POLICY BRIEF: Comparing DB and DC Plans

Scale
Key Policy Insight
In this brief, we evaluate public pension policy tradeoffs in terms of their impact on economic welfare of newly hired public employees. We examine a hypothetical policy change: switching new employees from the existing Defined Benefit pension system to a Defined Contribution system. In our example, wage increases of between 11% and 14% make the new hire indifferent between the legacy DB system and a DC system, contingent on the COLA policy.

I.  INTRODUCTION

Public pension policy focuses on the evaluation of tradeoffs. For any proposed policy change, it is important to identify the affected constituencies and determine how they are affected. This information then helps policy makers evaluate the costs and benefits of any policy change.

This brief is the first of four to evaluate public pension policy tradeoffs in terms of their impact on economic welfare. The brief begins by identifying the main constituencies. It then proposes a hypothetical policy change, and shows the effect of this policy change on one of these groups – newly hired public employees. Other constituencies are addressed in future briefs.

The main policy change investigated in this brief is switching newly hired public employees from the existing Defined Benefit pension system to a Defined Contribution system. Our main result from our model is that a small increase in compensation makes newly hired public employees indifferent between these two systems. The remaining three briefs will explore the impact of this illustrative policy change on the remaining constituencies.

It is important to note that we are not advocating for this particular policy change. Rather, the purpose of this brief and its example is to illustrate how policy choices can be framed and analyzed with modern tools of economic analysis.

 

II.  EVALUATING TRADEOFFS IN PUBLIC PENSION POLICY

By definition, any policy change affects identifiable groups of consumers in different ways. The immediate tasks are to identify the affected constituencies and define the principal characteristics of the current policy. 

For public pensions, the five main constituencies are:

  • Current public employees
  • Retired public employees
  • Taxpayers
  • Newly hired (or soon to be hired) public employees
  • Future generations of public employees and taxpayers

Each of these constituencies will be affected differently by any policy change. A reasonable analytic starting point is to use the current policy as a benchmark. By using the current policy as the benchmark we can ask whether each constituency is made better or worse off by the policy change. Using the tools of economic analysis, we can quantify the degree to which each constituency is made better or worse off relative to the benchmark.

The benchmark policy is the current public pension system. In that system, public employees are contractually promised a fixed retirement income. Historically, some pension systems have provided extra retirement income beyond the contractual level (e.g., a so-called “13th check). Additionally, retirees have had access to cost-of-living adjustments (COLAs) to their retirement income.[1]

Retirement benefits are financed by pension contributions during working years. Typical discussions of pension contributions characterize them as being paid by a combination of the employee or the employer. The level of the contribution depends on the solvency of the pension system – contributions increase when the system is not fully solvent and can decrease when the system is solvent.

To generalize, DB systems can be characterized as having variable contributions (or savings) but mostly fixed retirement income. The open policy question is the impact on each of the five constituencies of changing this mix. We will explore this issue with an illustrative example.

 

III.  SETTING UP A PRACTICAL EXAMPLE

This example assumes that the DB plan is closed to all new hires. Instead, new hires are given a DC plan. The exact characteristics of the plan matter, and will be discussed below. Note that the structure of a DC plan is almost the exact opposite of a DB plan: in our set up, volatile savings and fixed retirement income are traded for a fixed savings rate and more volatile retirement income. There are options in between these two extremes that introduce some volatility in retirement income in exchange for reduced volatility in savings (contributions).[2] Our model is flexible enough to explore those alternatives.

The main question for both policy makers and public employees is how much extra compensation do they need, if any, to be indifferent between the current DB system and a DC alternative. This is the question that we will explore in our illustrative example.

More concretely, our starting point is a DB plan for new hires. We will assume that a new hire:

  • Is paid $40,000 in salary.
  • Has expected wage growth of 2.1% annually (with some volatility).
  • Works a full career, i.e., 45 years (from 20 to 65).
  • Expects a retirement benefit equal to 75% of their final year’s salary. 
  • Contributes to Social Security and receives a fixed Social Security benefit.
  • Makes a mandatory contribution of 11.3% of total compensation to the pension.
  • Can make additional contributions to a separate investment account.

