by Oren Cass (@oren_cass)
Since its publication on February 20 by the Manhattan Institute, “The Cost-of-Thriving Index” has received widespread attention.
Christopher Ingraham, data reporter for the Washington Post, wrote that he had “never seen the discrepancy between what economic statistics say and how American families feel explained so clearly.”
Ryan Bourne, R. Evan Scharf Chair for the Public Understanding of Economics at the Cato Institute, was “convinced [COTI] captures one reason why the middle classes … feel jittery.”
Mitt Romney’s 2012 campaign strategist thought it “makes interesting and important points,” while President Obama’s speechwriter called it a “really interesting look at what is hiding beneath rosier economic statistics.” Will Wilkinson, an outspoken Trump critic at the Niskanen Center, wrote that, “a couple quibbles aside, this is great stuff,” while Chris Buskirk, an outspoken Trump supporter and publisher of American Greatness, called it “excellent.”
A small but vocal cadre of critics was less impressed and unleashed a broadside of variable quality.
“Captain Tako,” an anonymous Twitter user whose avatar is a frowning green blob with an eye patch, posted his own analysis that purported to correct the initial report and thereby refute its conclusions. The “analysis” was embarrassingly sloppy, among other things suggesting that the cost of a family health plan was no higher in 2015 than in 1995. It also became apparent that the anonymous analyst did not understand the terminology used to describe whether data were adjusted for inflation. Still, his work was promoted by researchers affiliated with the American Enterprise Institute, the Cato Institute, the Mercatus Center, R Street, the Joint Economic Committee, and the Committee for a Responsible Federal Budget. This was comical, but also indicated a distressing inclination to advance a specific narrative rather than pursue an open inquiry.
In a few cases, commentary progressed beyond the 280-character format to blog posts and op-eds — for instance, from Robert Verbruggen, Matt Yglesias, Mark Perry, and Michael Strain. Taking their criticisms as representative, I try to address the concerns that they and others have raised in three parts.
First, I offer some general comments on the contours of the debate, which I think provide useful context for both COTI’s value and the criticism it has triggered.
Second, I clarify why COTI was created and how I think it should be interpreted. This section draws heavily on quotations from the initial report, which while surely imperfect and less clear than it could have been, did make considerable effort to carefully define its specific claims. Just understanding what the COTI report says can hopefully go a long way toward resolving the objections that have been raised.
Third, I review the specific methodological objections raised and alternatives proposed. In many cases, these objections are fair and worth considering — indeed, nearly every one is raised in the report itself. Most frustrating in the debate thus far has been a tendency by critics to present their objections as the uncovering of objective errors when they are in fact disagreement about choices that I made and explained. Criticism in this context requires not just restating COTI’s methodology but refuting my arguments in its defense, which critics generally have not done. A better understanding of COTI’s actual claims, discussed in Part II, will hopefully make clear why the methodological choices discussed in Part III are reasonable ones.
I conclude by offering some suggestions for next steps that might move the conversation forward rather than leave us talking past each other or chasing in circles.
Part I: The Contours of the Debate
A few brief observations may provide helpful context for interpreting the particular disputes:
1.1 The Downward Ratchet. Criticism has run only in the direction of seeking to reduce COTI’s cost estimates. This is peculiar, because the COTI methodology seeks explicitly to adopt a moderate set of assumptions. In some cases, as the report notes, an assumption might overstate costs. In others, as the report also notes, an assumption might understate them. For instance, a focus on health insurance premiums ignores out-of-pocket costs entirely. A focus on public college costs ignores the much higher costs of private college. The use of Fair Market Rent estimates ignores the shift from a 45th percentile benchmark to a 40th percentile benchmark. For all the pixels spilled thus far, none has interrogated any of these issues. The goal of critics seems less to improve COTI than to disprove it.
1.2 Shooting Before Aiming. Relatedly, much criticism has worked backward from the starting point that COTI must be wrong and require refutation to a search for justifications of that view, whatever their quality. This is obvious from the enthusiasm for Captain Tako described above, but also from criticism launched by prominent commentators after having “only skimmed” the report, without knowing where to find it, or while “trying to look away.” Such a strong predisposition toward accepting only measures that tell a particular story, among the people whose jobs are to explore with open minds the nation’s economic conditions, suggests strongly that their preferred measures will be incomplete and alternative approaches like COTI are needed.
