Skip to main content
Blog Post

How the Middle Class Is Faring: My Long-Read Q&A with Jeremy Horpedahl


August 1, 2023

Does the typical American family today enjoy better living standards compared to 1985? We may have bigger TVs in our living rooms and smartphones in our pockets, but a recent report from Washington, DC, think tank the American Compass suggests the cost of a thriving, middle-class lifestyle has risen over the past generation. To discuss what that report gets right and where it falls short, I brought Jeremy Horpedahl on Political Economy.

Jeremy is an associate professor of economics at the University of Central Arkansas. He’s also the co-author, along with AEI’s Scott Winship, of the recent report, “The Cost of Thriving Has Fallen: Correcting and Rejecting the American Compass Cost-of-Thriving Index.” That report argues a better methodology shows modest gains for the typical American family.

What follows is a lightly edited transcript of our conversation. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.

Pethokoukis: Before we dig into the data, I want to take a moment to talk about public perception about how America’s been doing over recent decades. There was a recent Pew Research Center survey, and 58 percent of Americans say the American public is worse off than it was 50 years ago. What do you think drives that perception? Do you think that perception’s driven by actual economic conditions on the ground or something else—psychology, I don’t know?

Horpedahl: That’s a great question. I’ve thought about that myself a lot. In the wake of our report that’s come out, a lot of people have speculated as to why, if we’re right on the numbers, why do so many people feel this way? I think part of it is that there’s something to what we’ll talk about with this Cost of Thriving Index (COTI), that people might feel that even though we have lots of new and better technologies and things we’re buying, that in some ways it might feel like to buy all the things that you wouldn’t have had to buy 40 years ago, it might be a challenge.

I do think, though, some of it is driven by misconceptions about the data. Right now, the average American isn’t following all this data too closely, but I think most people have seen a statistic that over some 40- or 50-year time period, if you adjust for the [Consumer Price Index], it looks like median wages are flat or roughly flat. I think people have seen data like that, and then when they connect that to how they might feel about purchasing things they might say, “Well, yes, that lines up with how I feel. We must be worse off, or no better off.”

I think there are lots of interesting psychological things we could delve into. That’s outside my area of expertise, but I think people do generally have what we call “pessimistic bias.” I think that’s something that psychologists have identified. Economists have picked up on that Bryan Caplan in his book uses that as one of the biases that might drive people’s political decision-making. So I think that exists. But I do believe that some of it is just people misconstruing the data. What we’re trying to do in this paper is say, here’s what we think is the best data to represent what a typical family looks like.

I’m not sure how that poll I mentioned breaks it down, but I wonder if people are thinking about the aggregate different than their own specific situation. I just have a hard time believing that someone—and I’m a little bit older than you, maybe a lot older than you—like myself, who grew up with no air conditioning and a much smaller house and a TV that got three channels and some pretty bland dinners, thinks that somehow they’re worse off. I wonder if a lot of people think, “Maybe I’m better off. But overall, we’re not really better off and politicians have great statistics that tell us that.”

I think you see this in a lot of polls where you ask people about their own situation versus the national situation, whether it’s on economic striving, whether it’s on crime, whether it’s about the environment, whether it’s about public schools: People seem to think that they’re generally doing okay, not 100 percent of people, but that somehow the national situation is deteriorating. So I think there is some of that. And when you ask people a more generic question, they’re going to lean on what they think the national situation looks like. I may not be quite as old as you, but I grew up in the ‘80s and I remember my parents telling me, “We installed air conditioning the summer you were born because it was so hot in our house in the summers that we thought, we now have a child, we’ve got to have air conditioning.” And they paid for it. But before that, they didn’t have it. They didn’t grow up with air conditioning. It’s a new luxury that is now affordable to middle-class families, and people are willing to pay for it. I think that fits into what we’re doing in the data.

You, along with AEI’s Scott Winship, took a look at something called the Cost of Thriving Index, which was developed by a guy named Oren Cass at a think tank called American Compass. The goal was to compare the ease of affording a middle-class lifestyle now versus back in the 1980s. The way he went about it is he looked at the cost of five goods and services available on the income of a typical male earner.

The way he illustrated how things have gotten tougher is that it used to take 40 weeks to pay for those goods and services back in 1985. Now it’s up to 62 weeks, which shows the cost of thriving has gotten higher for people. The goods were transportation, food, housing, healthcare and education. Five important goods. First of all, what do you think just about approaching this question of, are we better off than were 50 years ago, by looking at five key goods?

