Consumer sentiment is at an all-time low because prices are at an all-time high. The UMich index isn't broken, popular government data just offer an incomplete picture of what people care about
Elliott, when I retired at the end of 2013, my husband and I had what we believed was enough to live comfortably, if not richly, for the rest of our lives. We have pensions, Social Security, an annuity, and an investment portfolio. Between 2015 and today, prices for housing, heating, fuel, and food have risen so steeply that we are no longer comfortable. Housing has risen the most steeply. If you want a metric, look at Zillow.
One more comment which has been said by others before. To most people, increases in wages feel like being rewarded for something they did. Inflation feels like something being done to them - a big difference in "how it feels". And in this case, especially after a very long (by historical standards) of very stable prices that a lot of folks grew up with.
I do think this is also tied to the K shaped economy. I strongly suspect that folks on the upper part of the K - helped by rising stock prices, etc. still see higher prices, but it has a much smaller effect as it is less of their budget. Folks on the lower part of the K shaped economy feel the hit of higher prices far more. And more of the population is on the lower part of the K. My guess is that what is described in the article would be even more pronounced if only looking at the lower part of the K.
The followup question - how does this feeling that prices are "too high" last?
A story to illustrate this: I was a magazine marketing guy early in my career and we constantly tested prices for an annual subscription for potential new subscribers. If we tested a panel with a price 10% higher than our standard offer (our control), we saw a 10% drop in the number signing up - and a constant in terms of dollars (nicely matching macroeconomic theory). These were identical samples in mailed offers, coming from the same mix of mailing lists with the same content except the process different. We actually tested a range of prices and saw this pattern of declining dignios with rising prices and constant dollar demand.
BUT our tests would reach a higher price where demand collapsed. Total dollars fell. A price cliff. For example, our price might be $7.95 and we tested 8.95 and 9.95 (yes these prices are from long ago). Demand collapsed at the $9.95 price and we learned we couldn't go above $8.95. So we'd go up to $8.95 and then be stuck there for a few years until we saw perceptions change and we could go to $9.95 without a demand collapse.
So consumer perceptions of a "fair price" will change but slowly.
I find your news sentiment study quite compelling. Given that there appear to be two entirely disjoint news spaces, namely Fox News and its adjacent players and "legacy" media, is there a difference in news sentiment between those two universes?
Or perhaps I'm completely out to lunch (it happens often) and the real separate universes are classical media (newspapers, TV, radio) and social media (FB, Instagram, Twitter, and so on). Does news sentiment vary between *these* universes?
And, worse yet, with the opaque precision targeting available with the social media platforms, is it even possible to measure the media sentiment for the social media world?
Great graphs, but I'm not sure they are measuring the level of anxiety after COVID and with AI all over the news.
Prices may still not be rising, but every time buying at the new, higher prices, people still remember how things were stable and at the price they "should be" not so long ago - so a reminder that things are unstable and out of my control, and could go roaring again at any minute - Anxiety. And now WAR! gas prices roaring totally out of control now - double anxiety (huge relief we went all electric in 2018, never looked back...)
Unemployment may be low TODAY, but the AI buzz is that all these jobs are going to be taken by robots and chatbots - and then what? More things the consumer has no control over. Predictability is at an all time low. But senses bad things are coming for all but the extremely wealthy. Anxiety abounds.
Some measure of perceptions of volatility and income disparity might be your missing link.
Dang, you’re so good! Love the numbers, analysis, and your ability to boil it down into something easily understood, a rare combination of abilities. I also agree with what others have said about the fat cat gap. The average person works just as hard and probably even harder than the wealthy 1% and they can’t get ahead. The game is rigged and they’re seeing it in real time. That’s not a vibe, but reality
Yes! Especially young people! Everything is stacked against them as they try to achieve their personal American dream, and, in most cases, it doesn't feel like trying harder will ever get them ahead. My son and friends are in their early 30s now, and starting families, trying to buy houses. They're all just treading water and hoping another wave (hello, gas prices!) doesn't push them under.
COVID really magnified who was essential. Wasn’t the C Suite. But once that was over those so called essential workers were kicked to the curb, many/most of whom are younger.
Today's SIN is a real service. Yes: price levels, price anxiety -- got it. These act also, however, as a proxy for other deep disruptions to one's living patterns. Ex. beyond-reach housing prices, esp. for aspiring 1st time buyers. Bs know that someone else [ aka 'fat cats' ] is profiting mightily from the stratospheric list prices even though they are shut out of the market. "Why should fat cats be able to succeed while I cannot [enter the market]?" -- unavoidable question. Through another prism: fighting over splitting up property during divorce proceedings isn't as much about the pieces of property, as it is about self-image, pain & anxiety, and seeing that your lifestyle is degrading/altering. US life will always be dogged by these losses underlying $$$ realities.
