Have prices have risen more quickly for people at the bottom of the income distribution than for those at the top? Lefty window-breakers wait impatiently while economists struggle to resolve this dispute.

Palko points us to this post by Mike Konczal pointing to this news article by Annie Lowrey reporting on research by Christopher Wimer, Sophie Collyer, and Xavier Jaravel finding that “prices have risen more quickly for people at the bottom of the income distribution than for those at the top.”

This new result counters an earlier study that got a bit of attention back in 2008, which I’ll get back to in a bit.

Before getting to the main topic of this post, which has nothing really to do with income inequality, let me talk about all the statistical and political challenges here.

First the political challenges. All the people mentioned above are coming from the left or center-left in the U.S. context, generally supporting economic redistribution, government regulation of the economy, and taking the side of labor in disputes with business. All these positions are relative, of course—I don’t think there are any Soviet-style communists in the room—but they have a general motivation to report that things are relatively worse for the poor. This would generally be the case, and it’s even more so during a Republican administration.

If a Democrat is president, the political motivations are more mixed: on one hand, people on the left will still want to emphasize the difficulties of being poor, but at the same time, people on the right might want to talk about rising inequality as a way to discredit the Democrats.

I don’t want to overplay this political point here. People on all sides of this discussion may well be addressing the data as honestly as they can, but we should still recognizing their political incentives.

The second political challenge is that I know two of the authors of this paper. I work with Chris Wimer and Sophie Collyer, and we’ve spent a lot of time talking about issues of measurement and poverty within households. I haven’t been involved in the particular work being discussed here—my contributions are with a survey of New York City families, and this appears to be a national study—but in any case I have this professional connection that you should be aware of.

The statistical challenge is that definitions of poverty depend, in large part, on survey responses and survey adjustments. I have no quick answers here, and I’ve not read this new study in detail. I just know that in economics, the data are not simply sitting there; they need to be constructed. And this can drive some of the differences in conclusions.

And there’s more. Going back to Konczal’s above-linked post and you’ll see a link to a post from Will Wilkinson in 2008 pointing to a Freakonomics blog post by Steven Levitt from that year. The link to Levitt’s post no longer seems to work, and the new link is missing the comment section, so I’ll point you to the Internet Archive version, where the rogue economist writes:

Inequality is growing in the United States. The data say so. Knowledgeable experts like Ben Bernanke say so. Ask just about any economist and they will agree. . . . According to two of my University of Chicago colleagues, Christian Broda and John Romalis, everyone is wrong.

Inequality has not grown over the last decade — at least not very much. What we think is a rise in inequality is merely an artifact of how we measure things.

As improbable as it may seem, I believe them.

Their argument could hardly be simpler. . . .

Let’s dissect this. The statement has to be an “improbable” surprise—“just about any economist” thinks the opposite—but yet it “could hardly be simpler.”

Levitt continues into a digression regarding “lefties” and “the sorts of people who break store windows in Davos,” which doesn’t seem to be so relevant, given that earlier he’d said that this new study was contradicting everyone, from Davos window-breakers to “just about any economist.”

Konczal also links to this 2008 post by sociologist Lane Kenworthy, who summarizes the argument of Broda and Romalis:

Income inequality has increased over time. But analysis of consumption data indicates that people with low incomes are more likely than those with high incomes to buy inexpensive, low-quality goods. In part because those goods increasingly are produced in China, their prices rose less between 1994 and 2005 than did the prices of goods the rich tend to consume. Hence the standard measure of inequality, which is based on income rather than consumption, greatly overstates the degree to which inequality increased. The incomes of the rich rose more than those of the poor, but because the cost of living increased more for the rich than for the poor, things more or less evened out.

The discussion then turns on the question of whether rich people are getting anything for these expensive purchases. Or, to put it another way, whether poor people are suffering for not being able to afford nice things. Kenworthy argues no.

Then again, I’ve collaborated with Kenworthy so maybe I’m more likely to hear out his arguments.