As you know and we’ll talk about, our paper is a critique of this index, but let me try to make the case for it. What Oren is trying to do in this report is to first of all reject what is normally done with inflation adjustments, which is that there’s been quality improvements. He says, if the goods you are buying—and he likes the example of, say, a Dodge Caravan—if that thing gets better features, better safety features, a V6 instead of a four-cylinder engine, which did happen with minivans, what happens with a standard inflation adjustment is they basically don’t count those as cost increases. They say that the vehicle is now X thousand dollars better so we’re going to say it costs the same amount. What Oren is saying in this index is that’s nice that it has those features, but someone still has to buy it. And you can’t now buy the minivan without airbags or whatever. It’s just not available. That quality adjustment is useful in one sense, but he wants to present this other perspective to say that if you can’t actually buy it, then those features aren’t worth anything to you. That’s what he’s trying to do.

He’s also using, as you mentioned, a one-male-earner household, essentially using the median earnings for a male 25 years or older. Another part of that is to say that maybe families can afford the minivan today, but that’s because they’re going to have, more than likely, two incomes. He’s saying perhaps that’s no longer a choice: If you want to buy these five things, then you’re going to have to have two earners. Because in the math, you mentioned 62 weeks, from one earner, that is not mathematically possible to do in a year. There are not 62 weeks in a year. Whereas 40 weeks, whether you think that’s too many or not enough, in 1985—at least there are 40 weeks in the year. Then you would have 12 weeks of your earnings to buy the other perhaps luxuries of life, if these are the necessities. So that’s his approach.

That’s the methodology.

That’s the methodology. We critique it in two ways, and we can get to it in a moment. But one is we just say, “Let’s take his framework as given, as a good framework, and just see if his numbers are right.” We find a number of ways in which we think the numbers are just wrong. Then we try to say, maybe that’s actually not the right perspective. The second part of the paper presents what we think is a better way to look at whether even a family with one earner is thriving or not. That’s the second critique. The subtitle is “correcting and rejecting.” The correcting part is to say, let’s stick with the framework and see if the numbers are right. And then the rejecting is to say, we don’t actually think it’s the right framework. We think that ours, a more traditional inflation-adjusted framework, is better. That’s what we try to do.

You’ve laid out a wonderful roadmap for us here. Let’s start with the correction, which is what gives them the 40 week versus 62 week conclusion. What would you tweak, assuming that is the approach you wanted to take, looking at those five goods and services?

You mentioned the five goods. He actually gets very specific on them. The transportation one is the cost of owning and operating a car. The education component is the cost of paying for tuition at a public university. The healthcare one is the cost of buying a health insurance package for your family. Those are the things he uses to get concretely at those five areas. We actually have corrections to basically all those areas, but one of the big ones is on healthcare.

This is one area where we say there’s a major error in what he’s doing. In both the 1985 and the 2022 numbers, he says we need to count the full cost of what it costs to buy a health insurance package for a family of four, or a family plan, which is something a little over $20,000. But that’s not how the median family buys their health insurance.

The median family is going to get it through their employer, in which case there will be a part that they pay and a part the employer pays. He’s subtracting all of that, in which case you are greatly overstating the cost of this to a typical family. Are there some families that in some cases might be paying all of that with their wages? Yes, there probably are. But they’re very small and most of those families are going to qualify for either [Affordable Care Act] subsidies or they might qualify for some other public assistance. We are not saying those are good or bad, we’re just saying we have to take those as given: What are people actually paying? And to say that someone with their market wages they’re being paid is then going to have to turn around and pay $20,000 in health insurance costs, this is just not the case. This is not reality. The reality is for most families, they’re going to be paying roughly half of it or so, depending on your employer share. You’re going to pay about half that. We are saying he is, in a sense, doubling the cost of that. And that’s one of the major costs. As most families know, that’s one of the big things your employer takes out of your paycheck. That’s also one of the things that has increased the most since 1985. By correcting that healthcare component, we find that alone is causing a big part of the increase. We make other changes to other parts.

We can go back to healthcare in a moment if we want, but just I’ll mention one more: the cost of college education. What he does is he looks at what’s the price you have to pay for a year of four-year public college. He amortizes that [because] you’re saving for it over a number of years. The major error that we think he’s overlooking is the fact that if you have kids that are of college age now, you know discounting of various sorts, merit and financial aid of various sorts, is a huge part of the component of what it costs to pay for college. If you just use the list price or the sticker price, today that vastly overstates what someone pays. In fact, there are calculations which we then use that many have done, which say, what’s actually the average price people pay to go to the school, after you account for all those discounts? They’re not accounting for student loans, but account for the discounts from the institution or from Pell Grants or anything else you get. Those existed in 1985, but they were much smaller part of the package. The discounting of various sorts is much more prevalent today. Then again, we’re thinking, what’s the reality for a family earning the median income or having one earner male income? They’re going to qualify for a variety of types of aid they’re going to get, so we need to take that into account.

