Can We Fix the Public Pensions Crisis?

Ranji Nagaswami ’86 is a visiting executive fellow at the Yale School of Management’s International Center for Finance. From 2010 through 2012, she served as chief investment advisor for the New York City Employee Retirement Systems under Mayor Michael Bloomberg. That followed more than two decades in private sector asset management roles, including serving as chief investment officer of the global mutual fund and multi-asset strategies units at AllianceBernstein L.P.

Q: Would you describe the scale of the pension fund crisis?

It’s important to grasp the scale of pensions. Globally, the 13 largest markets account for about $32 trillion in pension assets. In the U.S. alone, it’s about $19 trillion.

Pensions work when employers and employees contribute to a trust through every paycheck. If everyone contributes enough and the assets earn enough through compounding and wise investing, then the pension should be fully funded and available when people retire 20, 30, or 50 years down the road.

The problem today, to put it simply, is that the numbers don't add up.

Q: What about U.S. public pensions in particular?

Within the U.S., the public pension pool is about a quarter of that $19 trillion, and the underfunding problem is quite acute. Based on the actuarial assumptions that the plans use, we have a trillion-dollar underfunding problem. A commonly used assumption is an 8% discount rate—in other words, an expected 8% annual return. Many economists have argued that is overly optimistic, especially in a 3% bond-yield environment. Using a 4% discount rate, which is closer to what is used on corporate pension plans, the underfunding is $3.5 trillion.

The way to think about it is, if you acknowledge the $3.5 trillion underfunding, then the level of annual contributions made by states, local municipalities, and public employees and taxpayers to fund these future benefits needs to be 2.5 times what it is today. That is unfathomable because the individual workers are not people who have a lot of surplus savings to invest; neither can taxpayers afford to foot their majority share of the bill. Relative to a $16 trillion GDP, the dimension of the problem is acute, even if you use the $1 trillion underfunding.

Q: What has contributed to this situation?

I’d argue everybody and everything, including economists and other experts, were complicit in allowing this problem to develop. It is an issue we must debate rather than pretend that if we ignore it things will self-correct.

We have a system for saving during our income-generating years with the aim of assuring a dignified lifestyle during our longer and longer, non-income-generating years. It is in all of our interest that the system work. Pensions or defined-benefit plans are one way we do that. They are important social contracts. We don’t need to get rid of them; we just need to rethink their construction.

The representatives of the public interest on the public pension boards are elected officials. I’d argue that is a conflict of interest. Elected officials need votes and public employees vote in very large numbers. Elected officials are beholden to the beneficiaries of these plans to get reelected. That leads to election-cycle promises of, "Sure. I’ll give you a benefit, now," even though it back-end loads the impact on taxpayers.

A public system will always have tension around balancing cost to taxpayers and needs of public employees. The goal is to manage those conflicts openly. Beyond that, the elected officials and union representatives may be very thoughtful people, but they typically don’t have the skill sets you want managing assets. They certainly aren’t arm's-length decision-making fiduciaries.

Q: What is the impact of an 8% discount rate?

To put the issue in context, for decades interest rates were much higher than they are today. For example, in 1990 interest rates were close to 9%; an 8% return was a fairly conservative number.

At that point, 80 to 90% of pension fund assets were invested in bonds. Pensions are a string of bond-like liabilities. Match them with investments in bonds, and the mismatch is small. Measured at standard deviation, the volatility of those returns on an annual basis was close to 3%. It was a very comfortable world.

Then came the Greenspan era. Interests rates dropped precipitously, bond rates dropped, but the expected return for the public pension system in the U.S. didn't change from 8%.

Also, in the early '90s, there was academic work on the equity risk premium—investors who are willing to part with their cash in the short term can get higher returns in the long term if they are willing to take the temporal risk and risk of downside losses that comes with it. As fiduciaries became more comfortable with equity risk, pensions started moving into higher-risk assets, to the point where, today, about 57% of the entire pension system assets in the U.S. is invested in equities.

That means we've taken on a lot of risk relative to the bond-like nature of the promised pension benefits—the liabilities.