1.3 A Lack Of Understanding. In passing, critics have advanced a number of claims that suggest they may have an insufficiently firm grasp on the challenges that households face in the modern economy. For instance, one should not offer the Chevy Spark as an adequate substitute for a Toyota Camry or Dodge Caravan as a family’s primary vehicle. Nor should one generalize that employer-sponsored insurance is a standard element of the compensation received by a wage-earner. Nor should one suggest that if a family is sending both parents into the workforce while saying they would prefer if one could stay home, the correct interpretation is that really they prefer to both be working. Regardless of COTI’s particular methodological merits, arguments like these have left me more convinced than ever that economists and policymakers need to spend much more time thinking in COTI-like terms about the real world.
Part II: What is COTI and what does it say?
2.1 COTI, Defined. The COTI report provides a specific, carefully defined metric:
In 1985, the COTI stood at 30 — it would require 30 weeks of the median weekly wage to afford a three-bedroom house at the 40th percentile of a local market’s prices, a family health-insurance premium, a semester of public college, and the operation of a vehicle. By 2018, the COTI had increased to 53 — a full-time job was insufficient to afford these items, let alone the others that a household needs. (p. 4)
2.2 What COTI Is Saying. The COTI report uses this metric to provide a specific answer to a specific question:
Here, the question is how well the typical male worker can provide for a family. This report shows that his ability to do so has degraded dramatically. A generation ago, he could be confident in his ability to provide for his family not only the basics of food, clothing, and shelter but also the middle-class essentials of a comfortable house, a car, health care, and education. Now he cannot. (p. 6)
Because of several misconceptions, many criticisms of COTI have sailed far wide of the mark.
MISCONCEPTION #1: COTI claims to “overturn[] decades of research by expert economists” or show an “ongoing deterioration in the standard of living.”
2.4 COTI Does Not Reject Existing Statistics. The COTI report emphasizes that conventional statistics are valid and useful. It underscores that material living standards clearly have risen. It offers COTI not as a replacement or correction but as a supplement that can depict a dimension of economic reality that other statistics make no claim to measure. For instance:
Alongside formal “inflation” and a technical “cost-of-living” measure that aims to hold constant absolute material consumption, an accurate depiction of economic trends would track a more dynamic basket of the things that a family would need to retain the financial security and social engagement typical of a flourishing middle class. Call it the “cost of thriving.” (p. 6)
Products improve in countless ways that inflation ignores but people recognize. Consumers might benefit from better and more varied foods in the supermarket, brought to them by safer and more reliable supply chains. They might be able to communicate instantly with friends and family around the world using FaceTime, when in the past it would be a treat to see those relations once every few years. The material gains, viewed through this lens, seem endless; how could anyone say, holding money constant, that he would rather live in an earlier generation? That may all be true, but it remains incomplete. (p. 12)
The COTI tells only one part of the story of the economy’s evolution in recent decades — there is much it ignores, and many of its assumptions run counter to ones useful in answering questions about, for instance, monetary policy or technological innovation. The same, however, can be said for standard measures of macroeconomic inflation and the adjustments that they suggest to nominal wages, as well as for qualitative assessments of material living standards and technological progress. If we want to place price levels in historical context, inflation indexes are the closest approximation. If we want to know how much consumer surplus our households are capturing, evaluations of product quality can help. But if we want to understand what has happened to people’s ability to provide for their families, the COTI provides the more reliable guide. (p. 19)
MISCONCEPTION #2: COTI claims to show that families cannot make ends meet in the modern economy, i.e., that “the four essentials to a good life are out of reach” or that it is not “possible to finance a middle-class life on one income.”
2.5 COTI’s Focus Is The Structural Shift In How Families Make Ends Meet. The report emphasizes that families obviously still are finding ways to make ends meet. As discussed below (3.9), short of widespread bankruptcy, families are by definition always finding ways to make ends meet. But how they do that matters, to them and to society. COTI shows that the cost of a particularly important set of things a family would likely want to be able to buy has risen much faster than the wage with which they might seek to buy it, eventually overtaking the wage entirely. This leaves families’ ability to make ends meet contingent on many factors and the adoption of particular strategies, when in the past they had much greater agency, security, and flexibility.