I appreciate your analysis yet I wonder if the disconnect between consumer sentiment and aggregate economic statistics is not just a result of "vibes" and persistent high prices, but also record income inequality? Looking at unemployment, wages and the stock market doesn't account for the fact that the gap between the wealthiest Americans and the rest of the population is the widest since record-keeping began in 1989. Consumer sentiment is measuring a broader segment of the population who are not keeping up. I'm sure if you measure the sentiment of the top 1% they'll feel the economy is doing just fine.
It might be "too much " to ask GEM to do this, too, but if it be possible to sort the data by an additional factor, i.e., income deciles, without too much noise, this might well be informative.
Good Grief! I love how you write. I'm more a word person than a numbers person, which is why I appreciate your analysis. It gives both. I can't hope (and don't really want to) nerd out with every graph--which is one reason why I appreciate that you do. I can follow your basic methodology but then you sum it up so well with words. Many thanks!
I guess the next question from there, though, is how does this end? Prices aren't dropping back to 2019 levels, and in the past six years Americans have never gotten used to them (partly because Trump now is doing things to very visibly make them higher). I think one reason people don't like these data is because it cannot be resolved using anything we can see right now - either people will never feel good about the economy again, or people will go back to caring about prices less, and we have no idea when that will happen, why that will happen, or what they'll care about next.
(it's also not theoretically great for capitalism - prices never go down absent a recession, so if people feel worse whenever prices have gone up the only way to make people happy would be a command economy with prices set by the government).
Price attachment points are a great explanation for sentiment. As a person of a certain age, I remember (still remember) nickel candy bars and the horror of candy bars costing a dime. Almost impossible for me to spend two bucks on a candy bar yet here we are. Thanks for the heavy lift to find the dominant factor.
(1) I love this for the sheer data visualization nerdiness of it. Thank you. Speaking of high nominal prices, each month I've been pruning subscriptions, just because even though I am quite comfortable I gave myself a discretionary spending budget and my various newsletter subscriptions eat up a lot of it. Stuff like this is what saves you from those sheers. Some other substack had to go!
(2) As a heads up, gmail at least clipped this one in my email inbox. I could still read it but had to click through. I had never seen that before, but I guess there's some maximum message length. I don't know if that changes anything, just sharing in case y'all were also unaware of that as I was. My guess is if you see any sort of analytics drop off at the end of your piece, that might contribute. The message break is at the end of this paragraph block in section 2: "... It’s only when we add a measure of price levels — either the structural Bernstein-style cumulative price shock or the direct survey of how consumers feel about prices — that the models catch up to reality. Predictions from those models are shown as the red and teal lines of the main plot: ..." Then snip.
(3) Are you publishing this in a white paper format at all, like with academic style background and methodology sections? I get that the job is this substack style communication, but the empirical meat on the bones of this argument presented in a more buttoned up format could be worth something provided all the data and methodology stand up under a peer review. I'm not even directly suggesting submitting the work for academic publishing, but I think white papers move things in certain circles in ways that a substack post is still just a substack post.
(4) More of a philosophical point because I found your nominal price argument very convincing, but my personal ontology is that nothing is only one thing. So I don't think we just say it's nominal prices and stop.
I think nominal prices is a yes and thing, but I don't know how to go about a quantitative look at some of the other things, much less how you'd slap them altogether onto one chart.
-Job growth is at best flat, maybe negative, unless you're in healthcare. So that might make you sour if you don't work in healthcare.
-GDP growth is largely in AI, and AI financing is causing more and more "Are we in a bubble?" eyebrow raising think pieces. So, unless you're directly profiting from AI speculation and circular financing, you might not be feeling the GDP growth, and that might make you sour.
-Labor force participation is down, which means that employment statistics might not be representing the total pool of unemployed people who are feeling sour.
-Jobs numbers are not sufficiently distinguishing between people who are earning comfortably in secure jobs versus taking on gig work to pay bills. Under employed folks might be feeling sour.
I don't really love the general orientation around "Macro numbers good, economy good." It is very "The map is not the territory" and "Ceci n'est pas une pipe" to me. Just because a macro number is good, it doesn't automatically mean the thing being measured is good, just that so long as we are not misaligned and our measurements are good reflections of reality then we're good. But the possibility I don't see enough pieces tackling is that the measurements themselves are missing the reality of the economy.
There are methodological reasons it makes sense for the Fed to exclude them for the Fed's purposes. But for the rest of us, especially any Journalist trying to report out "why do people feel bad about the economy," they ought to consider food and energy. Agreed!