So, to summarize:

– In 2008 there was agreement, or tentative agreement, regarding the claim that the prices of things that poor people bought were going up more slowly than the prices of things that rich people bought. But there was disagreement about whether this should be taken to imply that consumption inequality was decreasing.

– As of 2019, it seems that the prices of things that poor people bought have been going up faster than the prices of things that rich people bought.

Is this a contradiction? I’m not sure. The time periods of the two studies differ: the 2008 study covers the 1994-2005 period, and the 2019 study covers the 2004-2018 period. So it’s possible that the poor people’s products had a relative decline in price for one decade, followed by a relative increase during the next. Also, the two studies are using different methods. It would be good if someone could apply the methods of the first study to the data of the second study, and vice-versa.

Putting this all together, you can see that the statistics and economics questions connect only tangentially to the political questions. Levitt was sharing an empirical claim, but it only took him a few paragraphs to start ranting about window-breaking leftists. Kenworthy accepted the empirical claim but refused to draw the same political conclusion. Now the empirical claim goes the other way, so the arguments about relevance can be spun in the opposite direction.

In saying all this, I’m not trying to imply that the economic questions are unimportant. I think it’s worth trying to measure these things carefully, even while interpretations can differ.

In this case, it’s a lot less effort for me to write a thousand words about the dispute, than to carefully read the two research articles and try to figure out exactly what’s going on. I skimmed through the Broda and Romalis article but then I got to Figure 4A which scared the hell out of me!

P.S. More here from Elena Botella.

82 thoughts on “Have prices have risen more quickly for people at the bottom of the income distribution than for those at the top? Lefty window-breakers wait impatiently while economists struggle to resolve this dispute.

  1. the prices of things that poor people bought have been going up faster than the prices of things that rich people bought.

    How does this measure inequality? You need to consider the rise in incomes as well…

    • It doesn’t measure inequality per se, but prices are an input into measuring inequality. If people in different income groups buy baskets of goods that have divergent price paths, one might want to consider using different price deflators to measure real income. For example, if everyone’s income growth was flat and the prices of goods the poor consumed grew at a slower rate than those the rich consumed, the use of different deflators will yield a higher income equality. Personally, I don’t think this is valid for various reasons, but it does seem to be a position that could be defended by a reasonable analyst.

      • It doesn’t measure inequality per se, but prices are an input into measuring inequality.

        After thinking about it more, I don’t see why you need to look at prices at all. Just look at net worth or whatever. Rich people will buy cheap or expensive stuff depending on their own preference while poor will be stuck with the cheap stuff… unless they make more money at which point they will go with the expensive stuff if they want.

        • The purpose of a dollar, or any currency unit, is to be an intermediate quantity in a calculation of a tradeoff… The currency unit is fully and completely arbitrary, you can never find out anything about the world by knowing *only* the dollar value.

          The same thing is true about the meter… you have no idea how far 1.2 meters is until you look at the definition of the meter, it just happens that this definition is time-invariant as far as we know because the speed of light is a constant and the cesium atom always emits the same frequency of light.

          The same is not true of nominal currency units through time. In the most famous cases it wasn’t even approximately true from one hour to the next…

          For example, in https://en.wikipedia.org/wiki/Hyperinflation_in_the_Weimar_Republic

          Of course, to measure inequality you can just compare the ratio of say an individual’s income to the median income… you don’t need prices of goods other than “labor” defined as whatever you sell to get income. But even that would be problematic measured over a year during the Weimar inflation… A Mark earned in Jan didn’t buy the same things in Jan as a Mark earned in Dec.

        • to measure inequality you can just compare the ratio of say an individual’s income to the median income

          Some people have relatively little income but a relatively high net worth (savings, etc). That is why I would not use income.

          But yea, I would just look at a histogram of networth to judge inequality. However, if you want to judge whether someone is poor/wealthy then you need to take into account the prices of at least basic necessities.

