Monday, May 14, 2018

Upside Down CAPM, Part 4: National Debt

The basic premise of the Upside Down CAPM (comment if you have a better name for the concept) is that there is a pretty stable expected real rate of return on at-risk assets.  This is about 7%-8%.  The problem with at-risk investments isn't that the expected rate of return changes much over time.  It's that realized returns over short time frames are highly volatile.  (This is why NGDP level targeting would be so beneficial.  Those short term fluctuations are waste.)

My hypothesis is that the waste isn't expressed much in the rate of return on at-risk assets.  Held over long periods of time, real returns are somewhat stable.  There are some persistent real shocks that change long term returns, so long term returns aren't completely stable, but variance on total returns to equities starts high and then declines regularly as the holding period increases.

Any asset manager knows this, and so savers with longer holding periods are more likely to hold equity positions.

In the context of the Upside Down CAPM, my point is simply that the basket of total assets, accounting for some maintenance reinvestment, is basically a perpetuity.  Diversified equities, the proxy for that basket of assets, are a perpetuity.  This long-term stability and mean reversion means that expected returns of equities tends to remain around 7%-8%, and long-term investments reflect this long-term stability.  The sharp changes in valuation are mostly due to short term shocks, real and nominal, and the fickle nature of the valuation of a perpetuity with minor shifts in cash flow and long term growth expectations.  The portion of ownership we call debt is simply a subset of this basket where some capital, seeking shorter durations and more cash flow certainty, pays a fee for that certainty.

So, right now, highly rated corporate bonds pay a little less than 4%, or about 2% in real terms.  The Upside Down CAPM approach says the proper way to think about this is that some subset of the capital base is paying a 5% annual fee in order to receive 2% annually rather than receiving 7% annually with short term fluctuations.  Debt supply is mostly a transaction where savers are paying a convenience fee for keeping their capital safe for future consumption.  It just happens that in developed economies, the base return on capital - 7%-8% real - is high enough that the fee paid for certainty still leaves a residual return that is greater than zero....most of the time.  As we have seen recently, this doesn't have to be true.  Zero is not particularly important, especially in real terms.

Equity holders currently expect about 9% returns (roughly 2% inflationary capital gains + 2% dividends + 3% buybacks + 2% real growth).  The 7% + inflation figure is pretty stable.  There are some small shifts between growth expectations and capital payouts over time.  So, for instance, in the late 1990s, growth expectations were high.  Since the real return of 7-8% doesn't change much, that mostly meant that dividends and buybacks were lower, and PE ratios were higher.  Since those growth expectations are negatively correlated with risk aversion, the fee lenders required for capital protection at the time was very low - maybe 2%, so that bonds paid close to 5% real returns after the discount.

If we think of national public debt through this framework, I think this argues for (1) less concern about the level of debt outstanding and (2) a higher standard of returns on public investments.  As a first conceptual step, I don't think there is much difference between funding public spending with taxation or borrowing.  Either way, $x in capital is removed from the private stock of capital.  If spending is funded with borrowing, that is just a separate transaction where the government is providing the service of capital protection.  So, deficit spending is really a combination of two transactions.  First, taxation and spending.  Then, a transaction where the government accepts cash and promises to protect it, with some interest, and pay it back in the future.

Thinking of public debt in this way, I think the typical understanding of deficit financed spending being stimulative is overstated.  The spending part of the transaction is the same either way.  If there is anything stimulative about the deficit financing, it is just that the government has a competitive advantage in providing capital protection services.  Comparing treasury yields to investment grade corporate bonds, this amounts to a little more than 0.5% on average.  The reason deficit spending is useful in a contraction is if that spread rises, then that is an indication that the public service of providing capital protection is especially valuable.

So, it is stimulative.  But, only to the extent that it provides this service.  It is better to think of this stimulus in terms of the spread between AAA-rated private securities and treasuries than to think of it in terms of the government borrowing cheaply to inject spending into the economy.  There is no injection of capital here.  It is just a transfer between a saver and a lender.  And, if spreads are relatively low, then it is a service that can nearly as easily be provided in the private sector.  For a few months in late 2008 and early 2009, that spread was more than 3%.  The spread could have been brought down by more accommodative monetary policy that would stabilize nominal activity.  Eventually it did.  But, lacking that, massive public debt expansion was valuable then, including actions like guaranteeing GSE MBSs to reduce spreads.

As long as public debt levels are low enough that they don't induce a credit spread, then the public benefits from having the government provide this service.  Debt outstanding is just a measure of the extent to which there is demand for that service and the government is meeting it with a supply of capital protection.

But, this argues for a high required return for public investments.  In private markets, at-risk investments are expected to return 7%-8%.  The spread between private bonds and public bonds is only about 0.5%.  So, public spending has a small advantage over private spending because of the government's ability to provide this service.  But, in order for public spending to be more valuable than private spending, it still needs to return something like 6.5%-7.5% or more to justify taking that capital out of private markets, even with the public advantage in providing capital preservation services.

This also means that high real interest rates are probably not something that we need to fear in the context of the national budget.  High real long term interest rates will only happen if risk appetites and growth expectations rise.  As in the late 1990s, this would be associated with rising growth, innovative investments, and rising federal revenues.  In the late 1990s, the relative weights of these factors was so favorable, that federal officials feared a shortage of treasury securities might develop among institutions that utilize them.


  1. Well, I read this post a few times, and I think I agree with it.

    In a way, this post argues for long-term money managers who borrow heavily (leverage up) at 5% and invest long-term for 7% equity yields. If they do not have to mark-portfolio-to-market in down years, then the idea should work.

    Interestingly enough, Oakwood Capital (Howard Marks) just issued preferred units paying 6.625%. I guess they think they can invest the money and make more than that.

    The game of arbitraging between short and long-term yields is an old one, obviously.

    You say the macroeconomic difference between federal borrowing or taxing is not much.

    What about QE? Do have a point of view?

    What about money-financed tax cuts (helicopter drops)?

  2. Check out the chart in this post. There are more 15 year periods where 5 year Treasuries outperform stocks than I think you'd expect. It was more than I expected, for sure. Related topic, while the variance of annual percentage returns falls as time periods lengthen, the dollar variance continues to rise. That is, for example, if the range of historic returns over 20 years narrows to 4% to 10% (that's a guess, so I could be way wrong), the ending balance at the 10% returns is three times as much as the 4% portfolio, so the future retiree with the lower balance is getting his perpetuity on a much lower number.

  3. I forgot the link! Sorry.

    1. Thanks bill. Interesting article. And, there is something to be said for managing for the human element of portfolio allocation and for looking ahead to potential problems, like the tendency to want to sell off when losses come (which is not entirely irrational).

      One issue with bonds is that they are a mixture of (1) a real hedge and (2) a speculative position on changing inflation. I think we would be operating on much more fruitful ground if the public norm was to treat real bonds as the pure asset class, and nominal bonds as a speculative subclass. Hopefully that will become more common over time.

    2. I had never thought of it that way, but that's a very good insight (real and nominal bonds).