next generation. He also notes that better-off retirees might choose to accumulate capital during retirement. He derives a basic formula for retirees who aim to leave a legacy with the same value as their initial nest egg and shows that their consumption rate (i.e., the share of capital that may be consumed) will equal the expected rate of return on their capital minus one-half its variance. He also derives a formula based in part on the annually recalculated virtual annuity (ARVA) devised by Waring and Siegel that calculates the consumption rate that could be expected to exhaust capital by the death of the retirees, and which takes mortality and investment risks into account.
The author uses these derivations to illustrate the impact on consumption of a legacy goal versus a consumption-maximization goal. Given a nest egg of $500,000, being prepared to exhaust the nest egg is estimated to increase consumption by about $11,400. Australians receive a means-tested pension call the Age Pension, whose amount would actually increase as capital was depleted, since it is subject to both income and asset tests. Retirees who want to be absolutely certain that they will not run out of money can buy an annuity, which the author estimates would increase their income by $2,000 above the income available with the legacy goal, but at the cost of loss of capital. The author concludes that segmenting the market by legacy preferences could be useful and would facilitate the provision of clear guidance to prospective retirees on the consequences of their decisions. Although the Australian second tier is not a carbon copy of its American counterpart, this article’s succinct analysis should be useful for investors in both countries.
The parlous state of the finances of some of the nation’s pension plans is a staple of business news. In “An Overview of the State and Local Government Pension/OPEB Landscape,” Alicia Munnell and Jean-Pierre Aubry address the question of whether the troubles of these plans are just the tip of the iceberg. They conduct a thorough analysis of the finances of state and local government pensions, other post-employment benefits (OPEB), and debt service. The coverage of this survey is very broad, comprising 50 states, 178 counties, 174 major cities, and 415 school districts.
The data on the pension plans of these entities were not always consistently derived: For example, plan discount rates ranged from 6.25% to 8.5%, and data on the actuarial required contribution (ARC) were not always available. The authors chose a uniform discount rate of 6%—which would tend to increase recorded liabilities—and derive consistent estimates for each plan’s normal cost and ARC. They first shift the obligations of local plans participating in cost-sharing plans from the states to the localities. Estimates of a plan’s unfunded liability were recalibrated using the chosen discount rate. The required plan payments thus calculated were expressed as a percentage of the concerned government’s own revenues (total revenues minus federal government transfers plus transfers to lower governments). This exercise makes clear the difficult position of a small number of states. Most states, however, have a ratio that is substantially less than 10%, and 24 states are below 5%. The situation for large cities and counties is less encouraging.
The estimates of the burden posed by OPEB liabilities were also calculated with an interest rate of 6% compared with an actual OPEB average of 4.8%. Consequently, the estimates of OPEB liabilities were reduced by the choice of the higher discount rate. The pattern of the distribution of OPEB payments across states is similar to the pattern of pension payments. A handful of states are incurring OPEB payments above 5% of own revenues, but nearly all are less than 5%, and many are close to zero.
As for the total burden—that is, the sum of required pension and OPEB payments and interest payment in relation to own revenues—the states with heavy pension obligations are those with high total payments. Illinois, Connecticut, New Jersey, and Kentucky have total payments that exceed 25% of own revenues. However, most states have payments that are less than 15% of own revenues. The diverse experience of the states is paralleled to a large extent by the nation’s large counties. Eight have a payment ratio of over 40% of own revenues (which can be small compared with transfers from their state), of which six are in California. The ratio of most large counties is less than 30%, however.
Finally, Meir Statman, a major contributor to the literature on behavioral investment theory and a member of The Journal of Retirement’s advisory committee, has written a substantial study on the behavioral approach to financial decision making—Finance for Normal People. I would like to invite readers to take a look at my review of his book, which is the last article of this issue. One whole chapter is devoted to retirement preparation. Overall, I found the book to be both comprehensive and enlightening.
TOPICS: Retirement, social security, legal/regulatory/public policy
George A. (Sandy) Mackenzie
Editor
Footnotes
Publisher’s Note
Institutional Investor, the publisher of The Journal of Retirement, wants to extend a special thank you to State Street Global Advisors for its support of The Journal of Retirement. Please note that no sponsor has influence on the editorial content found in The Journal of Retirement. Representatives from any firm are encouraged to submit an article to our independent Editor, Sandy Mackenzie, for review and prospective acceptance into the publication. All editorial submissions, acceptance, and revisions are the sole decision of Mr. Mackenzie. The editorial submission guidelines are found on the last page of the publication and online at www.iijor.com. Thank you, and I hope you enjoy this and future issues of The Journal of Retirement.
Dave Blide
Publisher, Institutional Investor Journals, dblide{at}iijournals.com
- © 2017 Pageant Media Ltd