I agree with the pushback here from the journalistic POV, but for the narrow predictive question in the article core PCE produces predictions of sentiment than headline PCE. Interesting to think about why that might be! But yes, when talking about prices, we should include the numbers that impact people at home
1. Prices are always at an all-time high with the exception of times of deflation (the gift that keeps on taking). It’s just the math.
2. Aggregates are a useful tool, but only a tool and typically a cruel one. In particular, they are blind to a K-shaped economy where folks are increasingly on the wrong leg of the K.
3. Sure, the numbers matter on their own. Twelve bucks for a plate of migas is jarring. But, often, it’s the sheer number of expenses — and the sheer number that keep becoming increasingly non-trivial — that get you. And when you’re *still* on the wrong leg of the K?
4. It’s the distribution, stupid. We’re forty-five years into the dismantlement (though not without some significant pauses) of the Great American Postwar Middle Class. The evil brilliance of the right has been to parlay the anger this has justly generated into support for the interests of the most rapacious among us. You know, the immigrant wants your cookie and all.
So maybe we can get to the next chapter without an intervening existential crisis?
And, if I may — something that gets to me at a meta-level across economics, politics, system performance and, hell, probably even sports, is that we (self included) constantly take valuable multidimensional metrics and collapse them down to scalars, getting rid of large amounts of their meaning in the process.
Historically, that was a necessity; we were bound to the printed page. But we can get past that and learn to present and learn to read graphical depictions of data in a more useful way.
Elliott, when I retired at the end of 2013, my husband and I had what we believed was enough to live comfortably, if not richly, for the rest of our lives. We have pensions, Social Security, an annuity, and an investment portfolio. Between 2015 and today, prices for housing, heating, fuel, and food have risen so steeply that we are no longer comfortable. Housing has risen the most steeply. If you want a metric, look at Zillow.
One more comment which has been said by others before. To most people, increases in wages feel like being rewarded for something they did. Inflation feels like something being done to them - a big difference in "how it feels". And in this case, especially after a very long (by historical standards) of very stable prices that a lot of folks grew up with.
I do think this is also tied to the K shaped economy. I strongly suspect that folks on the upper part of the K - helped by rising stock prices, etc. still see higher prices, but it has a much smaller effect as it is less of their budget. Folks on the lower part of the K shaped economy feel the hit of higher prices far more. And more of the population is on the lower part of the K. My guess is that what is described in the article would be even more pronounced if only looking at the lower part of the K.
The followup question - how does this feeling that prices are "too high" last?
A story to illustrate this: I was a magazine marketing guy early in my career and we constantly tested prices for an annual subscription for potential new subscribers. If we tested a panel with a price 10% higher than our standard offer (our control), we saw a 10% drop in the number signing up - and a constant in terms of dollars (nicely matching macroeconomic theory). These were identical samples in mailed offers, coming from the same mix of mailing lists with the same content except the process different. We actually tested a range of prices and saw this pattern of declining dignios with rising prices and constant dollar demand.
BUT our tests would reach a higher price where demand collapsed. Total dollars fell. A price cliff. For example, our price might be $7.95 and we tested 8.95 and 9.95 (yes these prices are from long ago). Demand collapsed at the $9.95 price and we learned we couldn't go above $8.95. So we'd go up to $8.95 and then be stuck there for a few years until we saw perceptions change and we could go to $9.95 without a demand collapse.
So consumer perceptions of a "fair price" will change but slowly.
I find your news sentiment study quite compelling. Given that there appear to be two entirely disjoint news spaces, namely Fox News and its adjacent players and "legacy" media, is there a difference in news sentiment between those two universes?
Or perhaps I'm completely out to lunch (it happens often) and the real separate universes are classical media (newspapers, TV, radio) and social media (FB, Instagram, Twitter, and so on). Does news sentiment vary between *these* universes?
And, worse yet, with the opaque precision targeting available with the social media platforms, is it even possible to measure the media sentiment for the social media world?
Great graphs, but I'm not sure they are measuring the level of anxiety after COVID and with AI all over the news.
Prices may still not be rising, but every time buying at the new, higher prices, people still remember how things were stable and at the price they "should be" not so long ago - so a reminder that things are unstable and out of my control, and could go roaring again at any minute - Anxiety. And now WAR! gas prices roaring totally out of control now - double anxiety (huge relief we went all electric in 2018, never looked back...)
Unemployment may be low TODAY, but the AI buzz is that all these jobs are going to be taken by robots and chatbots - and then what? More things the consumer has no control over. Predictability is at an all time low. But senses bad things are coming for all but the extremely wealthy. Anxiety abounds.