        • Sure, using net worth would be better for comparing wealth. You still want to look at a dimensionless ratio. So (Net worth / Some Meaningful Standard) like for example (Net Worth / Annual average household cost of a fixed amount of food and rent for a fixed property size/quality) and then look at that histogram… You can compare that histogram to the histogram of the same thing in China or in Germany or whatever, without any concern about currency units and major quality variation and soforth.

        • Suppose as your metric of inequality we choose the difference between the 99%tile income and the median income… we calculate this using dollars.

          In 1929 pretend this is $400 and in 2009 pretend it’s $98000

          obviously there’s way more inequality now right? more than 100 times as much!…

          NO that’s not right!

          now, we take the 1929 values and divide them by the median income value in 1929…

          then we take the 2009 values and divide them by the median income value in 2009…

          now we calculate the differences, in 1929 it’s maybe 4.2 and in 2009 it’s 3.9 (I’m making this up entirely just for illustration)

          this means that in 1929 the 99th percentile was 4.2 times as much as the median, and in 2009 it was only 3.9 times as much as the median… these are obviously comparable quantities, because they measure the same aspect of society in a way that doesn’t depend on arbitrary units of measurement, dollars vs cents vs swiss francs vs chinese yen for example, and obviously in 2009 the difference is only 93% as large as the 1929 statistic… inequality by this measure went down 7% not up 100 fold…

        • I wasn’t planning on looking at the absolute values I guess. I would be interested in the relative proportion of the pop in each percentile bin.

          Also, by normalizing the x-axis you would miss out on info like the net worth to be in 99% bin was 100x that for 10% bin in 1913, but is 1000x today (made up numbers).

    • Economists still haven’t gotten it drilled into their heads freshman year that *everything* should be dimensionless ratios…

      That being said, if you’re already segmenting on income, which is what people usually mean by “poor” and “rich” (as opposed to net worth), then you can hold income constant and look at the price of goods that the income group buys, so the dimensionless ratio is kind of half-baked in.

      This is problematic in part because it can’t handle the flow of people through the income group. For example, generally college students are very low income, then they graduate and get a full time job, and become medium or even high income… So the experience of the individual is different from the experience of the income group they’re in at any given time.

      All this being said, especially because I’ve been bashed by certain commenters for always complaining about social science in the last few days on this blog, I’d like to say that at the very least, this is a paper that tries to answer an important question and it doesn’t just take the “canonical” measurements as fact (like CPI and GDP and median household wage and etc). Economics is a very important field. Well constructed insights have the potential to dramatically improve quality of life for people who are suffering. We should encourage good work.

      • I’m a little confused by your response on dimensionless units

        To the original post, the reason why we economists care about prices in this context is that they are trying to create estimates of real income, in other words we adjust the nominal income for changes in the price level. This facilitates comparison over time, particularly due to inflation. If prices for one group have risen faster than another, then that means that their real incomes have not grown as fast, holding nominal income constant.

        In one sense, real income can be considered dimensionless because you are dividing dollars by dollars. However, usually they are presented as being in a particular year’s dollars. So for instance, real GDP right now is reported in chained 2015 dollars (where chaining is the methodology for converted nominal income to real income).

        I probably agree with the bulk of the rest of your post. Studies like this require a lot to get right

        • Yes, Economists like to take todays dollars, and then multiply by the dimensionless ratio of todays price of a particular basket to the price of that basket at some past point in time….

