On 5th December 1996, Alan Greenspan, then Chair of the Federal Reserve made a televised speech in which he questioned how central bankers should respond when asset values have been “unduly escalated” by “irrational exuberance” – in other words, by the actions of investors high on optimism but short on analytical rigour. Though he later explained that he had chosen his words carefully, investors were quick to perceive the implicit warning, and US equities fell c5% over the following weeks. However, exuberance soon regained its hold over analysis and the losses were swiftly made good – and then some: excitement over nascent internet and mobile telecoms technology drove a doubling of the market over the subsequent three years, with technology heavy indices up nearly 300%.
Seasoned investors know how this story ended. The US stock market peaked in March 2000, though the bubble only burst decisively in September. By October 2002, broad measures of US equity markets had halved while those more exposed to the technology sector had lost close to 80%. Devout followers of the latter would have had to retain their faith for fifteen years to recoup these losses: seventeen if one accounts for inflation.
Is there a moral to this story? Why did Greenspan issue his warning so far ahead of the bubble’s peak and, having done so, would investors have been wise to heed it? More to the point, why do we think this history is relevant now?
We see a number of parallels between the economic and market environment of the late 1990s and that which currently prevails. In fact, we see so many that this note will aim to do no more than summarise the argument, setting the scene for a series of future notes (identified in the footnotes) in which we will delve deeper into the troubling similarities, explore the differences and attempt to plot an appropriate course of action for rational longterm investors today. In our view, this course demands a cautious approach to the ample speculation evident in today’s financial markets. Happily – as evidenced by the outperformance of short-dated bonds over US equities over the first eight months of this year – watchful investors can still earn reasonable returns as they await the restoration of a more rational market environment.
Parallel Lines
The clearest similarity between the dot-com bubble and now is the excitement over a nascent technology. In the late 1990s, the world was enthralled by potential of the world wide web and affordable mobile telephony. Today, the excitement surrounds artificial intelligence. In both cases, investors identified an epochal change and raced to secure a share of the spoils.
Just like market bubbles before it, the technology driven gold rush of the 1990s triggered a mass misallocation of capital. Investment poured into businesses deemed to be at the cutting edge of the technological revolution and equity market valuations rose to levels never seen before or since. As the euphoria faded, investors were hit with the brutal reality: these companies’ assets were not worth as much as had been thought, they generated very little – if any – profits and, consequently, the wildly optimistic valuations ascribed to them came down to earth with a shuddering thump.
Fast forward to today, and the parallels are clear. Replace ‘the internet’ with ‘artificial intelligence’, swap a few company names, and market commentary from the late 1990s could apply almost seamlessly to today’s breathless headlines. Healthy scepticism is a non-negotiable requirement of a competent investment analyst – it is part of the job to hold the fort against those selling pipe dreams. Our analysis leads us to believe artificial intelligence is unlikely to usher in an era of “extreme abundance” (an extraordinary claim made recently in Forbes magazine) any time soon. Yet, as proved by the dotcom bubble, when the market is expensive, it does not take a failure of the new technology to leave investors nursing huge losses. We note that the US equity market is currently trading at 22.4x next year’s earnings – the closest it has ever come to the record 24.5x registered at the very peak of the tech bubble. The scene is set: anything other than world-changing profitability will see such demanding valuations fall in due course.
In another echo of the late 1990s, we believe the rush to invest in the AI revolution has macroeconomic consequences. We are used to thinking that the economic cycle determines the market cycle: a strong economy allows firms to earn strong profits, which are reflected in rising share prices. However, there are reasons once again to believe the natural order may have flipped. By 2000, rising asset values – working in tandem with the fear of missing out – had seen households all but double their exposure to the stock market. Consequently, the US economy became extremely sensitive to stock market performance. When the bubble finally burst, a recession followed. It is therefore troubling to note that household equity allocations are even higher now than they were 25 years ago (Fig.1). Furthermore, drawing yet another parallel, thanks to the US’ accumulation of huge capital inflows – the consequence, at least in part, of persistent trade deficits – the surge in US equity ownership has not been confined to domestic households, creating the conditions in which a decline in the US stock market could have global ramifications.
Fig.1: US Household & Non-Profit Organisations’ Equity Holdings as a Percentage of Total & Financial Assets
All these similarities feed one further historical echo, albeit one that is more difficult to precisely pin down. It relates to investor psychology and behaviour – the ‘irrational’ part of Greenspan’s warning. In his history of the 1929 US stock market crash, the economist John Kenneth Galbraith highlights the role that fraud plays in both fuelling a bubble and subsequently bursting it. He introduces readers to ‘the bezzle’ – the accumulation of ill-gotten gains built up over a period of unscrupulous behaviour “when the embezzler has his gain and the man who has been embezzled, oddly enough, feels no loss”. Distressingly, we see many aspects of the current market which match this description all too well . Such is the scale of today’s bezzle that it seems entirely possible to us that, when today’s victims finally realise their money has been lost, the fall out could rival – or even surpass – that suffered in the wake of the dot-com bubble.
This Time is Different
Those of a more optimistic persuasion will doubtless cast aside Sir John Templeton’s warning about “the four most dangerous words in investing” and claim that ‘this time is different’. Certainly, we have met with a number of strategists who identify the profitability and balance sheets of today’s AI juggernauts as a key contrast to the lack of identifiable revenues associated with some of the dot-com bubble’s headline acts. Today’s technology leaders can afford to invest hundreds of billions in pursuit of the next new thing without racking up huge debts. Equally, investors may point to claims that economic impact of artificial intelligence will dwarf anything that has yet been seen, thereby justifying elevated valuations for the companies involved in its development. Perhaps. We see scant evidence thus far, though we work hard to keep an open mind.
Less open to debate is the stark difference in the fiscal backdrop. By the end of the 1990s, the US had turned its budget deficit into a surplus and was even questioning whether the US government might ever need to issue Treasury debt again. Twenty-five years later, the contrast could scarcely be starker. US federal debt as a proportion of GDP is at levels previously only reached during times of war. Furthermore, bi-partisan projections suggest debt is only going to increase further from here. That the staggering amount of debt has been accrued during a period of generally strong economic growth is historically anomalous, and deeply troubling. It raises many questions : to what extent have tax cuts and government spending supported corporate profitability in recent years, and what happens if they are no longer sustainable? Will international investors remain willing and able to purchase ever more US government debt? And what if the US does tip into recession – will the Federal government be able to finance a recovery package?
Implications for Investors
In our view, the risks currently embedded in the US equity market are substantial. Though we are all too aware of Keynes’ warning that “markets can remain irrational longer than you can remain solvent”, this is not an argument for ignoring warning lights that are flashing ever more insistently. We suspect that we are closer to 1999 than we are to 1996 . Economic momentum is fading, valuations are extreme and the speculative nature of many of today’s most popular ‘investments’ is ever more apparent.
Thankfully, there is no need for investors to risk their solvency: short-dated investment grade corporate bonds offer attractive yields of 5-6% with, we believe, limited risk. Our model portfolios are therefore tilted away from speculative equity markets in favour of the relatively secure – if less exhilarating – prospective returns offered by high quality bonds. Those who read the financial press might not believe it, but each of our recommended short-dated bond funds outperformed the US equity market (in sterling terms) over the first eight months of this year. We are being paid to wait for rationality to return to markets. Though we endeavour to keep an open and honest mind, we are ever more convinced that investors today will be well-served by eschewing speculation, instead adopting a thoughtfully cautious approach to today’s equity market euphoria.