A generation ago, he could be confident in his ability to provide for his family not only the basics of food, clothing, and shelter but also the middle-class essentials of a comfortable house, a car, health care, and education. Now he cannot. Public programs may provide those things for him, a second earner may work as well, his family may do without, although his television may be larger than ever. The implications of each is surely worth pondering. But the fact that he can no longer provide middle-class security to a family is an unavoidable economic reality of the modern era. (p. 6)
As the COTI basket has become unaffordable, families have found workarounds, such as having more household members work more hours, making do without, borrowing, and relying on government support. Each of these workarounds comes with its own costs, undermines the stability of families and the rationale for their formation, and creates high levels of stress and uncertainty. While some may celebrate the increased role played by government in filling these gaps, the continued drift in this direction threatens to strip from the middle class the pride of earned success and self-sufficiency; it also serves to reorient society toward dependence on government support. (p. 19)
MISCONCEPTION #3: COTI is “expenditure data” or a “measure of consumer spending,” showing what “households choose to buy.”
2.6 COTI Looks At Cost, Not Spending. COTI considers the costs of things that families would likely want to purchase and, historically, could comfortably afford to purchase. This point, that COTI is not attempting to measure what households actually spend, is highlighted as a contrast with traditional measures and a contribution of its approach:
[COTI] tracks the cost of a basket of major items that a family of four would likely seek to buy. A comparison over time between the cost of that basket and a median weekly wage indicates whether economic trends are easing or compounding the challenge of making ends meet. (p. 4)
At first glance, some of the items chosen for the COTI basket will puzzle modern analysts: Do we really expect middle-class households to afford all the costs associated with home ownership, a car, a college education, and a health plan? This reflects the bias built in to standard inflation analyses, which work from what households are buying at the current moment in time. In the past, as the COTI demonstrates, it was perfectly plausible to afford all these things. And all are things that a typical middle-class family might want to have confidence that it can afford. (p. 15)
2.7 The Question Remains: Is COTI Useful? Accurately describing COTI is not a defense per se. It shows that many of the major criticisms are irrelevant. And, as explained below (Part III), it shows that many of the methodological objections are misguided. But one might still make the case that what COTI actually shows is unimportant or uninteresting. “Who cares,” the argument would go, “if a worker can afford to buy all these things himself? At the end of the day, most families have what they need. And while some things have become more expensive, others have become cheaper. So COTI tells us nothing. It’s a useless benchmark.”
2.8 The Argument Is Ultimately A Normative One. COTI presents easily verifiable data about the cost of various goods and services and the wages of the median male earner. As the report states, “The normative question of whether we should care if the typical man can confidently provide for his family is beyond the scope of this paper. The information presented here establishes only the descriptive reality that once he could and now he cannot.” (p. 17) Of course, the publication of the index implies an endorsement of the normative claim that we should care — that there was something good about an economy in which a worker earning a median wage could comfortably and independently pay the full cost of health insurance, rent a near-median three-bedroom house, operate a vehicle, and save while raising his children to put them through a four-year public college, and it is bad that this is no longer the case. Rejecting COTI as useless or its methodology as wrongheaded requires arguing that we should be indifferent to this change, or even that it is a change for the better . This is not a direction critics have been eager to travel.
2.9 What You Would Have To Believe. It is possible that critics have not tackled the normative question because it did not occur to them, but more likely the problem is that they share the normative beliefs from which COTI works. To reject the normative claim, one would have to minimize the importance of values like financial stability and security, self-sufficiency, and personal responsibility. For instance, before rejecting COTI we should ask:
· Do we care whether people are able to insure their own families or have to hope either that their employers insure them or else that they qualify for a government program?
· Do we prefer that people feel an expectation to save for their children’s future and confidence they can do so or that they feel like chumps when they try because the system operates to take every dollar saved but provides for anyone who has saved nothing?
· Should forming a stable family and becoming a productive contributor to society be the pathway to self-sufficiency and a middle-class life, or should people find themselves in equally precarious financial positions whether they do those things or not?
Precisely because we are not indifferent between such worlds, we need to consider a measure like COTI. We should want to know if a significant structural shift in our economy has occurred and we should be concerned that one has.
Part III: Methodology and Alternatives
Based in part on the misconceptions addressed above (Part II), critics have raised several specific issues with the methodology used for its calculation, focused primarily on health care and education but in some cases highlighting transportation as well.
These criticisms fail in two important respects: First, they treat as errors a set of choices that are described and defended in the report, failing to engage with or refute the justifications provided. Second, they fail to evaluate the methodology through the lens of COTI’s purpose and claims, in effect arguing some other measure would better serve some other purpose they prefer. For each issue, I present the explanation provided in the COTI report, the criticisms, and then a reply.