Yes, the cost of college education has gone up. This is the business I’m in. I teach at a university. I know it’s gone up. But we also know, both for me being on this side of it now selling the service, but also buying it 15 years ago as a buyer, if you’re in this business you know that discounting is a major part of it. That’s what we call the sticker price. Whereas the sticker price of a car, today maybe that is about what you pay, the sticker price for a college education is not what most people are paying. At some universities, not a single person is even paying that. It’s a figment of their imagination, that price. So that’s another area where we think there’s a big correction.

With your corrections—remember the number he came up with was 62 weeks, so more than obviously the number of weeks of a single year—so using your tweaks, but sticking with his framework, what do you think is a better number?

Just making the corrections—and we make some corrections to the 1985 numbers as well—there’s one more we’ll talk about, which is taxes, but before you even bring in taxes, it gets it down to about 31.5 in 1985 and then about 42 weeks in 2022. So just with those simple corrections, there’s still an increase. There’s still something to what he’s saying, going from about 31.5 to 42 weeks. But 42 weeks is still possible. Just based on what we think are his numbers being incorrect, it’s gone down from 62 to 42 weeks. Certainly feasible, even if we might think it’s a little worse. That’s our first cut of saying, at a bare minimum, these corrections must be made.

If I accepted that framework, and if I also accepted your corrections, my response would be, okay, maybe it was overstated, but directionally it’s correct. Things have gotten tougher. So now we’ll go to phase two: why you think that this is a substandard, a faulty, a less preferable way of looking at cost of living. What is the flaw with the framework and what is preferable?

So before we get to correction two, there’s a correction one B, which is that he does not account for changes in federal taxes. As anyone earning wages knows, this is another major thing that comes out of your paycheck: the federal taxes you pay. Of course, since 1985, there have been a number of tax reforms, including one right after that in ‘86. But there have been a number of changes which affect a family earning that median male salary, what they would pay. And a family actually pays dramatically less in federal income taxes today. So there’s a one B correction that we do. That even shrinks the gap even more, to where it’s only a five-week increase if you compare 1985 to 2022, if you include a sixth thing people have to pay for in addition to his five, which would be taxes. That would be the sixth thing.

That seems like a weird thing to leave out. I know that sounds like a scholarly economist phrase, but to me it seems weird.

Yeah, especially to me, because a lot of my research is on taxes. So I immediately thought, this has to include this. In our conversation that Scott and he had at AEI recently, this is one thing where he said, “Actually that was a big oversight, and right now we’re working on incorporating that.” Once we include that correction for taxes, we then find that the gap shrinks even more. Whereas it would’ve taken 40 weeks to buy these six things in 1985, if you’re “buying” taxes, and then just 45 weeks in 2022. A much narrower gap. Still an increase, though, we might say.

So then we’ll get to the correction number two, which is our preferred way, still working within his framework, not doing a full inflation adjustment, but some inclusion of some quality adjustments. So not a full CPI adjustment, but what we think are some reasonable things that you would need to do. Again, thinking about that Dodge Caravan: As a car, the fact that it has heated seats, maybe that’s a luxury. But the fact that it has better airbags, the fact that it has anti-lock brakes: Some of these things you do have to include. We do try to make some of what we think are minimum quality adjustments. If the size of houses is bigger, that’s something that, maybe people might prefer a smaller house, but that’s definitely a quality improvement. With the minivan, he says it’s got heated seats, you don’t really need those. But if the house is bigger, you have to account for that. So we think you’ve got to make some of these quality adjustments staying within that framework.

No one goes to the doctor and says, “Can I have medicines that were only in use before 1985? I don’t want any of the new medicines. I want the old medicines.”

And I’m willing to pay whatever the price was in ‘85. If something is still under patent in 1985, I’ll pay that.

I want the original, non-generic price.

We say you have to make some quality adjustments. It doesn’t make sense not to. We do for some of the components, and what we do then is we adjust up some of the 1985 numbers, particularly for food and for healthcare. Food and healthcare are actually the big ones that we make in this second change, adjusting some of those up. And when we do that and include taxes in what we call our preferred version of COTI…

You’re trying to stick with his basic thesis, the basic framework. First you wanted to improve it with some better numbers, but now you’re taking a more substantive correction, especially regarding using quality adjustments. That’s the big addition to the framework.

