ALM again

Household accounting; the consolidated balance sheet

Gosse Alserda is on a roll. He is writing nearly as many articles as I am on investing for the WTP, but unlike me he writes them in magazines that are widely read. Whereas I write exclusively for you, dear reader. Kudos to him!

His latest recent comment (€) about the benefits of hedging interest rate risk for young savers makes a case that puzzles me. (Gosse shared the entire article with me and as expected it is very sensible. I wrote this quibbling based only on the title.) To introduce my way of thinking, I start with how I learned portfolio theory.

1. Classical “Modern Portfolio Theory” of Markowitz / Tobin / Sharpe considers how to invest one’s assets for the best results given your risk appetite or risk aversion. Usually—at least in blog posts available for free online—we consider (at least) portfolios of stocks (equities), bonds and cash.

   But … liabilities?

2. In most tellings, the problem is formulated as optimizing for (utility of) wealth. In a pension context, one should target a future income or standard of living rather than a number on a bank balance. That matters because in long-term investing inflation degrades the consumption value of income, and interest rate determines the income yielded by any accumulated capital—and these factors don’t appear in wealth-maximizing utility functions. So, as Merton / Muralidhar / Mantilla-Garcia etc. have shown, the portfolio problem should be expressed relative to a real annuity—even if it is not investible. So pension savers have inflation and interest rate exposure in their implicit liabilities even when they don’t invest in bonds! The portfolio problem becomes one of ALM, or asset-liability management.

   But but … other assets?

3. There is a lot of discussion of ALM going on due to the restructuring of the Dutch pension system. In particular in personalizing this ALM thinking, as individual participants will be the focus of the renewed DC* contracts. But if you do pension-ALM for an indivudual (a.k.a. financial planning) you should not restrict your view to only their pension in isolation. Unlike a pension fund, individuals have a larger balance sheet with assets not committed to providing their pension. In particular, they have their salary and their home (also Merton)¹ and maybe more. Not only is there a lot of value in these other assets, they face sizeable risks that are not considered in portfolio theory above. Some of these risks—like unemployment or house fires—are idiosyncratic and not manageable on the capital market², but salaries and house-prices are related to inflation and valued at prevaling interest rates. Fine-tuning those risks in a (modest) investment portfolio without considering them in a (slightly less modest) total balance sheet³ might give precise but sub-optimal results.

I believe this last point is overlooked by pension-ALM modellers, like the arguments Gosse uses to argues young savers should hedge the interest rate sensitivity (Pensioen Pro, €) of their liabilities. Yes, they should take the interest rate risk of their future pension income into account (2); but then they should also consider the opposite and (possibly) larger, more immediate interest rate risk in the value of their disposable income (3).

Gosse rightfully adds that lifetime consumption is the goal of ALM and this is not immediately affected by the present value of disposable income or home equity. Nonetheless these asset have value when considering personal financial planning, and provide options (working longer, selling and renting) in low interest rate environments where retiring is relatively expensive.

My starting point is the Merton model⁴, that the balance sheet of younger savers is dominated by the fixed-income asset of their human capital, and participating in a pension fund is a way to transform part of that human capital into risky financial capital⁵. How would their balance sheet benefit from investing instead in fixed-income financial capital? Maybe it has a higher interest rate than human capital? It may be because financial capital is tradeable and contributes a return through rebalancing that human capital doesn’t. Or it may be because the specific features of the WTP allow taking more leverage when investing in fixed-income, rather than risky, financial assets.

Other suggestions welcome, to be continued.

1. An asset is any source of income. Earning a paycheck obviously provides income, and owning home pays your rent.

2. They may be correlated! Your employer might fail in a recession, your house might gain value if your region booms. So still don’t ignore them.

3. But but but … these assets are not themselves unencumbered. A house may carry a mortgage, and my salary supports my family and my gaming habit. So the pension is also not a household’s only liability. Use “disposable income” as the asset, and (net) home equity.

It does stop there, the consolidated household balance sheet is the endpoint. There will be no “but but but but”.

4. Although by now we have chalked up quite a list of instruments we need to model: stocks, bonds and cash to invest with, a forward-starting real annuity representing pension liability, and a real annuity representing human capital. (This is roughly the scope of the regulatory scenario set!)

5. As a mortgage is a way to transform some of that human capital into a real estate asset.