Health Care
From the report: The most reliable historical data on health-insurance premiums are available for the employer-sponsored market, which should provide a reasonable proxy for all comprehensive private health-insurance policies. … Individuals covered by employer-sponsored plans do not bear their full cost, but the share of the nonelderly population covered by employers has fallen to less than 60%. Among full-time workers with income up to 250% of the federal poverty line ($62,000 for a family of four), fewer than half have employer coverage. The share that will lack such employment at some point in time is substantially higher, and access to coverage in the event of job-switching is often a major concern. Even among those with coverage, the share of premium costs borne by the worker has been growing, and spending to meet deductibles has more than tripled in the past decade — the typical covered household incurred almost $8,000 in expenses in 2018. Despite skyrocketing costs, the share of compensation paid as employer contributions for pensions and insurance has barely increased, from 11% in 1985 to 13% in 1993. The share in 2018 was slightly lower than in 1993. (p. 16)
Verbruggen: This is the total cost of the average employer-sponsored health plan, including the portion paid by the employer. As Stan Veuger of the American Enterprise Institute noted on Twitter soon after the report came out, it’s misleading to put this on a chart against “income” measured the way Cass measures it — via the Bureau of Labor Statistics’ reports on weekly earnings, which do not include benefits paid for by employers. The implication of the chart is that these health-care costs, like transportation or housing costs, come out of “income,” but employers are actually covering most of the bill. (For those without access to employer plans, mid-tier Obamacare coverage for a family of four making $55,000 costs less than $5,000 this year, which the report deliberately ignores, in part because “government provision is not a substitute for self-sufficiency.)”
Yglesias: The figure for health insurance, which the median American family does have, suffers from a more subtle but critically important flaw. As Stan Veuger of the American Enterprise Institute pointed out on Twitter, Cass compares wages (excluding fringe benefits like employer contributions to health insurance plans) to the total cost of employer-provided health insurance (including both what the employee pays and what the employer pays). Consequently, rising employer-side spending on workers’ health plans registers in this chart as health insurance becoming less affordable. Since the growing cost of health care is the main driving force behind this chart, the fact that it isn’t calculated correctly is a huge problem.
Perry: Here’s where Cass’s analysis gets really problematic. His estimates of annual health care expenditures are about 4 times higher than the BLS’s estimates of actual out-of-pocket expenditures for the average consumer unit because Cass curiously uses the total cost of the average employer-sponsored health care plan, including the portion paid by the employer. … Cass unrealistically seems to assume that Americans bear the full burden of their health care costs, when that’s not remotely true.
Strain: Cass includes the charges for the annual cost of employer-provided health insurance for a family. In 2018, this was $19,616. But he does not include employer contributions to premiums in his measure of worker income.
3.1 Health Care Criticisms Ignore COTI’s Claims. Frustratingly, these criticism fail even to acknowledge the report’s pre-emptive discussion of the issues raised, so none actually provides a response. Worse, none frames its criticism in the context of how COTI is defined and what it seeks to show. All argue for a measure of how much the typical family is likely to contribute to employer-sponsored coverage, ignoring the problem that many do not have it. All imply that a family’s ability to purchase their own insurance should be irrelevant, without ever making the case for that view.
3.2 The Alternative Health Care Measures Proposed Are Incoherent. Several people have attempted to “correct” the COTI analysis by reducing the insurance cost to reflect only that portion typically borne by the employee. But, that ratio of employer-to-employee spending applies only to families with employer-sponsored plans. Using this as a reflection of the costs that families face implies that everyone receives such employer contributions, which is false. Whereas COTI makes the explicit and relevant choice to look at the full cost of buying health insurance, these alternatives present the incoherent hybrid of an average of what some households but not others spend. One might try to correct this by weighting the ratio for the share of households with employer coverage. (So, for instance, if 50% of families have employer coverage and employers pick up 70% of their insurance costs, then employers cover 35% of cost overall.) This would be even more nonsensical, presenting a hypothetical aggregate that applies to no actual households. Further, a focus on employer coverage would need to incorporate out-of-pocket costs to account for the increased cost-sharing that families with employer plans have faced — an adjustment critics have been uninterested in considering.
