“IT’S THE ECONOMY, STUPID,” famously exclaimed James Carville in urging campaign workers to emphasize pocketbook issues in Bill Clinton’s successful run for the presidency in 1992.

He was hardly the first to make the connection. Indeed, ever since the 1970s, the Yale economist Ray Fair has been making election predictions using a statistical model that leans heavily on macroeconomic performance. And while the model doesn’t have an especially good track record, few have doubted that variables ranging from unemployment to growth to inflation largely determine people’s views of the state of the economy, which in turn has an outsize impact on how they judge Washington policymakers’ success.

Until now, anyway. In the wake of the pandemic, the link between the economy’s performance and consumers’ perceptions about it appears to have broken, with implications for topics ranging from macroeconomic forecasting to the political returns from economic policy. In this essay, we attempt to answer the question: if the economy is doing so well, why are consumers so glum about it?

How're We Doing?

To start, it’s useful to review how economists measure consumers’ views on the economy. The most common approach is to ask them. One popular source is the University of Michigan’s Consumer Sentiment Index, which queries Americans on their perceptions of the current and future macroeconomy, as well as their current and future finances. The other main source is the Conference Board’s Consumer Confidence Survey, which asks about present business and employment conditions as well as expectations about business conditions, employment conditions and family income six months in the future.

While these indexes use established and reputable survey techniques, they are partially based on questions that cannot be answered by firsthand experience. Naturally, responses to these questions are influenced by where respondents get their news and their views about the political climate, opening up the possibility of a divergence between hard data and consumers’ views.

The divergence has only recently become an issue, with the gap emerging at the start of the pandemic and persisting until today. To show this, we constructed a measure of predicted sentiment using macroeconomic variables – including the unemployment rate, inflation, aggregate consumption and the performance of the stock market – and compare it to the actual sentiment measure. As shown in the chart below, prior to the pandemic, predicted sentiment, which is just a weighted average of the macroeconomic variables, moved in lockstep with the actual consumer sentiment measure.

Since the start of the pandemic, however, this relationship has fallen apart. For example, in the summer of 2024, the economy was exceptionally strong: unemployment was at near-record lows, the stock market was booming, the U.S. economy was expanding nicely and inflation had retreated to normal levels. Yet consumer sentiment was roughly at the same level as in winter 2011 when the unemployment rate was over twice as high and economic growth was sluggish.

While this decoupling is clear, the reasons behind it are not. Analysts and commentators often oversimplify the explanation, citing single factors like inflation, the media or general discontent as the sole explanation. The truth is more complicated.

Consumer Attitudes Versus Consumer Behavior

Before turning to possible explanations for the disconnect between sentiment and the macroeconomy, it’s important to note that Americans – despite their reported feelings – have been quite enthusiastic in their economic behavior.

For one thing, workers have been acting as though they were highly confident in their ability to find jobs. This confidence is best measured by the rate at which workers voluntarily leave their jobs. In the 20 years prior to the pandemic recovery, the “worker quits rate” – the percentage of workers quitting their jobs in a given month – varied between 1.2 percent and 2.4 percent, with the rate increasing with labor market strength. However, over the course of 2021 and 2022 the quit rate soared, reaching 3.0 percent in fall 2021. Workers were clearly confident that they could find other jobs if they left their current ones.

A second indicator of households’ economic outlook is their consumption behavior. Coming out of the pandemic, consumers eagerly spent their newfound savings, pushing household spending far above pre-pandemic trends. Between January 2021 and November 2024, real (inflation-adjusted) consumer spending grew by 3.5 percent annually, compared to 2.4 percent on average in the 2010s. And consumers weren’t just spending on necessities: expenditures on luxuries such as travel soared. Whereas real air transportation spending was typically around $80 billion to $100 billion in the decade preceding the pandemic, outlays soared to nearly $140 billion in 2022.

Another useful bellwether of consumers’ faith in the economy is entrepreneurship. If Americans think the economic outlook is positive, they are more likely to start new businesses and expand existing ones. Coming out of the pandemic, the U.S. experienced an entrepreneurship boom that matched the explosion in consumer spending. In the decade before the pandemic, an index of the “firm birth rate” – measuring the annual rate at which new businesses are created – was typically around 0.84 to 1.00. Over the course of the pandemic, Ryan Decker (Federal Reserve) and John Haltiwanger (Maryland) found that the firm birth index rose sharply, hitting 1.14 in 2022. And the entrepreneurship boom is likely to continue because new business applications in 2024 came in roughly 50 percent higher than the year before the pandemic.

These trends in consumer behavior only deepen the mystery of why U.S. sentiment was so low in the post-pandemic period. The economy was booming and people’s behavior embodied an optimism to match, with depressed sentiment in consumer surveys being the outlier. We now turn to possible explanations for why people reported feeling down on the economy when the economy’s performance – and their spending and work force behavior – suggested an entirely different picture.