          CostToday * (CPI in 2009) / (CPI Today)

          I suggest alternatively to just publish the dimensionless ratio:

          (Price of a thing of interest) / (Price of a relevant reference quantity)

          For example “how much does a hospitalization cost?” could be answered by :

          (Average Price of Hospitalization Today) / (Average household after tax income today)

          or

          (Average Price of Hospitalization Today) / (2000 working hours * median hourly wage today)

          or

          (Average Price of Hospitalization Today) / (Median Household Net Worth Today)

          etc… each one answers a slightly different question, but each one is immediately meaningful, it measures a tradeoff, whereas nominal dollars are ultimately meaningless, it really measures only half of a tradeoff, or some implicit tradeoff left to you to calculate… and as time goes on the intuitive ability to calculate because we “all know the price of tea (in china) today” is lost, since I have no idea how much a sandwich cost my grandfather at age 12…

          Part of the problem is the idea as Steve says below that you could somehow just define “the basket” by which all things should be compared… ie the CPI is basically opaque, you have to trust that the tradeoffs made by the CPI people are meaningful to you.

        • Suppose for example, I tell you that in Rogainistan the cost of a hospitalization is 15000 hairs… Is this a lot? I literally know *NOTHING* of use after hearing this fact.

          Now suppose I tell you that in Rogainistan the cost of a hospitalization is 15% of median household income…

          I immediately know that this is substantial but not ruinous. Furthermore if I find out that in Chickmanistan the cost of a hospitalization is 190000 feathers. is this more or less than in Rogainistan? But if I find out that in Chickmanistan the cost of a hospitalization is 36% of median household income… I can immediately know that it’s a much bigger deal in this country to be hospitalized than it is in Rogainistan, more than twice as expensive for a typical person.

        • Also notice how it is still totally useless to learn that a hospitalization today in Rogainistan is 13,000 in 2009 constant value hairs. in order to utilize this information you need to take your background knowledge of the prices of goods measured in hairs in 2009 and create the ratio yourself… so you can say “in 2009 I made 94,000 hairs per year… so 13,000 hairs back then would have been about 14% of my income… so that would have been a fair amount…”

          but by itself the knowledge “13,000 in 2009 constant hairs” is useless, and needlessly complicates comparisons across time and across countries and across goods.

        • Daniel said,
          “Now suppose I tell you that in Rogainistan the cost of a hospitalization is 15% of median household income…

          I immediately know that this is substantial but not ruinous. ”

          I don’t think it’s that simple . For example, if a family is already below the median household income, paying a hospital bill of 15% of median family income might put them below the income needed to proved the basics for healthy living — so they would need to get a loan, which would decrease their already marginal income for several years (and probably at high interest rates).

        • Sure, but you can at least see that in Chickmanistan things are a lot worse for a family at that same fraction of median income, and this without having to try to calculate purchasing power parity and convert everything to a common currency, which is all just a way of contorting yourself so you don’t have to think in dimensionless form… 15% is substantially smaller than 36%, hospitalization is cheaper in Rogainistan because you have to sell less than half the labor to acquire it.

    • Roberte:

      Rambling and disjointed . . . that’s what we do here! That’s what blogging is all about.

      Regarding your question: My point is slightly different. I don’t think that you can clearly separate “political ideology” with “sober economic analysis.” There are just too many ways of doing a sober economic analysis.

  2. We cannot adjust for inflation without a choice of a basket of goods. Theoretically, the basket represents an amount of utility, but that is fiction. So, the choice of the actual goods and services in the basket determines the inflation rate. That is not so problematic when we are figuring out the inflation rate from month to month, and the basket remains approximately unchanged, but, as I have said before, once the basket changes, the assumption that the basket represents an amount of utility is essential, it is no longer a useful fiction. If there is no real quantity with cardinality, then the inflation measure is meaningless. I cannot compare two completely different baskets of goods. Rich people and poor people don’t consume the same things. This is all meaningless, and not useful to get at the question that people want to know, which is did inequality rise. For that question, we need a general welfare function, which is impossible. So, it isn’t even a well-formed question. How about we start with questions that may not be quantitative but we can actually answer with real qualitative data. In what ways has life gotten harder for low income people? In what ways has it gotten better? Inequality is not susceptible to measurement, because we have to ask the next question which is “inequality of what?” We may not all agree on what should be equal and what shouldn’t, but at least if we were explicite, we would actually no what was at issue. Hiding behind inflation adjustments obscures the issues.