We’ve thrown out a lot of numbers now, so read the paper and look at the table. The key here is once we do those corrections, we include the tax calculation, in that scenario, we find that actually the cost of thriving has gone down slightly from taking about 46–46.5 weeks in 1985 to buy this package of stuff down to about 45 weeks. So there’s not a big decrease, but we do find that for using his preferred method and his preferred type of earner, which is the median male, 25 and older, working full-time, he is narrowly focusing on that, there we do find that it actually is slightly easier to thrive under this framework once we make all the changes within this framework that we think we you need to do and is reasonable to do.

This is your preferred version of this framework. But fundamentally, do you buy this framework? Or do you think it’s better to use what the BLS uses, the Consumer Price Index, or the Fed’s favored inflation index, the PCE index? Do you think economists are going to abandon that and flock to this because this is this a better way of looking at it?

Yes, we do prefer that framework, and we try to make the case as to why we prefer that in the paper. I think that economists are probably not going to flock to what he’s doing, but I think he wants to have this be a part of the conversation for a number of reasons. One being that he would prefer more support for people that choose to have one earner and have the other spouse stay home and take care of the kids, that there would be support in the tax code to do that. But we do think that we want to have a better understanding of all this, but there’s one additional correction we do without even getting to conventional inflation adjustments: We just broadened the set of workers beyond what he included to include women and to include full-time workers between 16 and 25. He’s using 25 and older men, but if we include both women—not dual-income households, just the median female income—and we include those that are still full-time but that are under the age of 25, doing that we find an even bigger increase.

Here we stop talking about the weeks of buying it. Doing all that, we say there’s actually a 15 percent increase in someone’s real income even within this framework, just broadening it beyond that traditionally male over 25. At the end of the day, though, we think that you do have to account for quality adjustments as it is done by the PCE index, which I think is probably the best one. But even in the CPI, we do think that that is the right way to do it and that economists have not been all “wet” all these years by just doing it that way.

In the paper, we try to strongly make the case as to why that is. I’ve learned a lot by thinking about his framework. I actually am writing up a blog post on Dodge Caravans coming out soon, looking at how those have changed over time. Even taking the quality improvements into accounts, but not adjusting for them, it actually is easier to buy one now than in 1996, which was the peak of minivan mania when the Dodge Caravan was the car of the year—the first time a minivan was. It actually is cheaper today, even for the median male, his preferred income measure. But it has gotten me thinking about those sorts of things a lot more. But I think ultimately we do settle on, at the end of the paper, that we think the way economists have been doing it is basically correct. We think that is the right way to think about it.

Would it be fair to say that even using the Consumer Price Index or the Personal Consumption Expenditure index, PCE, that those underestimate inflation? That perhaps the adjustments aren’t enough, they’re not aggressive enough, and perhaps people are actually even doing better than what some of the optimists say?

Yes, especially over very long timeframes—I think economists will always realize this, but I’ll always say this—but over a 100-year timeframe, I think the CPI inflation adjustment dramatically understates how much growth there has been. Just for example, using conventional inflation adjustments and things like that the average person in China today has roughly the same GDP per capita as the average American 100 years ago. But those clearly aren’t the same. Clearly the person in China is in many ways better off than an American 100 years ago. But what that means is we’ve understated how much growth we’ve had in the US. I think over long periods of time, they actually do understate growth. I think that’s widely recognized, though.

It’s just hard to incorporate new technologies, new services into a price index. Economists have clever ways of trying to do it, but I think we know that these measures do overstate, even not on a 100-year timeframe, even on a 20- or 40-year timeframe, the CPI probably overstates inflation by about 1 percent a year. That adds up a lot over 20 to 40 years. We do think that, at a minimum, you should use these inflation adjustments. Oren’s index suggests the opposite: that these things are overstating the gains. Although in fairness to him, he would say his measure, the COTI, is not a measure of the standard of living. He would say the standard of living has improved: Medicine is better, cars are safer, etc. He would agree to that, but what he’s trying to say is, “Yes, but for some families, it’s harder to buy that. The standard of living is higher, but for a family with one earner at the median, it’s harder to buy that.” We think it’s a useful thing to engage, because we spend a lot of time thinking about it and working through the numbers. We don’t think it’s something that should just be dismissed. As we started out this conversation talking about, a lot of people do feel that way and express it in surveys or just in casual conversation. So we think it is worth addressing. But we are clearly much more optimistic than either Cass or the general public about how well a typical family is doing.

Jeremy, mission accomplished. You walked us through I think some difficult concepts and data. Great job. Thanks a lot for coming on the podcast.

Thanks so much, Jim.