What Would LDI Look Like for Public Pensions?

There’s an old saying that history never repeats itself, but it often rhymes. I’ve frequently found myself searching for the rhymes amongst analogues when investigating applicability of various investment strategies.

When institutional investors think of liability-driven investing (LDI), the most frequent paradigm that comes to mind is allocating an increasing amount of the fund into long-duration bonds.

The objective is relatively simple, although the math can be complicated.

LDI is an attempt to hedge future liabilities by deploying highly funded pension assets into a duration-matched portfolio of bonds, either in the cash bond market or via Treasury swaps or futures, in order to reduce broad swings in the pension net asset value and ultimately accept more contribution certainty in the short run at the expense of reduced investment returns in the long run. By matching the duration of the assets with the duration of the liabilities, the pension funding status is no longer affected by moves in interest rates or, for that matter, in the equity market, either.

For a simplistic example, let’s say a pension has $10 billion in weighted average 20-year liabilities. If it is 100% funded, it can purchase $10 billion of Treasury bonds with a weighted average duration of 20 years and be done. If rates move, the assets and liabilities will be marked up and down together, and such a pension would be considered fully immunized, theoretically.

With an estimated $2.1 trillion globally allocated to LDI strategies (according to Pension & Investments Research Center), the approach has become increasingly popular among corporate pensions. And the recent rise in interest rates has certainly made it a bit more attractive for allocators to back up the truck in bonds. Today, the 10-year Treasury yield hovers around 3.0% versus roughly 2.3% a year ago.

But LDI activity originally surged after the passage of the Pension Protection Act of 2006, which among other things required that corporate pension liabilities be discounted at currently prevailing interest rates, essentially requiring companies to mark-to-market not only their pension investments but also their net funded status.

Moreover, the implementation of Accounting Standards Codification 715 in 2006 pulled pension accounting from the footnotes directly into the financial statements of corporate America. The goal of increased transparency moved the net assets and liabilities onto the balance sheet and changes in the funding status to the income statement in order to paint a clearer picture of the true financial impact of a pension plan for investors.

It should be no surprise that corporations soon began to adopt strategies to mitigate the impact of asset-liability mark-to-market volatility mismatch on their financials. Perhaps instead of LDI,  this strategy could more accurately be called funded status volatility minimization, but I suppose FSVM hardly rolls off the tongue.

And to be fair, that characterization is probably a bit uncharitable to corporate asset owners. According to recent estimates from Milliman, corporate pensions are on average 87.7% funded versus 70.7% funded for publics. And with a market-based average discount rate of 3.95% today as opposed to the average 7.5% rate assumed by public plans, corporates are much better positioned to go long bonds as an immunization strategy.

But does this mean public plans, with our lower funded status and higher discount rate, can’t implement a liability-aware investment strategy?

Of course they can, but the analogue is different.

The pension obligations that a public plan incur are discounted at a fixed rate. That is to say, they bear absolutely no relation to market interest rates. Moreover, the cash flow needs themselves are fairly predictable, using actuarial assumptions. The cash benefit needs for next year are based largely on the current retiree pool, and are known—at least for next month—with near certainty. The further out in time one goes, the less certain the cash needs become as demographics can change, such as payroll growth, increased retirement activity, or decreased mortality rates.

Perhaps for public pensions, a liability-focused asset allocation would be one that invests a substantial portion of the portfolio in moderate-duration, high-income-producing investments in order to meet the next few years’ cash benefit needs with near certainty, and the remainder of the trust into long-term growth assets. 

If, for example, a public plan needed to generate 3% net cash benefit payments for the next few years, 50% of the assets could be invested into a mix of strategies generating a 6% yield, a not impossible objective if one considered direct lending, mezzanine, drug royalties, asset-backed lending, high-yield debt and bank loans.

Such contractual cash flow generative strategies would allow the portfolio to meet the 3% benefit payment needs, reducing short-term volatility concerns or liquidity requirements on the remaining 50% of the trust. These assets could truly be invested for long-term capital appreciation, permitting a less than fully funded plan to still achieve the actuarial required rate of return and provide growth to meet longer future obligations which are not known with certainty.

While the resulting asset allocation may look controversial to many, the simple idea of a laddered income-producing portfolio to meet short-term cash needs and a separate aggressive growth portfolio to maximize long-term wealth shouldn’t be. In fact, it seems to rhyme quite a bit with portfolio construction best practices I learned in retail wealth management. Maybe, it’s time more public pensions gave that analogue a listen.