    • Steve,

      But what if we are comparing the same basket (approximately) for rich and poor people over time. That is, poor people 20 years ago and poor people today are consuming roughly the same basket, and the same can be said for rich people, in our analysis. Then this analysis seems at least as useful as any standard inflation exercise, which needs to deal with how to keep the basket of goods “constant” over time.

      • Matt said,
        “But what if we are comparing the same basket (approximately) for rich and poor people over time. That is, poor people 20 years ago and poor people today are consuming roughly the same basket, and the same can be said for rich people, in our analysis. Then this analysis seems at least as useful as any standard inflation exercise, which needs to deal with how to keep the basket of goods “constant” over time.”

        I’m not convinced you can always do this — the assumption that “poor people 20 years ago and poor people today are consuming roughly the same basket” needs to be examined carefully for a specific time period, location, and basket of goods.

        • Yes, and they aren’t consuming the same stuff. Just think about food, which is a good that seems easy, apples are apples, right. But, quality varies, and obviously it often varies with price. The Bureau of Labor Statistics tries to do regression analysis to adjust for quality changes likely exist and data is available. But, that is far from perfect and is ad hoc. Prices might be falling for the poor and quality falling even faster or the opposite, very hard (impossible?) to tease that out with the kind of data that economists are collecting.

        • I think the mistake is in thinking that the “officials” should “publish the one right answer” as opposed to say publishing data so people doing analysis can inform their own models.

          Given just the observed mean price across a sample of stores of a pound of apples/pears/broccoli/cauliflower/peaches/plums/nectarines/kale/collard greens/corn in each of 10 regions of the country, I’ll build my own index and do my own CPI.

          I don’t object to them doing some first pass, but I find it annoying that the more basic raw dataset is not really easily available. Most of what the BEA and etc publishes is rather “processed”, though I just did find the link (mentioned elsewhere on this page) that gives lots of “differently processed versions” at least.

    • I agree with a lot of this, but I disagree that we can’t measure inequality or inflation… we can’t measure inequality or inflation as a unique quantity, but we can measure various aspects of it, which would tell people that a multi-dimensional issue has changed in certain ways.

      Everyone likes car analogies, so I think your point is that you can’t say “what is the best car” because “best for what purpose?” can’t be answered uniquely.

      However, we can do things like measure the fuel consumption, the survivability of crashes at different speeds, the roughness of the ride over a fixed model of bumpy ground, the road noise, the acceleration, the braking, the stability of the car under wet conditions…

      We know that each of these has a “better” and a “worse”. No one wants to burn gas just for fun, or slide around and have long braking distances, or take forever to get up to highway speeds, or be needlessly killed or injured in minor collisions… But it’s also the case that different people value tradeoffs between these things differently.

      In my opinion, in asking and answering the question of how have things changed through time, we should be calculating straightforward dimensionless ratios between income and prices of different kinds of baskets of goods. Each basket should measure an aspect of life that is known to be important to people… We should also be looking at how much of each basket is consumed on average through time… And then we should be publishing these timeseries so that people can compare how different aspects of life have changed through time.

      • Also we should be looking at income, as well as income after taxes, and split up by educational attainment and industry…

        Some baskets of goods which would be useful to track prices of:

        Rent or owner’s-rent-equivalent for different square footages in different regions

        Cost of transport a fixed daily distance by various methods

        Cost of hospital treatment for non-life-threatening injury (falls, sports injuries, etc)

        Cost of medical treatment for chronic conditions like MS, diabetes, cystic fibrosis, etc

        Cost of life-threatening hospitalizations

        Cost of cancer treatments.

        Cost of 4 years of college education

        Cost of MS and PhD programs by an assortment of topics

        Cost of major appliance purchases (refrigerators, washers, etc, annual equivalent given observed typical lifespan)

        Annualized cost of housing maintenance

        Child care costs pre-school

        Child care costs school-age

        Child educational enrichment costs (sports, dance, music lessons, etc)

        Cost of food based on a fixed broad diet, at at least 3 levels of quality: healthy subsistence, moderate quality, and high quality as measured by a combination of physical properties of the diet and surveys on diet preferences.