Some measure of perceptions of volatility and income disparity might be your missing link.
Great graphs and words. The other crucial factor may be income distribution or the K-shaped economy. Go for it, Eliot!
Dang, you’re so good! Love the numbers, analysis, and your ability to boil it down into something easily understood, a rare combination of abilities. I also agree with what others have said about the fat cat gap. The average person works just as hard and probably even harder than the wealthy 1% and they can’t get ahead. The game is rigged and they’re seeing it in real time. That’s not a vibe, but reality
Yes! Especially young people! Everything is stacked against them as they try to achieve their personal American dream, and, in most cases, it doesn't feel like trying harder will ever get them ahead. My son and friends are in their early 30s now, and starting families, trying to buy houses. They're all just treading water and hoping another wave (hello, gas prices!) doesn't push them under.
COVID really magnified who was essential. Wasn’t the C Suite. But once that was over those so called essential workers were kicked to the curb, many/most of whom are younger.
Today's SIN is a real service. Yes: price levels, price anxiety -- got it. These act also, however, as a proxy for other deep disruptions to one's living patterns. Ex. beyond-reach housing prices, esp. for aspiring 1st time buyers. Bs know that someone else [ aka 'fat cats' ] is profiting mightily from the stratospheric list prices even though they are shut out of the market. "Why should fat cats be able to succeed while I cannot [enter the market]?" -- unavoidable question. Through another prism: fighting over splitting up property during divorce proceedings isn't as much about the pieces of property, as it is about self-image, pain & anxiety, and seeing that your lifestyle is degrading/altering. US life will always be dogged by these losses underlying $$$ realities.
I appreciate your analysis yet I wonder if the disconnect between consumer sentiment and aggregate economic statistics is not just a result of "vibes" and persistent high prices, but also record income inequality? Looking at unemployment, wages and the stock market doesn't account for the fact that the gap between the wealthiest Americans and the rest of the population is the widest since record-keeping began in 1989. Consumer sentiment is measuring a broader segment of the population who are not keeping up. I'm sure if you measure the sentiment of the top 1% they'll feel the economy is doing just fine.
It might be "too much " to ask GEM to do this, too, but if it be possible to sort the data by an additional factor, i.e., income deciles, without too much noise, this might well be informative.
Good Grief! I love how you write. I'm more a word person than a numbers person, which is why I appreciate your analysis. It gives both. I can't hope (and don't really want to) nerd out with every graph--which is one reason why I appreciate that you do. I can follow your basic methodology but then you sum it up so well with words. Many thanks!
I guess the next question from there, though, is how does this end? Prices aren't dropping back to 2019 levels, and in the past six years Americans have never gotten used to them (partly because Trump now is doing things to very visibly make them higher). I think one reason people don't like these data is because it cannot be resolved using anything we can see right now - either people will never feel good about the economy again, or people will go back to caring about prices less, and we have no idea when that will happen, why that will happen, or what they'll care about next.
(it's also not theoretically great for capitalism - prices never go down absent a recession, so if people feel worse whenever prices have gone up the only way to make people happy would be a command economy with prices set by the government).
Price attachment points are a great explanation for sentiment. As a person of a certain age, I remember (still remember) nickel candy bars and the horror of candy bars costing a dime. Almost impossible for me to spend two bucks on a candy bar yet here we are. Thanks for the heavy lift to find the dominant factor.
(1) I love this for the sheer data visualization nerdiness of it. Thank you. Speaking of high nominal prices, each month I've been pruning subscriptions, just because even though I am quite comfortable I gave myself a discretionary spending budget and my various newsletter subscriptions eat up a lot of it. Stuff like this is what saves you from those sheers. Some other substack had to go!
(2) As a heads up, gmail at least clipped this one in my email inbox. I could still read it but had to click through. I had never seen that before, but I guess there's some maximum message length. I don't know if that changes anything, just sharing in case y'all were also unaware of that as I was. My guess is if you see any sort of analytics drop off at the end of your piece, that might contribute. The message break is at the end of this paragraph block in section 2: "... It’s only when we add a measure of price levels — either the structural Bernstein-style cumulative price shock or the direct survey of how consumers feel about prices — that the models catch up to reality. Predictions from those models are shown as the red and teal lines of the main plot: ..." Then snip.
(3) Are you publishing this in a white paper format at all, like with academic style background and methodology sections? I get that the job is this substack style communication, but the empirical meat on the bones of this argument presented in a more buttoned up format could be worth something provided all the data and methodology stand up under a peer review. I'm not even directly suggesting submitting the work for academic publishing, but I think white papers move things in certain circles in ways that a substack post is still just a substack post.
