42nd Annual NABE Economic Policy Conference

President Goolsbee's Remarks

First, it’s great to be back at the NABE meetings and it’s great to see you all again. Thanks to Ellen Zentner for moderating. With the storms and the travel problems, Ellen never wants to get involved with me. The last time Ellen moderated a NABE event where I was involved was on Zoom in 2020 during Covid, before I was ever at the Fed. Our panel was at lunch time. My driver’s license was expiring so I had to go to the DMV, and I got there the minute it opened. But I ended up stuck in a Covid-procedures line for three and a half hours where you weren’t allowed to use a phone or leave the line. I missed the first half of the panel and ended up having to do my presentation from my car in the DMV parking lot. Sorry, Ellen. This time it wasn’t my fault.

I. Introduction

Taking stock of 2025 overall, I think growth was solid and the labor market mostly held steady despite low headline job numbers; however, we stopped making progress on inflation. Some observers have interpreted the same data to mean the economy is teetering and may fall apart without more rapid action by the Federal Open Market Committee (FOMC). I haven’t found that persuasive so far. So today I wanted to make four points about the economy and the data and what they might mean for monetary policy.

Before I do, I need the usual disclaimer that these are my own views and not those of my colleagues on the FOMC or in the Federal Reserve System.

II. Consumer Spending

For all the talk about artificial intelligence (AI) data centers, I believe the main driver of growth in the economy in the last year was solid, broad-based consumer spending.

During the second half of 2025, real consumption expenditures grew at nearly a 3% annualized rate. January looked maybe a little softer, but some of that may have been the brutal weather. Fundamentally, though, a solid consumer remains our key strength going into 2026.

And I think it is broad-based. Remember, it’s hard for consumer spending to be as strong as we’ve seen without it being broad-based. The U.S. Bureau of Labor Statistics (BLS) distributional accounts data from 2000 through 2023 indicate that the bottom 90% of the income distribution does 75% of the spending in the economy. That share barely changes over time. The New York Fed recently published some new, more updated economic heterogeneity statistics, and they show, similarly, that over the past three years, the cumulative real growth in spending has been pretty similar across income groups, with differences of only around 1 percentage point between high- and low-income consumers.¹

Certainly, many households face important challenges, and I don’t want to minimize these. At the broader macro-level, though, my take is that we have seen broad-based consumption growth—and not something driven solely by the outsized stock market gains among the wealthy.

There is one darkish cloud here. Measures of consumer sentiment are pretty depressed, which might be a sign of reduced spending to come. But remember that sentiment measures’ accuracy as a leading indicator has broken down in recent years. Indeed, at the Chicago Fed we have looked at how adding sentiment measures to forecasting models of consumer spending changes their accuracy and found that, in recent years, adding sentiment indexes literally makes the spending projections worse!²

So let’s not over-index on sentiment numbers. As long as actual consumer spending stays solid, it’s a recipe for stability.

III. Artificial Intelligence (AI)

AI investment has not been as big a driver of the economy as some have portrayed. There has been a great deal of discussion that growth in non-AI-related sectors has been measly and that the capital investment associated with AI data centers has accounted for a dramatic share of gross domestic product (GDP) growth. These discussions seem rooted in fears that the popping of an AI bubble might end the only thing we have driving growth.

I just told you that I think the data suggest that the American consumer is the key driver of growth. But let’s unpack the AI numbers a little. First, it doesn’t make sense to just sort of cover up one eye, look at the non-AI parts of the economy, and say that’s how the economy would be doing if there were no AI boom. The AI and non-AI parts of the economy compete for the same factors of production—land, electricians, computer chips, HVAC equipment. All kinds of different inputs would be cheaper and the non-AI sectors would be doing better if they weren’t paying elevated prices caused by high AI demand.

But second, adding up the direct capital investment parts of AI overcounts how important it has been to GDP growth. Yes, the investment contributes to GDP, but much of that investment is buying imported goods, so we need to subtract that to get the overall GDP effect. For example, a recent paper by economists at the Board of Governors estimates that AI investment in data centers accounts for a tad over four-tenths of GDP growth in 2025, but that properly subtracting off the imported content meant that the net impact of data center investment cut it by more than half—to less than two-tenths of a percentage point.³

I’m not saying there is no AI bubble to worry about. My point is just that GDP growth in 2025 was not primarily driven by frothy AI spending.

IV. The Labor Market

The job market has been steady.

Certainly, a sudden deterioration in the labor market, like those at the start of a recession, could change the policy calculation pretty quickly. And we have experienced an extended period of low aggregate job creation, which some have interpreted as a sign of a weak and weakening labor market that could fall apart in a business cycle sense.

That isn’t my take. Yes, there was little net job creation in 2025, and the first estimates were substantially revised down as the full data came in. But I don’t think that says much about the business cycle or labor market slack. The immigration crackdown clearly affected population growth and labor supply last year, and that reduced the break-even level significantly. Aggregate job creation at a moment like that is one of the least helpful data series for understanding how much slack there is in the labor market.

Exactly the same thing happened, though in the other direction, in 2023 and 2024: Payroll job creation was surprisingly high, but it was not a sign that the economy was overheating or the labor market was especially tight. The unemployment rate was rising for much of that time.

The unemployment rate now, except for a couple of months of statistical pollution surrounding the government shutdown, has basically been steady for a year. In February 2025, it was 4.2%, and going into February 2026, it is 4.3%. Our Chicago Fed Labor Market Indicators, which combine real-time public and private data sources, have also continued to show general steadiness; we will get the model’s forecast for the February unemployment rate on Thursday.

The modest aggregate job numbers can’t be far from the true break-even point or else the unemployment rate would have been rising. That’s basically the definition of what the break-even point is!