        Cost of utilities: heating, cooling, electricity, gas, water… regionalized.

        Imagine if https://fred.stlouisfed.org had all of these indexes available, how useful it would be to aid understanding of the state of economic welfare of individuals or groups?

        I HATE it that for the most part we have CPI and a couple modifications to it, where they make all of these choices for us, the underlying basket changes without us really understanding the effect of the changes, and they leave in and out various things like energy costs on the basis primarily of trying to eliminate volatility in the index or whatever… what if volatility is what we are interested in? The different CPIs are primarily based on different demands based on different population sub-groups: all urban consumers, all urban consumers city weighted average, all urban consumers less food and energy… etc

        It’d be much better to be able to build your own index from 20 or 30 relevant time-series. The CPI is a hold-over from a 1940’s concept when “computing” meant hiring a room full of “girls” with adding machines, and it was imperative to pre-calculate as much as possible:

        https://www.nasa.gov/feature/when-the-computer-wore-a-skirt-langley-s-computers-1935-1970

        • For some reason that link points to the wrong table. Search for

          Table 2.3.4. Price Indexes for Personal Consumption Expenditures by Major Type of Product

          at the BEA website.

        • Thanks for that link, it’s pretty nice. But notice how they’re all arbitrarily scaled index values… In other words, if I want to know say how the purchase of a piano in 1980 compares to the purchase of a 4 door sedan in 1980… I’m SOL. These are dimensionless ratios arbitrarily scaled to themselves at a particular snapshot in time. They are totally in-comparable across goods.

          What’s really needed is a dimensionless ratio of the item scaled to say median household income at that time, or median hourly wage, or some other fixed standard. In fact for the purpose of forming indexes, you can actually just publish nominal dollars… Let someone build the dimensionless ratio of interest to them by dividing two dollar values.

          It’s nice to see that they’re trying though.

        • If I poke around in there perhaps I can find what I’d really want, which would either be not-seasonally adjusted nominal dollar values, or maybe not-seasonally adjusted “real” dollar values using a well publicized index, like say GDP deflator or CPI all goods all consumers or something. Then I could just un-deflate them back to nominal and construct new nominal indexes for use in the denominator of some ratio of interest.

        • Daniel wrote,

          “Some baskets of goods which would be useful to track prices of:

          Rent or owner’s-rent-equivalent for different square footages in different regions

          Cost of transport a fixed daily distance by various methods ..”

          At the very least, I’d replace “would be useful” with “might be useful”, since there might be confounding factors. Just looking at the first two examples Daniel gave:

          “Rent or owner’s-rent-equivalent for different square footages in different regions”: As housing becomes more expensive, people at low incomes often become more likely to live in housing with lower square footages.

          “Cost of transport a fixed daily distance by various methods”: As housing becomes more expensive, lower income people may be forced to live further from work to find affordable housing, producing an increased commute distance.

        • yes, those are true things. No one statistic can do it all. Having the denominator reference a fixed physical quantity transported across time makes the statistic interpretable in terms of what it’s *possible* to afford, rather than what people typically actually *choose* to afford… This is actually a problem with the CPI type calculation, what’s in the basket changes as people change what they buy. So if you are trying to figure out the inflation of a good whose demand has changed significantly through time… it’s impossible with the CPI.

