By Mohamed El-Erian*
Recent political and asset-market developments have given investors hope that we may be reverting to a more familiar and comforting landscape in which stock-market gains are no longer dependent on a single sector, and diversification into United States Treasury bonds provides reassuring risk mitigation. But this judgment seems premature. Rather than assuming that traditional asset-allocation guidelines apply again to current conditions, investors should position themselves for more volatility, significant variations in returns across asset classes and sectors, and opportunities stemming from thematic market moves.
The conventional wisdom on portfolio diversification has been fundamentally challenged by two major developments: the dramatic dominance of tech stocks in delivering equity market performance, and the inadequacy of US Treasuries as risk mitigators. Although these are distinct dynamics, the rapid development of AI systems has played an important role in both.
The promise of huge payoffs from AI has attracted massive capital allocations, and for good reason. As Google’s James Manyika reminds us, the technology’s potential extends far beyond mere productivity enhancements. In addition to being the next general-purpose technology (like the steam engine or electrification), it is also an “invention for inventing,” owing to its ability to unlock new scientific discoveries. We are entering a period with parallels not just to the Industrial Revolution, but also to the Enlightenment.
The United States is well positioned—uniquely so, for the moment—because its capital markets have proven both willing and able to finance the AI revolution. The problem, however, is that no one has made room for a new wave of massive bond issuance. With the US federal government still relying on debt markets to finance its own immense funding needs, borrowing costs have increased across the board, pushing down the price of US Treasuries and eroding their traditional risk-mitigating role in diversified portfolios.
Then came the US midterm elections earlier this month. Worries that tech stock prices had sprinted too far, too soon, were compounded by concerns that the Democrats’ sweeping victory would fundamentally alter the industry’s operating environment. The market priced in fears that the Trump administration’s “let it rip” approach would be challenged by the new Congress and state governors who have promised to throw sand in the gears of AI infrastructure construction. Still, as a silver lining, the same concerns took some pressure off borrowing costs, allowing Treasuries to act as a temporary buffer against falling tech stocks.
No wonder market analysts have shifted their recommendations. Having previously urged investors to go in big on tech stocks—“let the trend be your friend”—many are now advocating a “rotation trade”: moving capital out of tech and into everything else. Moreover, this new push for greater equity diversification has been accompanied by renewed confidence in the traditional benefits of stock-bond diversification.
But, again, investors should not presume that we are returning to the old normal. Too much of the market remains heavily levered to AI, both in finance and across the broader real economy. Confidence in AI’s promise to drive higher productivity has consistently shielded markets and the economy from the impact of adverse shocks—like geopolitically driven oil-price spikes. But if that confidence starts to dissipate, the economic and financial implications could be profound.
Since the launch of ChatGPT in 2022, the AI trade has powered the economy through massive capital expenditures that have in turn driven employment. It has also indirectly spurred growth via higher consumption among better-off segments of the population, through what economists call the “wealth effect” (the increased inclination to spend as a result of feeling richer on paper). Both these forces have compensated for the increasing risks that come with a K-shaped economy, where the fortunate grow richer while the less fortunate tread water.
Without the AI engine that has been powering growth in recent years, the outlook for both the US and the global economy would deteriorate, creating headwinds for the overall stock market, and this less favorable environment would also bring higher risks of financial instability. Remember, the AI sector has distinguished itself through widespread “circular financing” arrangements, where one company lends its customer money to buy its product. Such sectoral interdependencies increase the risk of contagion, where localized challenges at one firm suddenly create system-wide problems.
Worse, the challenge facing the bond market now extends well beyond the need to make room for AI-related bond issuances. Sticky inflation remains a concern for many investors. Having consistently missed its inflation target (on a monthly basis) for more than five years, the US Federal Reserve will show no patience before it hikes interest rates again. Then there is the potentially more prolonged, structural challenge of what seems to be a diminishing appetite for US debt among previously reliable buyers and holders of Treasuries, including both adversaries and allies in Asia, Europe, and the Middle East.
Investors therefore should think twice before concluding that recent market developments constitute a durable return to historical norms. This is still an age of structural uncertainty. The architecture of both the global economy and markets is in transition, owing both to top-down geoeconomic pressures and bottom-up technological forces. Such circumstances serve as a caution to investors keen to revert to the portfolio strategies of the past.
Looking ahead, successful portfolio management will require focusing on balance sheets and on companies with pricing power, using more cash (and fewer bonds) for risk mitigation, and embracing a more thematic, granular approach to asset allocation. It also means looking for opportunities where investors are paid to complete markets, such as in private capital in emerging markets.
*Mohamed A. El-Erian, President of Queens’ College at the University of Cambridge, is a professor at the Wharton School of the University of Pennsylvania and the author of The Only Game in Town: Central Banks, Instability, and Avoiding the Next Collapse (Random House, 2016) and a co-author (with Gordon Brown, Michael Spence, and Reid Lidow) of Permacrisis: A Plan to Fix a Fractured World (Simon & Schuster, 2023). Copyright: Project Syndicate, 2026, published here with permission.
We welcome your comments below. If you are not already registered, please register to comment
Remember we welcome robust, respectful and insightful debate. We don't welcome abusive or defamatory comments and will de-register those repeatedly making such comments. Our current comment policy is here.