J Studios
AI is not the future of trading, it’s the present. In the US, Europe and some Asian stock markets, up to 70% of transactions are now executed through AI-driven algorithmic trading. That has fundamentally changed stock price behavior. Rapid and massive money flows have increased volatility and the magnitude of short-term price movements. Algorithms are binary: either sell or buy until the order is executed, no matter the price or value of the business.
Algorithms react to news. They don’t care for the “vision thing”. What matters is hard facts, usually looking back. A company’s disappointing earnings report triggers immediate sell orders, even if lower profits are due to opportunistic investments or a one-off extraordinary expense. Small cap growth stocks are dumped on the news of higher interest rates, no matter how strong their secular growth prospects. Computers, it seems, are not yet programmed to look forward or to understand short-term pain for long-term gains.
This brave new world is confusing at best, but it also offers opportunities for the patient investor.
MercadoLibre’s (MELI) recent volatility is a case in point. Latin America’s leading eCommerce and fintech company is growing by leaps and bounds. In Q4 of 2023, revenues were up 42% and operating earnings 77%, beating even the most optimistic estimates. However, the stock went down 10% the day of publication because of a one-off tax provision. A long-standing and still-disputed tax liability in Brazil caused a temporary dip in published earnings.
There was a time—not so long ago—that investors would easily have seen through such temporary expenses. But, instead, MELI drifted another 15% lower over the quarter as no fresh news was available. Unsurprisingly, earnings bounced back three months later, triggering a sharp bounce in the stock price.
This is just one example among many. Artificial Intelligence is still in its infancy and seems to lack nuanced thinking. As a result, the market seems to have lost its price discovery function. It is now mostly reacting to backward-looking, unfiltered numbers.
Over time, AI Machines may become more sophisticated but, for the time being, existing mechanisms tend to reinforce themselves. With stocks being excessively punished or rewarded by short-term thinking, many are reluctant to look past the next news cycle. Herein lies the long-term opportunity but short-term volatility.
Algos are programmed to react to news, either micro news during earnings season or macro events like interest rate moves. If interest rates go up, computers have learned to sell growth stocks across the board. If interest rates are anticipated to go down, the reverse happens instantaneously. Algorithms are directional, up or down, regardless of valuations.
The price discovery mechanism of the market is further hampered by index ETFs. Large cap stocks disproportionately benefit from the flow into index ETFs. The larger the weight in the index, the more money is allocated to it. That money pushes the stock higher, which increases its allocation in the ETF even further, thus creating a not-so-virtuous circle.
Even when stock pickers refrain from buying these large cap stocks when richly valued, ETF money continues to blindly support the stock disproportionately. The buying is automatic and goes a long way to explain the concentration of performance in a few select stocks. There was the Magnificent Seven at first, but today, all the money seems to be concentrated in one stock: Nvidia. One thing is for sure, small caps have been left out.
History teaches us that the music eventually stops and traders are then reminded that a stock is a share in a company, not an instrument of some computer game. Remember the computer “insurance policies” that led to the 1987 crash? Computers were programmed to sell whenever a correction reached a certain percentage. It was a smart way to limit the downside of one’s investment. Unfortunately, it ended up doing the reverse when too many people came on board. Then it fed on itself and provoked the first crash of the digital era. The Long Term Capital debacle is just another example of a computer-driven strategy that ended being the victim of its own success.
When the dust settles, fundamental investment will once again prevail. AI driven money flows notwithstanding, companies with secular growth left behind get cheaper with time. The more earnings grow, the cheaper they get.
