Recent Posts

Sunday, September 30, 2018

Liability Matching Versus Return Maximisation And Bonds

Before we discuss investment strategies, we first need to decide what our ultimate objectives are. When we are investing, the two usual objectives is that we are either attempting to achieve maximum returns, or we are matching against liabilities, using a broad sense of liabilities. A boring non-levered bond portfolio is a reasonable choice for liability matching, but it is much harder to see how it fits within a pure returns maximisation view.

Background on this Article

I have sent my book on breakeven inflation analysis to be edited, and so I am now starting to think about upcoming writing projects. (The firm that handled my paperback printing and distribution has been merged with the ebook distributor, so I am currently unsure how long the new publication workflow will take.) My plan was to write a text on business cycle analysis, but I am also thinking that a book on the basic principles of how bonds fit into portfolio management would would be a good project. I prefer to concentrate on the business cycle book, but will work on ideas on the bond portfolio project at the same time.

(In case any readers work for a firm that produces bond indices, I would be interested in getting my hands on data and permission to refer to particular bond indices in the book.)

Return Maximisation versus Liability Matching

We can view the objective of a portfolio in two ways:
  1. Achieve the highest possible return at any given time.
  2. Look at market pricing and see what returns are implied by aggregate valuation. Assume that our returns will be relatively close to those implied returns, and so contribute to the portfolio based on our projected cash flow needs and those returns.
I will discuss these in turn.

Return Maximisation

For someone outside investment management, particularly those who come from a physical sciences background, maximising returns seems like a no-brainer. In fact, "maximising returns" or some variant will be a standard piece of boilerplate text in most investment pitches.

We need to look at this concept more closely. If we wanted to be extremely literal, returns maximisation would involve implementing the following simple (hah!) procedure:
  1. Forecast returns for all assets over the next month.
  2. Put 100% of our portfolio into the asset with the highest expected return. (And then apply as much leverage as possible...)
 If one were able to do step #1 with a decent level of accuracy, one would end up with phenomenal levels of returns -- probably at least 50% per year, year after year. Starting with $10,000 investment, someone could end up purchasing all financial assets for sale in the United States after a couple of decades.

Needless to say, there are not a whole lot of people starting with $10,000 and purchasing all the financial assets in the United States. So we need to temper our expectations slightly.

In the current low inflation environment, the usual story is to aim for a 20% return, which is the hurdle rate pretty much everyone in the private sector uses for returns on equity (no matter where the central bank sets interest rates). This 20% return (or more) can be achieved "legitimately" in a number of way:
  • stock picking;
  • options strategies;
  • hedge funds (or leverage);
  • emerging markets.
Investment manias of course promise such high levels of returns, such as condo investments. One may note that people pushing gold or crypo-currencies typically dangle extremely high returns ("gold is going to $10,000!") in their pitches, even though the story is supposed to be that these investments are a stable store of value.

Liability Matching

This approach is based on accepting that we cannot impose our desires for future returns, we have to accept what is implied by valuations. Since we lower our return expectations, we need to ask ourselves: how large a portfolio do we need to meet objectives? We need to determine what future cash flows the portfolio needs to provide, and see what contribution level is needed to those cash flows. In most cases, the cash outflows are determined by things like desired retirement income, and so we refer to them as actuarial liabilities. However, for levered financial institutions, the outflows are determined by contractual obligations, and so they are liabilities in the usual sense.

Coming up with expected returns is a task that ranges from the trivial to extremely complex. For government bond portfolios, we can often guarantee a return for relatively long investment horizons, although things are obviously more complex when the investment horizon is much longer than the maturity of the bonds we invest in. Risk asset returns, particularly equities, are obviously much harder to forecast.

This mode of thinking is standard in investment management, and would be considered so obvious that I will not spell out the details. My point here is that although this mode of thinking is unconsciously assumed by most professionals, it does not align with the thinking of the broad public, who will have more of tendency to assume that "returns maximisation" is how finance operates.

Of course, professional investors do attempt to "maximise returns," but they are working within risk limits relative to a benchmark. That is, their portfolio returns are supposed to be somewhat above market averages, but our expectations should end up being close to where broad valuations suggest.

Whither Bonds?

Bonds, particularly government bonds, do not fit into a view where we expect extremely high returns from a portfolio -- unless we apply leverage via derivatives or borrowing. Instead, they are only of interest if we are interested in meeting future cash flows.

Even so, the very low level of risk-free yields in the developed countries (courtesy of our New Keynesian friends), makes bonds a difficult asset class to work with. Government bonds returns that are guaranteed to be much lower than even fairly pessimistic equity return forecasts over long investment horizons (although one can always make a case for a near-run bear market). Why drag down expected portfolio performance by including bonds in the mix?

The usual answers are not what most people would expect.

  1. Government bonds provide diversification as that their prices tend to rise during financial crises, when risk assets all tumble across the board. Although not all recent crises were of the same magnitude as the 2008 Financial Crisis, all the post-1990 recessions were associated with financial market stresses of one sort or another. This is for reasons pointed out by Hyman Minsky decades ago.
  2. Other asset classes generally cannot provide much in the way of near-run guaranteed cash flows. This matters for things like pension portfolios that are in run-off mode: just paying out cash to beneficiaries.
The second factor is obviously situational; only the first provides a general principle. However, if the objective of the bond portfolio is to provide risk diversification, it casts a shadow on one extremely common practice: reaching for yield by taking credit risk.

Concluding Remarks

I have kept this article short as it is not providing revolutionary insights into portfolio management. However, it is aimed more at an audience that has not really thought about this distinction; the "maximising return" viewpoint seems so self-evident it is a surprise that it is not how the bulk of the financial profession is organised.

(c) Brian Romanchuk 2018

1 comment:

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.