(4) More of a philosophical point because I found your nominal price argument very convincing, but my personal ontology is that nothing is only one thing. So I don't think we just say it's nominal prices and stop.
I think nominal prices is a yes and thing, but I don't know how to go about a quantitative look at some of the other things, much less how you'd slap them altogether onto one chart.
I saw someone post somewhere that the number of sellers in real estate is way higher than buyers right now which is a major recession indicator (https://www.threads.com/@fluent.in.finance/post/DXFhnzJEcWO/just-in-home-sellers-now-outnumber-buyers-by-thats-the-largest-gap-ever). Or labor force participation has still not recovered since covid (https://www.brookings.edu/articles/the-us-labor-market-post-covid-whats-changed-and-what-hasnt/). Or we know that a lot of our job growth is lower quality "gig economy" type jobs, which means folks who are being counted as "employed" in surveys might be underemployed and quite sour on the economy (https://qz.com/gig-workers-hiring-jobs-economy). And we know a lot of the actual employment growth is in one sector, healthcare, which not everyone participates in and is subject to weird fluctuations like striking nurses going back on the job (https://www.barrons.com/livecoverage/february-jobs-report-data-2026-today/card/strikes-hit-health-care-job-growth-t4gGxUdeshcXtsBCrUiZ). And that a lot of GDP growth is a bit of a sugar high from increasingly circular and super sus (to borrow some youth parlance) AI + semiconductor financing (https://www.stlouisfed.org/on-the-economy/2026/jan/tracking-ai-contribution-gdp-growth, https://www.bloomberg.com/graphics/2026-ai-circular-deals/).
So, if we take a step back:
-Job growth is at best flat, maybe negative, unless you're in healthcare. So that might make you sour if you don't work in healthcare.
-GDP growth is largely in AI, and AI financing is causing more and more "Are we in a bubble?" eyebrow raising think pieces. So, unless you're directly profiting from AI speculation and circular financing, you might not be feeling the GDP growth, and that might make you sour.
-Labor force participation is down, which means that employment statistics might not be representing the total pool of unemployed people who are feeling sour.
-Jobs numbers are not sufficiently distinguishing between people who are earning comfortably in secure jobs versus taking on gig work to pay bills. Under employed folks might be feeling sour.
I don't really love the general orientation around "Macro numbers good, economy good." It is very "The map is not the territory" and "Ceci n'est pas une pipe" to me. Just because a macro number is good, it doesn't automatically mean the thing being measured is good, just that so long as we are not misaligned and our measurements are good reflections of reality then we're good. But the possibility I don't see enough pieces tackling is that the measurements themselves are missing the reality of the economy.
"the Fed's preferred inflation that omits noisy changes in food & energy prices"
As if people don't eat, heat & cool their homes, & put gas in their cars.
There are methodological reasons it makes sense for the Fed to exclude them for the Fed's purposes. But for the rest of us, especially any Journalist trying to report out "why do people feel bad about the economy," they ought to consider food and energy. Agreed!
I agree with the pushback here from the journalistic POV, but for the narrow predictive question in the article core PCE produces predictions of sentiment than headline PCE. Interesting to think about why that might be! But yes, when talking about prices, we should include the numbers that impact people at home
1. Prices are always at an all-time high with the exception of times of deflation (the gift that keeps on taking). It’s just the math.
2. Aggregates are a useful tool, but only a tool and typically a cruel one. In particular, they are blind to a K-shaped economy where folks are increasingly on the wrong leg of the K.
3. Sure, the numbers matter on their own. Twelve bucks for a plate of migas is jarring. But, often, it’s the sheer number of expenses — and the sheer number that keep becoming increasingly non-trivial — that get you. And when you’re *still* on the wrong leg of the K?
4. It’s the distribution, stupid. We’re forty-five years into the dismantlement (though not without some significant pauses) of the Great American Postwar Middle Class. The evil brilliance of the right has been to parlay the anger this has justly generated into support for the interests of the most rapacious among us. You know, the immigrant wants your cookie and all.
So maybe we can get to the next chapter without an intervening existential crisis?
Please?!?
And, if I may — something that gets to me at a meta-level across economics, politics, system performance and, hell, probably even sports, is that we (self included) constantly take valuable multidimensional metrics and collapse them down to scalars, getting rid of large amounts of their meaning in the process.
Historically, that was a necessity; we were bound to the printed page. But we can get past that and learn to present and learn to read graphical depictions of data in a more useful way.
(…sits back down…)
Cheers.
I hadn't thought about the impact of number of expenses. I'll see about that