Q: My impression is that in the wake of the very large losses suffered in the financial crisis, the response has been to double down, to take on even more risk to make back those losses. Is that accurate?

The problem that is created by an expected return number that is too high is that to get that return, you start taking bigger and bigger risks. Introducing more and more equities into the portfolio kept that 8% return assumption going. But the volatility of those returns quadrupled. In 1990, the 3% annual volatility meant two thirds of the time the returns would be between 5 and 11%. Fast-forward 15 to 20 years, and plan return volatility is at 12%—so two thirds of the time, you are going to experience returns between -4 and 20%.

Literally in my first days at the New York City pension, I was seeing asset allocations that were likely to be presented to the board that increased equity holdings even more—it was the only way of achieving 8% returns—which would have sent the 12% volatility to even higher levels.

In keeping the 8% chimera alive we accepted more and more and more volatility without realizing it because the '90s were the perfect decade of high growth and moderate inflation. We know that, over time, markets don't grow much more than productivity plus economic activity. We also know that as bubbles are created, the downside can be quite dramatic.

The economy since 2000 has demonstrated that. An economist would argue that was highly predictable, because the variance had gone from 3% to 12%. That's what I mean about having the right voices at the table. If someone says we have to expect big drops with this portfolio, maybe the board says, "Wait a minute. We can't take this kind of risk."

Q: How would you change things?

The question that I think trustees need to answer is, what level of losses are we willing and able to tolerate on the downside? That forces a discussion about risk. The solution isn’t to constantly take on more risk. The solution is to question unrealistic assumptions. When that happens, redesigning the system gets put on the table, as it has been across the country. The message to take away from this is we need to reset how we think about risk, how we think about benefits, and how we think about the intersection of assets and liabilities.

In New York, we dropped that actuarial return assumption to 7%, and even that was higher than what I recommended. At 7% we needed stepped-up contributions by the city, so it was hard to take it lower because the taxpayer ultimately bears that burden.

But, if the promises are based on exaggerated or unrealizable assumptions, beneficiaries also pay because they lose peace of mind, at the time when they need it most, wondering whether their pensions will be paid. If we don't rethink some of the basic economic structural assumptions about pensions, everyone will be hurt.

Q: How could risk be approached differently?

Generally, the discussion ends up being, how much do we put in equities versus bonds? That tends to have a mean-variance-approach-driven answer. But, to me, risk is not standard deviation. Risk is more than the volatility on an annual basis of your return stream. Risk is how much of a downside can you tolerate, how exposed you want to be in various and unpredictable economic environments.

The financial crisis served as a tremendous lesson. Many pension plans around the world lost 30 to 40% of their market value during the crisis. The funded status deteriorated even more sharply because even as the pension fund assets—essentially long equity positions—performed dismally, the pension fund liabilities, basically equivalent to a short bond position, performed very well. Said another way, in effect, the liabilities grew while the assets shrank, making the mismatch even worse than the 30 to 40% loss to the asset values alone.

Preventing future similar drops requires thinking about downside risk, thinking about where you want to set the probability of losing a certain amount of money. An example might be, "I want a 90% probability of not having losses of greater than 15%." Well, that leads to a very different portfolio than the one that experienced a 40% peak-to-trough loss. Pension trustees need to get their heads around this idea of, “How much am I willing to lose relative to the liabilities, and what does that mean for my allocation?”

The second risk that people don't think about enough is what one might call “environmental risk.” Over time, in this quest for higher and higher returns, trustees have moved into more “growth-oriented” investments, essentially equity-like asset classes. They thought they were making diversified bets by spreading investment across stocks, private equity, and real estate; it turns out they weren't. Fundamentally, all of these asset classes are exposed to economic growth. In an environment, like the '90s, where you had higher-than-expected growth and lower-than-expected inflation, those portfolios did exceedingly well, so it became the normative portfolio. That’s the 60-40 portfolio that we have heard about for the last 20 years.

But in an environment like the '70s or '80s, that portfolio actually underperformed quite dramatically, because there was a negative surprise both in growth and in inflation. Another variation is a period where you have low growth, like the decade from 2000 to 2010. Investors were terribly surprised to have poorly performing portfolios. But equities generally do poorly in those environments.

What about an environment where you have lower-than-expected growth and higher-than-expected inflation? It's going to be even worse than the financial crisis, because we have insufficient exposure to assets that do well in steeply inflationary environments. Equally, if you look at the long history of the world, even in the U.S., there are long periods of deflation. The long cycle in the Japanese economy, which we describe as an anomaly, is actually more common than people realize. Yet we are insufficiently exposed with assets that do well in periods of deflation.

Downside risk and environmental exposure are not considered sufficiently in discussions about how assets should be invested. I don't think our portfolios are immunized to a range of economic outcomes that are much more probable than many people realize.

Q: What can pension fund boards do differently?

Governance plays a critical role. I take nothing away from the integrity, commitment, and passion of the representative trustees at the table today, but the issue simply is whether we have the right people at the pension board table to advise and direct the policies of the pension investments while remaining at arm's length from the actual selecting of managers and individual investments to avoid conflicts of interest.

Pension funds are spending more and more time telling corporate boards how they should work, about the importance of having thoughtful, diverse boards that can be objective about governance policies. The tough question I am raising is, have the same funds directed the governance mirror at themselves transparently and with enough accountability to see how they themselves might work better? Academic theory and economic theory have evolved so much, but many pensions seem anchored to a way of doing business that is very dated.

Q: Is there a model for doing this right?

There are many systems that have survived the crisis better. The Dutch, Danish, and Australian systems are examples. The one that people keep referring to, and certainly the one that I have studied the most, is the Canadian model.

To me, a better system means better outcomes for plan beneficiaries— this comes from investment policies, a governance board, and an organizational structure which result in a plan where the assets are less volatile, the funded status has been addressed very specifically as the benchmark, and there is alignment among all stakeholders. This is, perhaps, the most important factor. There must be tremendous alignment of mission, of shared objectives, and of shared views on how best to manage the funds among the employers, the employees, the fiduciaries, and the staff who are charged with managing the assets.

The Canadians, for example, created what they call a Crown Corporation for pensions. These are arm's-length entities that are run as a business—they are not influenced by the electoral cycle, that are not populated by elected officials, but constitute a system that is independently governed by a board, which nominates its successors. The board composition is driven by getting the right stakeholders and also the right skills represented at the table—that means employer and employee representatives but also those with expertise in legal, back office, and investment.

Where the model really differs is the organizational structure. They have brought a lot of the asset management in house. They are direct investors in many asset classes. They are able to recruit high-caliber investment professionals who are attracted to the mission of the organization while pursuing their intellectual and professional interests as investors at the highest levels. They pay these employees at the senior level very well relative to most public salaries, even though they do not attempt to compete with the most exorbitant and, some might say, egregious private-sector salaries. They pay 4 or 5 and sometimes 10 times what a public employee might make, but not the 50 times that a private-sector CEO might make.

On the cost side, bringing the asset management in house reduces costs tremendously because they are not paying agency fees to external managers. To date, they have done an extraordinary job running the Ontario Teachers’ Pension Plan, the Canada Pension Plan, the Healthcare of Ontario Pension Plan, and the Ontario Municipal Employees Retirement System, among other examples.

These pension plans have, more or less, fully recovered, whereas most U.S. public pension funds are nowhere near having fully recovered and particularly not if you use market-based economic assumptions, which, I must add, all of these other plans use.

Now, there are examples, even in the U.S., of plans that are very well managed. The State of Wisconsin Investment Board is definitely an example, where the staff the investment approach and policies, the mission statement, and implementation is world class. TIAA-CREF in the United States is another example of a multi-employer pooled organization that is well managed and has done well for its beneficiaries.

Again, I'm not saying that all public pension funds in the United States need to change, but there are examples all around us of plans that have not learned from the crisis. Even though there are examples of the change working well, change is very hard.

Interviewed by Ben Mattison. Edited by Ted O'Callahan.