There are two additional benefits that are not guaranteed. The first of these[3] is retirement income in excess of the guaranteed income (e.g. a so-called 13th check), and the second is a COLA. Each of these depends on the level of the pension system’s solvency.[4] For illustrative purposes, we’ll assume that the DB plan is 80% funded (i.e., the value of assets is 80% of the value of pension liabilities). Under this assumption, the value of a potential unexpected increase in retirement income decreases. Finally, we’ll assume that the COLA is 60% of realized inflation.

 

IV.  DC PLANS AS AN ALTERNATIVE – HOW MUCH EXTRA COMPENSATION?

As discussed, a DC plan trades savings volatility for volatility in retirement income. The main question is how much additional compensation does the new hire need to accept this trade, if any. To analyze this issue, we need to first discuss the specific features of the DC plan.

Historically there have been three main issues with DC plans. These are: insufficient savings during working years; improper investing during working years, and no clarity about what to do with accumulated assets during retirement years. Investment managers and consultants have provided options to corporate DB plans for the first two of these. Consequently, this brief will focus only on those two.

DC plan design at large U.S. corporations address the savings and investment issues with three specific policies. First, they automatically enroll employees in the DC plan. Secondly, they include an auto-escalation feature for pension contributions.[5] Finally, they use a so-called Target Date Fund as the default investment option.[6] Importantly, employees can choose to opt out of both of these defaults.

To compare the DC plan option with the DB plan, we need to compare wages and total compensation between the two plans. In our example, we assumed that the entry salary was $40,000. In addition, the plan participant receives a pension benefit, which is financed through a contribution on the plan participant’s behalf to the pension fund. For this example, we calculated the additional contribution as equaling $4,500. Thus, total compensation for the DB worker is $44,500.

For the DC worker, we will assume that they are paid $44,500. A fixed and escalating percentage of total compensation is deducted every year and invested in a Target Date Fund-style investment. Upon retirement, the worker invests their accumulated wealth to support retirement income.

Exhibit 1 shows the results of our analysis. The exhibit compares the economic welfare and wage levels for the DB worker with three different DC worker cases. In the first case, the DC worker receives the same total compensation as the DB worker. As is evident from the exhibit, when the DC worker receives the same total compensation as the DB worker, welfare decreases (albeit modestly).[7]

Exhibit 1

For the second and third DC cases, the worker receives an increase in compensation. In case two, the wage increase is sufficient to match the economic welfare of the DB worker. The exhibit shows that this wage increase is around 11.5%. The third case finds a wage increase so that present value of consumption spending is matched between the DB worker and the DC worker. Not surprisingly, the wage increase is higher in this case, 14.0%.

It is important to note that the required wage increase in the example depends on the initial policy, and in particular, the existence of the COLA. Our example assumed that the beneficiary received a COLA that covered 60% of inflation. That is, if the inflation rate is 2%, the beneficiary would see a 1.2% increase in their benefit. Thus, the provision of the COLA would improve welfare for beneficiaries, but decrease welfare for taxpayers. The magnitude of the welfare changes depends on the level of the COLA.

What Happens to Retirement Income?

The main reason we considered a fourth case is because of the differences in the path of retirement income between the DB worker and the first two DC examples. These are illustrated in Exhibit 2. The exhibit shows that consumer spending is higher during the worker years for the DC worker, but lower during retirement years. The exhibit also shows that when we find a wage that matches retirement income, consumer spending increases even more during working years.[8]

 

Exhibit 2- Expected Path of Consumption

From a beneficiary’s perspective, the main reason we need to increase wages is as a compensation for the potential for extra volatility in retirement income. The DB worker receives a guaranteed fixed benefit plus any additional investment income generated by savings outside of their pension. The DC worker’s retirement income is subject to variation in equity market returns during working years and retirement years. The incremental volatility is compensated for by a wage increase. This point is shown in Exhibit 3. The exhibit shows the average risk in retirement consumption, after normalizing for the level of consumption.[9]

Exhibit 3

 

V.  CONCLUSIONS

This brief has shown how the tools of modern economics can be used to analyze public pension policy changes. We chose to illustrate the use of these tools with a specific policy change  – switching newly hired public workers away from a legacy DB system and into a DC system. We chose to use a DC system as an example not because we are proposing it, but rather because it is an easy example against which to delineate choices. Other types of systems exist that combine the savings (contributions) volatility of DB plans with the retirement income volatility of DC systems. These types of systems can easily be analyzed in our model.

