Via the excellent Damein Boey at Credit Suisse:
Another day, another set of US President Trump and China headlines. Overnight, we heard reports from Chinese officials that US officials were prepared to rollback existing tariffs as each phase of trade negotiations progressed. We later heard US officials confirm these reports from their perspective, although there has been some suggestion of strong internal opposition to rollbacks within the US delegation. In response to the constructive news, bonds sold off sharply, the yield curve steepened and equities rallied. Within equities, value factors outperformed, while momentum and defensive factors underperformed. Essentially, the rotation story from the past few months has continued.
Our thesis is that bond yields will continue rising, inflicting pain on passive and risk parity investors, who are very overweight bonds and related exposures. For a short time, value equities should continue to do well as a diversifying asset class, because their earnings should receive a boost from improving global growth momentum, they are relatively cheap (by definition), and most importantly, because their performance has not been overly correlated with other major asset classes (given trend underperformance in recent years). However, should passive and risk parity styles start to really underperform, we would ultimately expect market de-leveraging forces to move us into a broader risk-off phase, even in factor land. For further discussion of our investment thesis, please see our recent articles Tail risk or apocalypse” dated 5 September 2009 and “Uncorrelate me” dated 30 August 2019.
In this “in between” phase, much of the discussion in investor circles is likely going to revolve around valuations, and whether we can find reasons to justify extremely stretched equity market valuations with shifting fundamentals. In this update, we highlight a few of these.
For context, at current levels of non-financial market cap-to-GDP for the US market, the S&P 500 is priced for 10-year annualized returns of less than -0.5%. Valuations and future returns are inversely correlated on long-term horizons, because it is always the same uncertain future we are buying (with wide error bands) – it is just a question of how much we choose to pay for that future. A higher (lower) valuation starting point should be consistent with lower (higher) future returns. But the key thing to note about the relationship between valuation and future returns is the power of mean reversion. Depending on how we construct our market valuation measure, we will get stronger or weaker mean reversion at different investment horizons. Our preferred metric relies on the power of two mean-reverting series rather than one – we assume that prices revert back to earnings in the long-term, while profit margins mean-revert back to long-term average. Therefore a price-to-revenue metric should have better (faster) mean-reversion properties than a standard price-to-earnings ratio. Intuitively, the more “through-the-cycle” we can get with our valuation, the better we will be able to look through short-term volatility, and focus on the longer-term picture.
In order to deny the power of mean reversion within our framework, one has to believe that:
- Interest rates will not rise too much, too quickly, supporting multiples.
- Profit margins will stay elevated for longer, rather than compress.
- Long-term productivity growth will hold up, despite some very nasty-looking demographics.
Let us leave aside the rates question for now, because from a stock selection perspective, rising rates are very clearly good for value equities, and very clearly bad for quality equities. We are more interested in the latter two prospects, because they could potentially impact our currently positive view on value and cyclicals.
On profit margins, the so-called Kalecki equation is very helpful. The Kalecki equation is a national accounting identity. It merely states that the sum of saving across all sectors has to equal zero. For every lender, there must be a borrower. Re-arranging the identity, we can establish that corporate saving must equal the sum of the trade balance and fiscal deficits, less the household saving rate. As it turns out, in the US, we can say something causal about this relationship too – the sum of fiscal and household balances has historically been an inverse leading indicator of corporate profits, a sub-component of corporate saving. When expressed in terms of GDP shares, we can say that corporate profit margins lag behind the inverse of government and household saving rates. In order to believe in “stronger- for-longer” corporate profit margins, rather than mean-reversion to lower levels, we have to believe that fiscal deficits will remain large, and that households will keep dis-saving. Can we rule out these possibilities near-term? Probably not. The Fed’s Senior Loan Officers’ Survey reveals that household loan demand is very strong, consistent with dis-saving. And fiscal deficit spending in the US could continue at its currently brisk pace of 6% of GDP per annum, and even accelerate, if policy makers become serious about infrastructure investment. Austerity ideologues are always a problem of course, particularly heading into an election year. But our point is that it is possible to see a scenario where profit margins stay high for longer, allowing valuations to stay stretched for longer.
On productivity growth, we note that in the long-run, real earnings per share growth for S&P 500 companies has averaged 2% per annum through time – roughly the same rate of growth as productivity, whether measured as real GDP per capita, or real GDP per hour worked. We have demonstrated in previous work, that trend productivity growth is largely a function of demographics.
The more rapidly the population ages, the lower the ability for society to create and absorb new technologies in a shrinking working-age population environment, and the lower productivity growth becomes. Based on demographics alone, productivity should be shrinking right now, and over the next decade. If the forecast materializes, we would expect real earnings growth to slow significantly, meaning that it will not be safe to assume that stock market investors are buying the same uncertain future as they have in the past, because of a trend shift down in the profile of earnings. Equities would appear even more expensive than they currently do on a “growth at a reasonable price” basis. What might be done or said to stave off this fate? We can think of a few possibilities:
- Increased immigration, especially in younger working age segments, could help to lower the weighted average age of the population for a short period of time. But the sugar hit would be short-lived, and in any case, Trump’s current policies do not favour a more rapid pace of immigration.
- Revise up retirement age, so as to keep older, more experienced people in the work force for longer. Perhaps older cohorts might choose to work for longer anyhow, given the bleak outlook for long-term investment returns …
- More public investment in capital stock (for example, again, infrastructure).
- To make the US a net exporter again, bringing back high “gross value added” manufacturing onshore. But of course, this would require a step up in trade wars, and potentially damage the financial system through a USD liquidity drain.
- Hope for a technology revolution, with the detrimental effects of “fallacy of composition” – the perverse under-weighting of growth among start-ups in winner-takes all markets, because their growth occurs at the expensive of larger-cap incumbents.
Interestingly, much of the spread between actual productivity, and the level implied by demographics alone, can be explained by the sum of trade and fiscal balances … But the gap is becoming too large to explain, even with leads and lags, or prospective changes in policy settings. Therefore, it is critical that interest rates do not rise much, or even fall in the very long term to preserve equity valuations on a “growth at a reasonable price” basis.
In conclusion, we will search for many reasons to justify grossly-stretched through-the-cycle valuations. We may even convince ourselves that some of these reasons are valid – but the question is for how long, especially when these reasons are dubious.