3.3 A Problem With Averages. The employer-coverage issue underscores a challenge of measuring household well-being solely with averages. Here, we have a very irregular distribution of costs: a family either has employer coverage or it does not (of course, there is also variability within employer coverage). Focusing on the average — the cost a family faces if it kind-of-sort-of has employer coverage — describes no one’s reality. Nor would choosing a “typical” family that is only one kind or only the other, when in fact many if not most families are like to find themselves in each category at different times. An economistic analysis that tries to find the average of the average that shows the “average person” is doing all right provides one useful perspective and its trend-line is of particular relevance, but it hardly captures economic reality in full. Looking at what health insurance costs and noting that many workers who could in the past have afforded it may now be unable to, is useful as well and deserves a place in the conversation.
Education
From the report: Two children pursuing four-year degrees would require a combined 16 semesters of college, so a household preparing for those costs would need to save roughly one semester’s worth of cost per year before the children reached college age. … The one-semester estimate may overstate costs in some respects — for instance, a family would likely have 20 or more years between the birth of a first child and the college graduation of a second. And in practice, many children do not ultimately attend college (though a small and, it seems likely in recent decades, declining share has chosen from a young age not to consider that path). But it also understates costs by considering only public college costs; private college costs are more than twice as high. Note also that the cost of public college tuition already incorporates the substantial public subsidy provided by the state government. Inescapable in any discussion of college costs is the question of whether sending all kids to college, especially where the return on investment might be poor, makes sense. The answer to that question is no. But here the question is what costs a household faces and, so long as the nation’s education policy continues to advance a message of ‘college for all,’ saving for college will remain at the forefront of parents’ minds. (p. 16)
Verbruggen: Importantly, though, this number could be brought down significantly by the financial aid a student from a family in this income range would likely receive; it includes room and board, when many students at four-year public schools live off campus, quite frequently with their parents; and it assumes the kids won’t be earning any money of their own, winning scholarships, or taking out a single dollar in loans. Additionally, most American kids don’t actually go to four-year colleges — 15 percent of kids at public high schools don’t even graduate that on time, and of kids who do graduate from high school, about a third don’t enroll in college right away and another quarter go to two-year schools, which tend to be much less expensive. Some kids choose to go to more expensive private schools, too, but on balance it seems fair to say a father supporting a family of four on $53,000 doesn’t need to put aside $10,000 every year to make it possible for his kids to continue their education after high school.
Yglesias: The college portion of the “cost of thriving” index is the most obviously problematic item on the list, simply because only 30 percent of young people enrolled in a four-year college — a share that’s basically at a record high. In other words, the median wage earner may not have kids who are attending a four-year college, so it’s not clear why you would make this comparison. Cass’s statistic also includes room and board along with tuition, which, again, many students don’t pay. But perhaps most egregiously, he’s looking at the sticker price of college, even though most students receive grants to partially defray the cost of college.
Perry: Cass’s estimates of consumer expenditures on Education are wildly inflated by a factor of about 7X because he considers the cost of college (including room and board) as spending on “Education” under the flawed assumption that the average American household or family faces those educational costs year after year when that’s not close to reality.
Strain: Cass charges the worker the cost (measured by the sticker price) of college tuition, but does not include financial aid as part of the family’s income.
3.4 The Financial Aid Omission. Critics make a fair point insofar as the report does not explicitly acknowledge that students receive financial aid. The text should have highlighted that and explained why it does not factor into COTI: First, because as noted above (2.9), the price discrimination practiced by colleges to vacuum up every dollar saved is itself a pernicious change in economic conditions that discourages saving. Second, because many students want to attend (far more expensive) private colleges and a family, when making ex ante decisions about saving, cannot know what a child will ultimately need.
3.5 Coping With Uncertain Costs. This question of how much a family should be able to save toward their children’s future, especially when they will only find out later how much they had needed to save, is a particularly challenging one. It returns us to the problem of averages described above (3.3). Some children will not attend college (though we should not, as the critics do, mistake not being in college at a moment in time with not having ever enrolled), some will earn full scholarships as football stars. Some will need room and board, some will live at home. Some will attend private four-year colleges that offer little in the way of grants or scholarships, some will attend public two-year colleges and get generous financial aid packages too. Just as with health insurance, where COTI uses a standard that seems desirable (“afford health insurance”) and was once achievable, here COTI uses “save for the full cost of a four-year public college” as a standard that seems desirable and was once achievable. Critics are welcome to say that’s not worth aspiring to, but they cannot deny that it once was reality.
3.6 On Education, Critics Again Make An Obviously Incorrect Choice. As with health insurance, critics inexplicably offer an alternative that makes sense only if the goal is to reduce costs artificially to the lowest possible estimate without consideration of the challenges households face or the aspirations they were better able to fulfill in the past. One could argue that, rather than use COTI’s public-four-year benchmark, the best measure would be a weighted average across all the scenarios described above (3.5). What one cannot argue is that the best estimate would be to still only look at public-four-year colleges, ignore the room and board costs frequently incurred, add in all of the financial aid one might receive, and ignore that many people attend (or at least want to attend) private schools whose costs are much, much higher.
Transportation
From the report: Most households have one or two vehicles and drive them between 10,000 and 20,000 miles per year. The COTI adopts the 15,000-mile figure used by BTS to report the total cost of ownership, holding this figure constant over time. A more refined analysis could account for changes in vehicle usage over time for particular household types and, especially, particular regions. Generally speaking, vehicle-miles-traveled (VMT) per capita rose 37% from 1985 to 2005 and then fell 2% from 2005 to 2018. (p. 17)
Yglesias: Cass also includes an odd argument about the way the government calculates inflation to drive his conclusion that car ownership has become less affordable when official data says it’s become more affordable. His basic argument is that the BLS deflates the price of cars on the grounds that the quality has improved, but that this doesn’t help you if you’re strapped for cash. The fact that a bespoke suit might be in some quality-adjusted sense a “good buy” doesn’t change the fact that it’s expensive, and the same, Cass thinks, applies to cars. But this isn’t right. The reason automakers don’t sell low-end new cars like an $18,000 minivan is that the low-end is now served perfectly well by pre-owned cars — because cars have gotten better. You can’t travel back in time to get a brand-new $18,000 Grand Caravan from 1996, but you can go on Carvana and buy a 2018 Dodge Caravan for $16,800.
3.7 COTI Measures Total Cost Of Ownership, Not Sticker Price. Yglesias simply misunderstands COTI here. The sticker price for a Dodge Caravan plays no role in the index. That example is provided as an illustration of how products that increase substantially in cost can appear in an inflation index as not increasing at all in price. COTI uses the estimate from the Department of Transportation for the annual total cost of ownership for a vehicle, which measures annual depreciation rather than sticker price as well as costs like gas, insurance, maintenance, and taxes.
Mark Perry’s Alternative
3.8 Perry Misunderstands COTI As A Spending Measure. Perry writes that “Cass creates his own manufactured measures of annual consumer expenditures on four major spending categories — housing, health care, transportation, and education — perhaps not knowing that the Bureau of Labor Statistics has been reporting those data since 1984.” I am aware that BLS reports those data, but did not use them because the purpose of COTI is not to measure annual consumer expenditures (Misconception #3). Laboring under this misconception, Perry goes on to describe COTI as a “fictional, contrived approach” that uses “an improvised, manufactured measure of consumer spending” and suggests instead “comparing average annual out-of-pocket consumer expenditures on major expenses by consumer unit to average annual income by consumer unit.”
3.9 Perry’s Alternative Is Perverse. As I noted on Twitter when Perry first floated this approach, proposing to measure whether it is becoming harder or easier to provide for a family by analyzing actual expenditure data is downright cruel. Consider what Perry is actually saying by replacing an analysis of what things cost with one of what a household spends. That spending is necessary capped (plus debt) at what the household earns. And they need a lot of things besides our “big four.” So, no surprise, their spending on the big four is limited. Understand how perverse this is: Looking at what a household actually spends and noting it fits within their earnings will always be true by definition. Using this as proof that costs have not risen faster than wages or that making ends meet hasn’t become harder is nonsensical. Try it in conversation —
Family: It keeps getting harder to manage these costs.
Economist: Well, let’s look at your budget. Have you cut back to avoid going into debt?
Family: Yes, that’s what’s made it a struggle.
Economist: Great job, I’m glad to see things are going well.
3.10 Perry Also Treats Non-Market Labor As Worthless. Perry’s other methodological adjustment is to switch from one worker’s earnings to household earnings, claiming credit for the added income of converting a stay-at-home parent into a wage earner. If a parent at home just sits around eating bon-bons, this might be a fair adjustment. But if the parent was working hard caring for children, managing the household, contributing to the community, etc., then that is all lost when the extra income is gained. So where does Perry’s analysis account for the lost utility from the non-market work? It doesn’t. The implicit assumption is that stay-at-home parents are not only sitting at home eating bon-bons, but also having no fun doing it. Presumably this is not what Perry thinks. Then he shouldn’t make a chart based on that premise.
3.11 Perry Declined The Opportunity To Address These Issues. Perry proceeded to post his analysis on the AEI blog four days after my comments on his approach, neither acknowledging nor addressing any of the issues. Based on the fact that households continue to spend in line with their income, he concluded that, “the average American household is flourishing and thriving today compared to 25 or 35 years ago and is actually able to cover a year of major expenses with two fewer weeks of income than in 1985.”
Michael Strain’s Alternative
Strain: Conservative populists commonly argue that typical male workers today have great difficulty supporting their families, whereas in the past they could do so much more easily. Is that true? Data from the Bureau of Labor Statistics show that from 1990 to 2019 the median male worker’s wages grew by 23%. The bottom 10% of male workers saw their wages increase by 36% over this period. Male workers in the 20th percentile — those who earn more than the bottom 19% but less than the top 80% — had 30% wage growth. All these figures are adjusted for inflation. That is, they account for increases in the prices of housing, health care, education and transportation — Cass’s four categories — but also for the prices of many other goods and services. The upshot: It is hard to argue that price increases are overwhelming the wage gains of male workers.
3.12 Strain Offers No Alternative, Only Restates Standard Inflation-Adjusted Measure. While Perry’s analysis neither addresses COTI’s concerns nor makes sense on its own merits, he at least attempted to bring new data to bear on the subject. Michael Strain rejects the suggestion that BLS inflation-adjusted data tells an incomplete story, and that another perspective could help, on the grounds that this disagrees with the BLS inflation-adjusted data. This is neither surprising nor insightful, and certainly not a refutation of the usefulness or relevance of COTI.
Conclusion: Making Progress
My goal in preparing this document was to consolidate and organize criticisms in one place for people following the debate, to show that most were unfounded, and to address those that were on point. I hope that, in doing so, I have fairly represented the arguments of the critics and engaged their strongest arguments directly (in part by quoting them at great length, a courtesy I hope they will repay in the future). I will try to continue doing so. This a living document that I will update as needed over time. If I have misrepresented your argument, please tell me so that I can make a correction.
The exercise has also, at least in my mind, clarified the core point of disagreement that warrants further debate. With some room for adjustment at the margin, the methodological choices embedded in COTI are the correct ones for the purpose that COTI was intended to serve. The interesting question is whether this purpose and its normative implications are valid. Is it useful to put the infinite variation in what particular households need and considerations of the many mechanisms that fill the gap between what things cost and what households can afford, and examine the structural question of how costs are moving relative to wages?
I say the answer is yes, which is why I created COTI. To recap, I say yes for the following reasons:
1. Not only making ends meet in a given year, but also achieving financial security and confidence in the ability to provide for a family is important. So while it is helpful to understand that when everything goes right a family with employer-sponsored insurance and kids attending public colleges with good financial aid packages would have seen wages keep pace with costs, it is also helpful to understand that a family that would in the past have been well-positioned to weather any number of less favorable circumstances has seen that capacity erode. Even if the average still looks OK, we should care if the variability and risk become higher and, especially over an extended period of time, if the share of families that would rightly worry about and face such challenges is increasing.
2. Our society and economy should facilitate and reward the achievement of self-sufficiency and the agency of controlling one’s own fate and planning for one’s future. When providing for one’s family becomes contingent on what others choose to offer or on receipt of government benefits, people will rightly sense a loss of ownership, efficacy, and dignity.
3. Personal responsibility, family commitment, and hard work should be the sine qua non of achieving middle-class comfort. If those things become neither necessary nor sufficient, our understanding of privileges and obligations, rights and duties, achievement and failure will shift for the worse.
For progressives, who tend to regard a government safety net and programs of redistribution as valid substitutes for earned success, a reluctance to accept any of these three premises and an insistence on emphasizing solely post-transfer consumption would make some sense. For conservatives, it makes none. Attending to the texture of our social fabric, the vitality of our institutions, and the influence of culture and incentives on behavior requires a willingness to listen to what a measure like COTI has to say.