In our example, wage increases of between 11% and 14% make the new hire indifferent between the legacy DB system and a DC system, contingent on the COLA policy. The differences are due mainly to whether expected retirement income in the two systems should be the same. It goes without saying that the results are also dependent on whether a state changes its policy on COLAs. Our analysis assumed that COLAs were paid at the 60% of full inflation. However, paying full COLAs would make the DB system more attractive to the new hire and consequently increase the required additional wage contribution.

Our next two briefs will continue this example and assess the impact of changes in pension policy on current workers, retirees, future generations and taxpayers. To reiterate: the policy changes considered in these briefs are not recommendations, but rather an illustration of how these tools can be used to evaluate policy changes.

[1] COLAs adjust benefit payments for the impact of inflation. The specific COLA adjustment is expressed as a percentage of the inflation rate.
[2] The state of Wisconsin’s hybrid system has volatile savings rates and volatile retirement benefits. Both are less than what is implied by either the DC or DB extremes.
[3] The contribution rate was calculated by (a) finding the present value of DB retirement income, and then determining the savings rate from total compensation that (a) matches the present value of contributions and (b) leaves the initial salary as take-home income. Total compensation is the sum of salary and pension contribution.
[4] The intuition behind why these benefits depend on solvency is straightforward. Public officials always have the option of not giving COLAs or extra retirement benefits. The value of this option decreases when the fund is underfunded, and increases when it is overfunded. Moreover, the value of the DB plan depends partly on the guaranteed benefit and partly on the value of these two options. Hence, the value of the DB plan to the plan participants also depends on the plan’s solvency.
[5] The auto-escalation feature means that savings can start low and increase through the working years. This policy is supported by results from the lifecycle portfolio choice model.
[6] A target date fund automatically reduces the level of portfolio risk as the worker ages. With a TDF, young workers hold more of their assets in equities, while older workers (especially those nearing retirement) hold more of their assets in bonds. This result is also supported by results from the lifecycle portfolio choice model.
[7] The main reason welfare decreases is because the DC system we’ve set up here imposes a constraint on the DC worker’s behavior. Theoretically they would want to rebalance their savings/investment choices and their investment portfolio choices when market conditions change. Our set up, with defaults, precludes that possibility. However, the reason for the defaults is that without them there is a potential moral hazard of workers reaching retirement with insufficient assets.
[8] This result shouldn’t be too surprising. At least anecdotally, in some negotiations between public union officials and policymakers, future pension increases have been offered in lieu of salary increases. It could be argued that current public employees would have preferred higher wages, but settled for future pension benefits.
[9] These results are illustrative, as they depend on (a) the specific wage process; (b) whether the DB worker choses to invest in addition to the DB benefit and (c) the structure of the DC investment glidepath.

AUTHORS

Kurt Winkelmann is a Senior Fellow at the University of Minnesota where he leads the Heller-Hurwicz Economics Institute’s Pension Policy Initiative. He is also founder and CEO of Navega Strategies, LLC. Prior to starting Navega, he was Managing Director and Global Head of Research at MSCI and a Managing Director at Goldman Sachs Asset Management. Kurt earned his Ph.D. in economics at the University of Minnesota, and is the chair of the Heller-Hurwicz Economics Advisory Board.

Jordan Pandolfo worked as a graduate research assistant before earning a Ph.D. in the Department of Economics at the University of Minnesota.

This policy brief was made possible through generous support from Arnold Ventures. Policy briefs from the Heller-Hurwicz Economics Institute are intended to convey policy relevant research done by researchers associated with the Institute and do not represent the position of the Institute or its funders.
Share on: