The federal government is not quite ready for David Swensen.

David Swensen was Chief Investment Officer (CIO) for the Yale University Endowment for roughly 45 years, from 1985 until his untimely death in 2021. When he started as CIO, the endowment totaled about $1 billion; by 2021, the endowment exceeded $42 billion (and that is net of more than $20 billion in outflows over those years to support the funding needs of the university). That’s an annualized return of more than 13% - outpacing a simple 70% equity/30% fixed income portfolio return of about 9%. In dollar terms, had Yale just followed the simple 70/30 portfolio strategy, the endowment in 2021 would have been closer to an estimated $20 billion (less than half compared to the $42 billion that Swensen achieved).

How did he do this?

Swensen introduced what has since become known as the “Yale Model,” a multi-asset class portfolio that relies heavily on private equity (e.g., venture capital, buyout), real assets (e.g., cash-flowing assets such as real estate), and marketable alternatives (e.g., non-correlated hedge funds). The fundamental theory of the Yale Model is that – if you have a long investment time horizon (i.e., an endowment is expected to last essentially forever) and a sophisticated team who can access and appropriately price investments that are not marketable daily – you can meaningfully outperform daily-liquid equity and fixed income assets.

As Swensen showed, the model worked.

And over the past 25 or so years, the Yale Model has been largely adopted as the gold standard by most institutional investors, albeit with some variations in the actual percentage of portfolios allocated to alternatives.

There is some debate today about whether the Yale Model can persist for another 25 years, but what is not up for debate is the fundamental principle of diversified asset allocation. That is, building a portfolio with exposure to multiple asset classes is generally recognized as the best way for institutional investors to achieve long-term equity growth, coupled with a sufficient ballast of non-correlated assets to weather the inevitable storms that will come periodically. Nobel Prize winning economist Harry Markowitz summed it up nicely: “Diversification is the only free lunch in investing.”

But the federal government is not a Swensen acolyte

The federal government manages a number of trust funds, but the two largest are the Social Security Trust Fund (about $2.7 trillion today) and the Civil Service Retirement and Disability Funds (CSRDF) (about $1.1 trillion today), the latter of which we at the Office of Personnel Management (OPM) have the privileged of overseeing.

CSRDF provides the funds to pay out the pensions for federal employees. Current obligations are funded by federal agency and employee contributions, but any excess goes into the fund itself to grow for the benefit of pensions to be paid out in the many years to come. It has positive cash flow today (meaning we take in more money from employee/agency contributions and returns on the investments than we pay out in current pension obligations) but there is a shortfall in the out years from previous years of underfunding. To offset this shortfall, we will either need to earn a higher rate of return on our investments or increase the amount of contributions that are made by the federal government and its employees.

By statute these trust funds are allowed to invest only in Treasury bonds; we are not legally permitted to invest in a diversified portfolio. So, 100% of the funds are invested in various maturities of treasuries – compared with about 10-15% that the Yale Model would traditionally allocate to treasuries.

What does that mean in terms of returns on the portfolio? Well, if Swensen achieved about 13% annually throughout his tenure, and a simple 70/30 portfolio has returned about 9%, the federal government trust funds have averaged about a 3-4% return.

What does that return delta mean in terms of actual dollars? Let’s look at the Social Security Trust Fund as an example.

As you know if you follow the news, the Social Security Trust Fund is at risk of running out of money. It’s $2.7 trillion in assets are forecast to run out some time mid-2030’s; that’s not too far from now. However, instead of investing simply in treasuries, had we invested in a simple 70/30 portfolio, the trust fund would have a current balance of between $10 and $20 trillion today. That would mean we could likely extend the life of the fund for up to another 30 years past the current forecast expiration date!

So, what can we do if we can’t diversify into other asset classes?

It’s unlikely we are going to get Congress to eliminate the Treasury-only investment restrictions on the Social Security Trust Fund or the CSRDF. One can dream, but we all know that hope is not a strategy. However, all hope is not lost – and we at OPM have partnered with the Treasury Department to change what we can to the benefit of the American taxpayer.

Enter “duration.”

A very quick lesson on the economics of treasuries (and bonds more generally). Treasuries have varying expiration dates (or “duration,” as the finance nerds prefer). For example, you can buy Treasury bills (less than 1-year in duration), Treasury notes (with 2, 3, 5, and 7-year durations) and Treasury bonds (with 20 and 30-year durations). It’s not always the case, but in general the longer the duration, the higher the interest rate the government will pay you on the bond. Today, a 30-year duration bond pays about 5% in annual interest, whereas the 10-year pays 4.5% and a 1-year pays 4%.

Historically, the CSRDF has had an average duration across its entire Treasury portfolio of about 7.5 years. That means, that we had lots of different maturities, ranging from less than 1-year to up to 15-years, but the average was right in the middle. As a result, we have achieved the roughly 3-4% annualized returns that I mentioned above on the overall fund balance.

Why ask why?

Consistent with our cultural precepts at OPM, we decided to ask why? Why did we not have longer dated bonds (e.g., 20 and 30-year durations) and why did we have an average duration of only 7.5 years (versus a longer one to reflect the true period over which our funds will get paid to retirees)?

And it turned out the answers to those questions were not very inspiring. The short answer is that nobody had really looked hard at the investment portfolio and returns for a very long time – it was essentially operating on autopilot for the last 45 years.

And, even more interesting, regarding the question of why we don’t have a longer duration and, in particular more 30-year bonds, the answer was really illuminating. It turns out the 30-year bond wasn’t even introduced by the government until 1977 and it didn’t really become a staple investment alternative until the mid-1980’s when the 30-year became non-callable (meaning that the government couldn’t arbitrarily buy it back from holders prior to its 30-year natural expiration).

OPM itself was formally established in 1978. So essentially, we hadn’t “asked why” in any meaningful way post the mid-1980’s introduction of the non-callable 30-year Treasury. And, as a result, we didn’t have any 30-years in our portfolio and thus had only a 7.5-year average duration on the overall portfolio.

What’s done is done but tomorrow is a new day

Partnering with Treasury, we concluded that a longer average duration was appropriate for the CSRDF from a risk and pension liability perspective and would yield a better economic return to the fund. As a result, OPM has increased our duration from 7.5 years to 15 years by adding longer-date treasuries (e.g., 20- and 30-year bonds) to the portfolio.

Who cares? Well, this enabled us to increase the expected investment return of the portfolio to approximately 4.4%.

And why does that matter? Because a higher investment return reduces the amount of money the federal agencies need to contribute annually out of their own budgets, directly saving taxpayer dollars.

We estimate based on current headcount that the higher investment returns on the CSRDF will reduce by $5.5-6 billion annually the amount of money agencies would otherwise have to pay into the fund. On a present value basis over the life of the CSRDF, those savings could approach $100 billion!

It’s not for me to say where those savings go, but every dollar that the agencies do not have to contribute to the CSRDF could ultimately benefit the American taxpayer by lowering the deficit. And it’s pretty clear we could use more of that!

This is the kind of taxpayer-first thinking that the Trump Administration is doing on behalf of the American people. Asking “why?” in pursuit of making sure we are stewards of taxpayer dollars, instead of running the federal government on autopilot.

We are a long way from fully honoring the memory of David Swensen and his many contributions to institutional investing, but today we are at least taking some baby steps – to the tune of $100 billion in potential savings!