When the population and labor supply have question marks—like they did in 2023 and 2024—we found that rate-based measures like the vacancy rate, the layoff rate, and the unemployment rate are better indicators of the labor market. Over the past year, most rate-based measures have shown steadiness (again, other than some of the statistical pollution issues) and at levels comparable to normal expansions previously.

The one weak component in the rates data has been the hiring rate. It is low and that is a warning flag. But it has been relatively steady for a year and a half. And the layoff rate and new unemployment insurance claims have also been low for quite a while, which are normally positive signs.

Remember, low hiring with low firing is not a normal business cycle combination. But it is exactly the combination you would expect if businesses are dealing with heightened uncertainty. When we talk to businesses across the Midwest, we still hear a lot about uncertainty and “let’s see what happens.” The question marks surrounding the impact of last week’s Supreme Court opinion on tariffs and where things go next make it feel like this dynamic may continue.

V. Inflation

A steady real economy should put the focus back on inflation for near-term monetary policy decisions.

That shouldn’t be a surprise given the past five years. The public’s attention never really turned elsewhere. People remain especially concerned about prices.

The inflation rate is well above our 2% target and has been for almost five years now. We made progress in 2023 and 2024, but that progress stalled in 2025. Core PCE inflation for the year was around 3%.

With inflation at 3%, it is not obvious that our interest rate policy is even restrictive. The real federal funds rate measured that way is pretty close or even a bit below the consensus view of long-run r*. And the real rate isn’t much higher if you measure it against year-ahead inflation forecasts.

Three percent inflation is not good enough—and it’s not what we promised when the Federal Reserve committed to the 2% target. Stalling out at 3% is not a safe place to be for a myriad of reasons we know all too well. We need to make more progress.

There have been some encouraging developments on the inflation front. Housing, for example, has improved nicely. But there have been some warning signs, too. Goods inflation went the wrong way in 2025. Multiple studies show that the U.S. has absorbed the vast bulk of the cost of tariffs.⁴ Precisely how much has found its way downstream to the consumer so far and how much more is on the way is not yet known.

The optimistic case is that if a lot of the increase in inflation came from tariffs and if the tariffs have already passed through to consumers, the inflation effect will be transitory. If so or if the Supreme Court’s decision leads to lower overall tariff rates, we could be back on path to target inflation in the near term.

It’s not just about goods inflation, though. Core PCE services excluding housing inflation has been running stubbornly high, at 3.3%, over the past year. That is very unlikely to have come from tariffs, and it’s harder to make an optimistic case that high services inflation is just transitory. So we need to be vigilant.

I remain optimistic that there can be more rate cuts this year. But that hinges on seeing actual progress on inflation that shows we are on a path back to 2%. Forecasters have projected that we will see progress on inflation in the near term. Goods inflation, particularly, should start fading.

Since we have been burned by assuming transitory inflation before and since there are plausible scenarios where inflation again proves more persistent than we forecasted, we should be careful not to put ourselves in a difficult position.

The fact that over the past several months, the date at which the forecasts say inflation will start falling keeps getting pushed back is not a great sign.

I feel that front-loading too many rate cuts is not prudent in that circumstance. In every economic survey we have seen and in the many meetings we have had with businesses and consumers across the Seventh District, people express that prices are one of their most pressing concerns. Let’s pay attention. Before we cut rates more to stimulate the economy, let’s be sure inflation is heading back to 2%.

VI. Conclusion

So that’s my overall view. Economic growth and the labor market don’t seem especially fragile, but there is a nagging bit of inflation that I hope, and expect, will subside soon and will allow us to get back on the golden path.

In recent years we’ve had to face many unexpected and even unprecedented shocks, so I’m sure 2026 will have a few surprises up its sleeve. But we will keep at it. As I always say—and it’s worth remembering in the midst of a record-breaking blizzard—our motto at the Chicago Fed is that there is no bad weather, only bad clothing. Though, that was easier to stomach last week when it was 61 degrees in Chicago.

Notes

¹ Chakrabarti et al. (2026).

² Brave et al. (2025). This unpublished special memo is available upon request.

³ Brandsaas et al. (2025).

⁴ See, for example, Amiti et al. (2026), Gopinath and Neiman (2026), and Hinz et al. (2026).

References

Amiti, Mary, Chris Flanagan, Sebastian Heise, and David E. Weinstein, 2026, “Who is paying for the 2025 U.S. tariffs?,” Liberty Street Economics, Federal Reserve Bank of New York, blog, February 12, available online.

Brandsaas, Eirik Eylands, Daniel Garcia, Robert Kurtzman, Joseph Nichols, and Adelia Zytek, 2025, “Estimating aggregate data center investment with project-level data,” Finance and Economics Discussion Series, Board of Governors of the Federal Reserve System, No. 2025-109, December. Crossref

Brave, Scott A., Benjamin L. Henken, Katherine Jolley, and Walker D. Ray, 2025, “How much signal should we take from consumer sentiment,” Federal Reserve Bank of Chicago, unpublished staff memo, April.

Chakrabarti, Rajashri, Thu Pham, Beck Pierce, and Maxim Pinkovskiy, 2026, National Consumer Spending Report: Updated Through December 2025, Economic Heterogeneity Indicators, Federal Reserve Bank of New York, February 3, available online.

Gopinath, Gita, and Brent Neiman, 2026, “The incidence of tariffs: Rates and reality,” National Bureau of Economic Research, working paper, No. 34620, revised February 2026 (originally issued January 2026). Crossref

Hinz, Julian, Aaron Lohmann, Hendrik Mahlkow, Anna Vorwig, 2026, “America’s own goal: Who pays the tariffs?,” Kiel Institute for the World Economy, policy brief, No. 201, January, available online.

The views expressed today are my own and not necessarily those of the Federal Reserve System or the FOMC.