It's Inflation, Stupid

The most obvious explanation for the disconnect between consumer sentiment and the macroeconomy is inflation. On its face, this explanation has some merit. Between January 2020 and June 2022, year-over-year prices rose by 6.5 percentage points, topping out at 9 percent, the highest level in the U.S. since the early 1980s. And Americans were not shy in reporting their displeasure: a 2022 Pew survey found that 70 percent of respondents regarded inflation as “a very big problem,” a figure that put it substantially ahead of any other public policy issue.

Yet there are reasons to question inflation as formally measured as the sole driver of discontent. First, ascribing all of the decoupling to inflation implicitly assumes that other factors – the five-decade-low level of unemployment, historically high balances in checking accounts owing to unprecedented government support during Covid-19, and a booming stock market – don’t matter.

Second, the relationship between consumer sentiment and the macroeconomy already accounts for inflation. Our predicted measure of sentiment includes the historical impact of changes in inflation on consumer attitudes. Third, the gap between predicted and actual sentiment continued through 2024 despite year-over-year inflation dropping to around 3 percent, within the normal range over most consumers’ lifetimes and approaching the Federal Reserve’s 3 percent target. Fourth, households widely believe the inflation rate is substantially higher than it actually is. In a household survey conducted between December 2023 and January 2024, economist Stephanie Stantcheva (Harvard) found that when asked about the level of inflation over the past year, the average respondent reported it was 7.1 percent when the true value was 3.3 percent.

This gap begs the question: are consumers actually angry about inflation as formally measured, or are they angry about something else that they conflate with inflation? Our view is that inflation played a major role in sentiment decoupling because people care not only about inflation – which is the speed at which prices are increasing – but also about the level of prices. While formally measured year-on-year inflation had dropped to 3.3 percent when Stantcheva conducted her survey, prices were up 19.2 percent since the start of the pandemic. Importantly, it was the high levels of prices that people referred to when they said they were mad about inflation, indicating that cumulative inflation contributed to their economic anguish.

Consumers do eventually adapt. Although a bottle of Coke cost a nickel from 1886 until the 1950s, consumers have since adjusted to its much, much higher price. But how quickly do consumers acclimate?

To answer this question, two authors of this essay (Cummings and Mahoney) crunched the numbers and found that the negative impulse from inflation “decays” at a rate of one-half per year. This suggests that it takes roughly three years for the impact of higher prices to mostly work its way through the system. To put this in concrete terms, since inflation peaked at 9 percent in June 2022, the model predicts that consumers would not fully adapt to 2022 prices until 2025.

Republicans Cheer Louder and Boo Harder

Following the latest election victory of Donald Trump, consumers’ views on the economy underwent a remarkable shift. In the December 2024 consumer sentiment survey, Republican views on the economy immediately shot up while Democrats’ views plunged. The shift was not a surprise. Analysts observed a similar phenomenon when the White House changed hands following the 2020 elections and with prior charges in political party control. Less widely recognized, however, is the partisan skew in the shift. Analyzing the full series of data, Cummings and Mahoney found that the partisan effect for Republicans is roughly 2.5 times that for Democrats:

When a Republican is in the White House, Republican survey respondents feel about 15 index points better than predicted about the economy, whereas Democrats feel around 6 index points worse. When a Democrat is President, Republicans feel about 15 index points worse than the economy, but Democrats only feel around 6 index points better. This roughly +/-15 point swing for Republicans versus the +/-6 point swing for Democrats is what we term asymmetric amplification.

In other words, Republicans cheer louder and boo harder, with the “asymmetric amplification” depressing sentiment when a Democrat controls the White House and inflating sentiment when a Republican is in control. During the Biden presidency, this force depressed sentiment by seven percentage points, explaining roughly one-third of the gap between observed and predicted sentiment. Today, while we don’t have enough postelection data for precise estimation, the same calculations suggest that excess enthusiasm from Republicans could be giving sentiment a significant boost.

It's the Media's Fault

Along with the residual impact of elevated inflation and asymmetric amplification, the media likely played a role. First, legacy media – TV, radio, newspapers – may have become more negative over time, dragging down sentiment. Second, changes in our media diet – namely, larger servings of social media – may have had a depressing effect.

To shed light on the potential role of increasing negativity bias, it is helpful to study the economic news index produced by the Federal Reserve Bank of San Francisco. This index uses a technique called lexical analysis to characterize the positive or negative sentiment of economic news stories from two dozen major U.S. newspapers, including The New York Times, The Washington Post and The Wall Street Journal, at a daily frequency.

In work with economist Aaron Sojourner (Upjohn Institute), one of us (Harris) analyzed whether the trajectory of economic news sentiment followed the path predicted by economic metrics, including economic growth, unemployment, inflation and stock prices. The analysis revealed a sharp divergence between actual economic news sentiment and predicted news sentiment, with a substantial negative bias beginning in 2018 and accelerating from 2021 to 2023. By late 2023, the gap had mostly closed, but the six years of atypically negative messaging likely impacted consumers’ views on the economy.

