Danger ahead
Not since 1929 have the markets looked so precarious
There was a pleasant surprise for me in the post the other day. The bi-annual statement detailing how my pension investments are doing showed another hefty gain, as it has done pretty consistently for some years now and as would be the case for more or less everyone who has their pension pot invested substantially in equities. Stock markets have on the whole been doing well for nearly 10 years now, remarkably so when it comes to the US, which is about to enter the record books with the longest bull run in history. As ever, the question is for how much longer can it last?
I don’t want to dwell unduly on the extended duration of the current economic expansion, suffice it to say that it is already exceptional by historic standards. The law of averages alone suggests that it must soon draw to a close, as likely as not ended by Federal Reserve tightening, to the growing frustration of Donald Trump, who told a Republican fundraiser in the Hamptons at the weekend that he thought he had appointed an advocate of cheap money when he made Jay Powell the Fed chairman, but instead finds an independently minded policymaker of apparently hawkish tendency. Like president Recep Erdogan in Turkey, Trump expects a compliant central bank. As it is, it seems ever more probable that Fed tightening is going to shift the global economy into an at best still growing, but fast slowing, disposition.
Yet there are at least four other factors, somewhat divorced from the usual ups and downs of the macroeconomy, which are also flashing amber, if not outright red – buyback activity, valuation, computer-driven trading, and the related issue of artificial intelligence, or machine learning.
All these things break new ground, in that they point to profound changes both in the structure of stock markets and the way stocks are traded. It scarcely needs saying that where there is rampant financial innovation, there is nearly always trouble in the making.
Start with buybacks, the subject of a fascinating recent analysis by Bernstein Research. Obviously, there is nothing new in companies buying back their own stock, but the scale of the present surge is almost without precedent. For some stock markets, buybacks have indeed become the major source of demand, and it has led to a very curious phenomenon – combined with relatively low levels of stock issuance and IPO activity, it means that the stock of listed equities globally is actually shrinking, and at a rate never before seen.
Nor is this just a Us-driven thing. According to Bernstein, UK and Japanese corporates are also net buyers of stocks. What’s more, the rest of Europe is in net buyback territory for only the second time in 25 years.
Developing Asia, where you would expect to see substantial net issuance, is also on the cusp of shrinkage, for the first time ever.
Bernstein prefers not to draw particularly negative conclusions from all this. “Buybacks are not at record levels as a proportion of market cap,” it points out, while “US buyback activity has also only used a small proportion of the possible benefit from repatriation of cash from overseas”.
Somewhat more questionably, the authors make the assumption that companies act as rational agents. Buyback activity can therefore be seen as a signal to investors that corporate insiders consider current prices to undervalue likely future cash flow.
In other words, if the company is buying, they know something the rest of us don’t.
I wouldn’t be so sure. Virtually every chief executive I have ever come across thinks his stock is undervalued. On the whole, they make poor judges. Buyback activity is in many respects just another symptom of ultra-loose monetary policy, enabling companies to boost their earnings per share by borrowing cheaply to shrink the equity base.
Executive share options and bonus schemes, moreover, give management a big incentive to engage in the practice. Potentially gainful investment of profits in the future has given way to the immediate gratification of what is in effect a share price ramp.
Indeed, far from being a signal of undervaluation, low to negative levels of net issuance are a relatively reliable indicator of market peaks; the last time
‘At the turn of the century, the value of listed stocks was about the same as GDP. Now it is three times as big’
they reached today’s levels was March 2008, just before the worst financial crisis in history.
This brings us to valuation. If you think Robert Shiller’s cyclically adjusted price-earnings ratio – today at a higher level than just before the Great Crash of 1929 – is a misleading guide to valuation, just consider this rather more striking statistic.
At the turn of the century, the value of listed stocks worldwide was about the same as global GDP. Today it is three times as big. Even assuming that the share of corporate profit in GDP continues to grow – and by implication, that labour’s share falls further – this stretches credulity.
To these factors need to be added potentially highly disruptive changes both to the structure of stock markets and the way shares are traded. The aforementioned Powell plans to address some of these later this week at the Jackson Hole Symposium, an annual gathering of the grand mufti of global central banking, in a speech entitled “Changing Market Structure and Implications for Monetary Policy”.
This is expected to focus primarily on whether multinational tech giants such as Amazon and Google are changing the dynamics of the economy.
Answer: they are, and often not in a good way. By gobbling up great swathes of previously diverse economic activity, they are likely to be reducing competition and therefore productivity growth. By the by, they are also a large part of the leap in both stock market value and valuations.
But the Fed chairman might just as instructively be homing in on changes in the way stocks are traded. Computer trading – quantitative, algorithmic, high-velocity, smart beta, exchange traded funds, and so on – is now 50pc to 60pc of daily volume in US equity markets. Quant funds alone are set to surpass the $1 trillion mark this year, double the size of just three years ago.
Often the algorithms behind these funds – effectively black box operations – are designed by IT specialists with little or no understanding of market and economic fundamentals.
The idea is that they reduce risk and guarantee returns, but they said that about collateralised debt obligations. Algorithms of mass destruction, more like.
In any case, the average holding time for US stocks has fallen from eight months at the turn of the century to two months in 2008, and just 20 seconds today, according to data collated by Leo Chen, portfolio manager at Cumberland Advisors.
Ownership is changing hands at speeds measured in tiny fractions of a second. Investment banks and fund management companies have come to think of themselves as more tech companies than conventional financial services businesses.
The advent of artificial intelligence, or machine learning, further distances the process of stock market trading from ordinary human endeavour and control. It is admittedly quite hard to articulate exactly why this construct is bound to fail, but instinctively we know it to be unsafe, as do even many of the so-called rocket scientists at the forefront of this brave new world. The flash crashes of 2010 and 2015 were, I suspect, only harbingers of worse to come. Nobody can tell you when the next big train wreck might be, but that it is coming is not in doubt.