“If prices are obviously wrong, then you should be rich.” With this unflinching retort, Eugene Fama—father of the efficient market hypothesis (EMH)—has long dismissed critics who see stock markets as irrational, manipulated or wildly inefficient. Yet these days, some of his former allies are beginning to waver. The price of Nvidia, a chipmaker that has become a poster child for artificial intelligence, doubled in five months last year—before falling by $1trn, and recovering much of that again, in less time than it takes to finish a macroeconomics course. Other AI darlings have seen similarly vertiginous rises, often with little discernible news behind them. Meme stocks like GameStop and AMC, whipped into a froth by Reddit threads and TikTok influencers, appear to move on vibes alone.
If markets are efficient, as Fama famously argued, then prices should reflect all available information. But today asset prices appear to reflect less information and more narrative. Retail traders trade on Reddit-fuelled conviction. Venture capitalists claim to spot AI winners years in advance. Central banks set policy using the same macro data that everyone else can see, yet markets still lurch wildly when policymakers speak. Are prices really right?
In a world seemingly driven by hype and herding, EMH should be on its last legs. But paradoxically, the more the market seems divorced from fundamentals, the more useful the hypothesis becomes: not as a description of perfect rationality but as a warning against hubris.
Fama’s idea, first formalised in 1965, was that if millions of investors are all trying to spot mispricings and trade on them, then any information worth trading on gets rapidly baked into prices. A stock might be too expensive or too cheap in hindsight, but it is roughly “right” at the time, given what is publicly known. That insight reshaped finance. It undermined the idea of skillful stockpickers, ushered in passive investing and won Fama a Nobel prize.
The theory has never lacked detractors. Behavioural economists such as Richard Thaler have long argued that investors are prone to cognitive biases, and that prices are distorted by waves of greed and fear. Market crashes, from the dotcom bubble to the global financial crisis, are regularly used as cudgels against EMH. The latest weapon is artificial intelligence—both the market mania surrounding it and the technology itself, which some believe will allow investors to digest information faster than ever and beat the market using brute-force computing.
Yet data continues to mount in Fama’s favour. Over the last decade fewer than 10% of actively managed US equity funds have outperformed the market after fees, according to S&P Global. In Britain, the figure is below 20%. In Europe and Asia, it is a little better. Hedge funds, supposedly home to the sharpest minds and fastest algorithms, have struggled to justify their fees. For every year that active funds underperform, the appeal of cheap, passive index funds grows, now holding more than half of all US equity-fund assets. If markets are truly inefficient, then those exploiting the inefficiencies are being awfully modest about it.
Still, the EMH was never a claim that prices are always correct. Rather it suggests that prices reflect available information, not that they anticipate future truths. This distinction is crucial. The price of Nvidia reflects a collective belief about future AI profits, but no one knows how the market will shake out. Investors may be wrong, but they are not irrational. As Fama has pointed out, the dotcom bubble mispriced many firms but it also underpriced Amazon.
Moreover, the hypothesis is less a claim about outcomes than about process. An efficient market is not a stable one, it is a competitive one. The presence of bubbles, fat tails and sudden corrections is not necessarily evidence against EMH, but a feature of markets in which thousands of people are constantly updating their beliefs and reallocating capital. Prices may move on flimsy news, or no news at all, not because markets are broken, but because all known news is already priced in. A rumour, or even the absence of bad news, can influence expectations.
Indeed some of the apparent inefficiencies in modern markets stem from rational incentives. Retail traders on Reddit might appear ill-informed, but they have limited capital at risk. Institutions meanwhile are increasingly incentivised to ride bubbles rather than fight them. A fund manager who avoids Nvidia on valuation grounds might lose clients long before being proved right. In this context, prices are not wrong so much as politically, socially and institutionally constrained.
The more powerful critique of EMH today is not that it fails to describe markets. It is that its success has changed them. Passive investing, once a niche strategy justified by the EMH, now dominates equity flows. Index-tracking funds do not evaluate company fundamentals, they just buy the market. Some worry this has created feedback loops: more money flows into the biggest stocks, pushing up their prices, attracting more flows. If everyone behaves as if prices are fair, who ensures they actually are?
That fear is likely overstated. Index funds do not set prices; they take them. As long as active traders still exist—and they do—the price discovery mechanism holds. And in practice the dominance of passive funds may enhance efficiency by reducing noise. Speculative investors chasing flavour-of-the-month stocks have always been around. What is new is the sheer scale of passive capital sitting on the sidelines, unaffected by emotion.
Even so, there are limits to how far EMH can be pushed. Fama himself has always stressed that the hypothesis is “just a model”. All models are wrong, as the statistician George Box famously said, but some are useful. The market is not a giant calculator arriving at the “correct” price. It is an ever-shifting consensus view, forged by humans with incentives, biases and constraints. Sometimes it is wildly off the mark. But not in a predictable way.
What makes EMH enduring is not that it is perfectly true, but that the alternative—the belief that one can reliably outsmart the market—is more often false. The hypothesis does not say that markets are fair, or just, or wise. It merely warns that for most investors, they are too hard to beat. As Fama dryly notes, if mispricings are so obvious, where are the fortunes?
The lesson remains unfashionable. AI traders now promise precision-timed entries and exits. Meme-stock gurus offer high-conviction trades based on little more than vibes. But investors would do well to remember that markets are made up of millions of people trying to make money—and that as a result, prices tend to reflect what is knowable.
Believe you can beat the market? Fine. But prove it—and consistently. Otherwise, the market is probably right enough. ■