There is other evidence supporting a systemic negative shift in the media. For example, research by Jules van Binsbergen (Wharton) and colleagues examines over 200 million pages from 13,000 local newspapers to characterize economic news sentiment and finds that local news’ characterization of the economy has been trending downward since 1960 – with a sharper downturn since the turn of the century.

In another piece, two of us (Cummings and Mahoney) with Stanford’s Giacomo Fraccaroli identified a growing negative bias in TV news reporting on gas prices. Analyzing over one million transcripts from popular news channels from 2004 to 2023, the authors found that coverage of gas prices spikes whenever nominal prices pass $3.50 per gallon. Since this is an increasingly common occurrence due to inflation, it means that the real threshold at which TV news sounds the alarm about gas prices has plummeted by roughly one-third between 2004 and 2023.

The other potential explanatory factor is the increasing presence of social media. Sentiment is strongly correlated with the medium of news consumption, with those who primarily receive their news from social media showing lower consumer sentiment than those who get their news from traditional media. Of course, this evidence demonstrates correlation rather than causation, and it is still unclear whether and to what extent social media has a causal effect on negative sentiment.

That said, there are a number of potential channels through which increased use of social media could depress sentiment. It may cause unhappiness, which in turn may cause people to report a more general negative outlook, including views on the economy and their own financial prospects. Another possibility is that social media is increasing the amount of disinformation in circulation, which distorts consumers’ economic perceptions – likely in a negative direction.

Whether due to the growth of social media consumption or some other factor, Americans frequently report dramatic mischaracterizations of the macroeconomy. To take one pertinent example, a Morning Consult poll administered in November 2022, when the U.S. economy was growing rapidly, revealed that 65 percent of Americans thought the economy was in a recession, and less than one-quarter correctly reported that it was not.

This schism between the actual economy and Americans’ perceptions of it presents a major challenge in interpreting consumer sentiment. For example, in November 2022 when the Morning Consult survey was conducted, the Michigan Survey of Consumer Sentiment Index stood at just 56.7 – roughly the same level observed in the depths of the Great Recession. But the low sentiment coupled with the sweeping mischaracterization begs the question: were households displeased with the actual economy, or instead reporting unhappiness over an economy that was only imagined to be in recession?

Does the Paradox Even Matter?

Like the famous murder on the Orient Express, the decline in consumer sentiment is a victim with many perpetrators. The largest contributor, in our view, was the lingering effects of inflation, which we estimate account for one-third of the gap. Asymmetric amplification played an important role, explaining perhaps another third. Systematic negative media bias and the ongoing emergence of social media likely contributed, although it’s hard to pin down the amount.

Other factors have made a difference, too. Since older Americans generally hold more negative views, population aging provides modest downward pressure on sentiment. One of us (Cummings) found that a shift in survey collection methods (from phone to online), which occurred only recently, is depressing the current readings by more than 10 percent. Wall Street Journal columnist Greg Ip has another hypothesis he calls “referred pain”: general displeasure about domestic and geopolitical events bleeding over into economic perceptions.

Drivers of the paradox aside, it’s reasonable to ask whether divergence between the economy and sentiment even matters. We think it does for several reasons.

To start, this paradox not only describes a divergence between the economy and sentiment but portends a broader slide into pessimism. It’s tough to pinpoint the culprit, but it’s clear that Americans are becoming less content with their lives. Economists David Blanchflower (Dartmouth) and Alex Bryson (University College London) have found that the left side of the historic U-shaped pattern for life satisfaction – where happiness was high in the very young and very old – has flattened, with young adults no longer exhibiting the elevated happiness they once did. Economists Anne Case and Angus Deaton of Princeton have chronicled the disturbing rise in deaths of despair – drug overdoses, alcohol abuse and suicide – that emerged in the late 1990s and have steadily driven up midlife mortality rates, especially among those without college degrees.

Second, consumer sentiment has traditionally been seen as a harbinger of future consumer spending: when people feel good about the economy, they spend more. By the same token, lower sentiment has meant less spending and a possible recession. If current economic conditions have a diluted impact on sentiment and sentiment becomes disconnected from future spending, then the leading indicators of macroeconomic performance have become less reliable.

Third, the disconnect has implications for conventional wisdom on the connection between the economy and electoral outcomes. Economist Robert J. Gordon (Northwestern) looked at the links between economic indicators and sentiment, approval ratings and elections and found that prior to 2000, two economic variables – “excess” inflation and real GDP growth – accurately predicted election outcomes. He has also found that this relationship has deteriorated in recent years. In a world in which economic performance maps less directly to consumers’ views and electoral outcomes, politicians may shift their attention away from economic well-being and toward social issues and other topics.

* * *

There is still much we don’t know about the disconnect, but some lessons are clear. Americans abhor inflation until they adapt to higher prices – so much so that elevated inflation obscures other positive aspects of the economy. Intense, asymmetric political polarization on the right has important implications across society, including the ways people feel about the economy. And along with its much-discussed social and political implications, the rise of social media likely muddies Americans’ economic perceptions even more.