Key insights

• Bessent and Hassett focus on the tech investment surge of the mid-90s but leave out the macro conditions that enabled it: cheap Chinese imports holding down inflation and a government cutting its own borrowing to free up money for the private sector.
• Globalization was the invisible engine of the 90s boom. Today's tariffs and trade war create the opposite dynamic, potentially making an AI productivity boost inflationary rather than disinflationary.
• Technological revolutions don't happen in a vacuum. The macro context (trade policy, fiscal policy, interest rates) matters as much as the technology itself.

In Brief

Chris Anstey, a senior editor at Bloomberg News, appeared on Bloomberg This Weekend to push back on a comparison being made by the Trump administration's economic team. Scott Bessent, the U.S. Treasury Secretary (the top official in charge of the country's finances and tax policy), and Kevin Hassett, Director of the National Economic Council (the White House body that coordinates economic policy), have been arguing that trillions of dollars in AI investment could supercharge US productivity the same way the internet boom did in the 1990s. Anstey's response: the parallel only holds if you ignore the most important part. The 90s boom was not just about technology. It was also powered by globalization, cheap Chinese imports, and a government that was cutting its spending, not expanding it. None of those conditions exist today.

---

The 90s parallel the White House is selling

Bessent and Hassett's argument goes like this: in the mid-1990s, a wave of investment in internet and computing technology boosted worker productivity. That let the US economy grow faster without sparking inflation. Alan Greenspan, then chair of the Federal Reserve (the US central bank, which sets interest rates), recognized what was happening. He was "very prescient," Anstey says, and kept interest rates low for longer than most economists expected, allowing the productivity boom to run. The conclusion drawn by today's team: AI investment could do the same thing now.

Anstey does not dispute the technology part of that story. Greenspan was right. The internet did have a profound effect on productivity. Economic growth accelerated sharply compared to the previous decade. But he argues that focusing only on the technology part and ignoring everything else gives a misleading picture.

---

What they're leaving out

Globalization: the invisible engine

The biggest missing ingredient is globalization, specifically the early years of incorporating China into global supply chains. The flood of cheap Chinese goods into the US kept consumer prices down even as the economy was running hot. That created disinflation (prices were still going up, just more slowly than before). Disinflation gave Greenspan room to keep rates low. Low rates meant cheap borrowing for businesses investing in technology.

Anstey has a personal story that captures how central this factor was. When he was the incoming US economy team leader at Bloomberg, Greenspan came to speak at the bureau. It was a meeting arranged by the then bureau chief, Al Hunt, a longtime contact of Greenspan's. Eager to impress his new colleagues, Anstey blurted out that, looking back over the previous decade, Greenspan's most important metric for disinflation had been the declining price of Chinese goods imported into the United States. Greenspan glanced over, scowled, and asked: "Do you want to continue with the presentation?" Anstey apologized immediately. But the incident stuck in his mind as proof of how significant that factor was. And, he notes, it is completely absent from how the Trump team talks about the 1990s experience.

Tariffs run in the opposite direction

Today's trade environment is not just different from the 1990s: it is the opposite. Where the 90s brought lower import prices and less inflation pressure, today's tariff policy is pushing import prices up. If AI investment does boost productivity, those gains could be partly or fully offset by higher prices caused by tariffs. That changes the whole calculation for whether the Federal Reserve can afford to keep rates low.

Fiscal policy: the other missing ingredient

The 1990s also saw fiscal contraction, meaning the US government was cutting its borrowing and spending less. When the government borrows less, it leaves more room in financial markets for private companies to borrow cheaply. Businesses investing in new technology could access low-cost loans. Today, the US is running large government deficits (spending significantly more than it takes in). That can push up interest rates for everyone else, making it more expensive for companies to borrow and invest in AI.

---

Why this matters beyond the debate

The White House framing is not just an academic argument about history. It shapes real policy decisions. If policymakers believe that AI investment will automatically produce a 1990s-style boom, they may underestimate inflation risks and resist raising interest rates when needed. They may also overestimate how much the labor market will benefit, since the 90s boom was accompanied by a job market that was adding millions of new positions, not one where AI might replace workers.

Anstey's point is not that AI won't improve productivity. It may well do that. His point is simpler: technological revolutions don't happen in a vacuum. The macro conditions around the technology matter as much as the technology itself. In the 1990s, those conditions lined up unusually well. Right now, several of them are pointing in the wrong direction.

---

Glossary

---

Sources and resources