          Buggy whips for example… I might actually want to know my income in terms of what it costs to buy 1 buggy whip today, not the 0.00001 buggy whips per person people typically buy on average today… by that measure my income is going to infinity, even if I can only half the number of buggy whips today as I could 10 years ago…

        • Daniel said, “I might actually want to know my income in terms of what it costs to buy 1 buggy whip today”

          And just why might you (or anyone else) want to know your income in terms of what it costs to buy 1 buggy whip today?!? Do you perhaps have some inside information that buggywhipmania is beginning? Or are you trying to start it? ;~)

      • We might be in agreement. Inflation and inequality are not unique quantity, agreed. But, I think where your idea breaks down is that we always have a basket of goods that does not change over time (or geography or social class, etc.). Sometimes sure, but often not. There is substitution in the basket (over time, place, social class, etc.) And, I don’t think that you can say that comparison of one basket can render a dimensionless comparision except by fiat. A basket of rental units is not the same everywhere. A square foot is not the same everywhere. People who want to live in New York City are different than people who want to live in Wyoming, and their utility functions for all sorts of other goods and services are built into the price of a square foot of a rental unit in both places. A person who loves living in New York and gets fired and has to move to Wyoming may be paying rent in Wyoming that delivers the same “utility” as the much higher rent in NYC did, while another person making the same switch just got a huge raise because of the much lower rent.

        • I don’t disagree with the utility aspect. I don’t think basket prices are supposed to give utility, they’re supposed to be inputs to calculating your own concept of utility. I also think aggregating at the level of the whole country is problematic, we should aggregate at about the level of a Census PUMA, that is around 100k people.

          Suppose you want to know whether you should take a job opportunity in Montana. They offer you 10% less money, but it’s doing a job you like to do. Prices in Montana on things you care about are not equal to prices where you live today… How should you evaluate whether you should take this job?

          One way would be to ask yourself the question “with this income in that location, what could I buy? and how can I compare it to what I can buy where I live today?”

          Not, what does the BEA think the basket should be, but what actually is the basket I want to buy? So it’s of relevance to you to calculate things like how much housing you can afford, what cost of transportation is like, what cost of internet connectivity is like, what cost of gym membership is like, what cost of sending your kids to private school is like, whatever you care about… maybe it’s having a restaurant grade kitchen to cook gourmet meals in or river rafting trips… whatever…

          With sufficiently broken down price histories on goods, people can make their own cost of living indexes. I think this is critical to enable better economic decision making. People are moving out of the SF bay area to places like Pittsburg and finding it a HUGE relief and improvement to their quality of life to take a 50% pay cut…

          With a sufficiently broad spectrum of localized price information, people can construct imperfect but at least much better measures of quality of life, and inequality in quality of life… we could have *SO* much better measurements than we do today.

        • I don’t disagree (I think), but what we began taking about was interpersonal comparisons (inequality). Of course, I can construct a welfare function for myself. I know my preferences, and if I’m wrong who’s going to know. I think that Amartya Sen’s work is relevant here. Having lots of different measures would be good, but assumptions have to be explicit and those include things like not everyone has the same ability to enjoy the same goods and services. I would love to see all sorts of measures as long as they were presented not as reporting objective facts but as one perspective on what matters for social welfare. But, I am suspicious that economists will ever present their measures that way (even if they do, someone else will present it as the correct way to measure social welfare.

        • With lots of measures put together you can then use surveys to construct a distribution over baskets of goods and construct Bayesian posteriors over the frequency distribution of a ratio whose denominator is uncertain… when most of the probability mass indicates wider frequency distribution you can be relatively sure of greater inequality… I don’t have much hope for econ adopting this view, but if they’d publish the right numbers I could bypass the gatekeepers ;-)

        • Steve said,
          “Having lots of different measures would be good, but assumptions have to be explicit and those include things like not everyone has the same ability to enjoy the same goods and services. I would love to see all sorts of measures as long as they were presented not as reporting objective facts but as one perspective on what matters for social welfare. But, I am suspicious that economists will ever present their measures that way (even if they do, someone else will present it as the correct way to measure social welfare.”

          +1

    • Steve said,
      “I cannot compare two completely different baskets of goods. Rich people and poor people don’t consume the same things. This is all meaningless, and not useful to get at the question that people want to know, which is did inequality rise. For that question, we need a general welfare function, which is impossible. So, it isn’t even a well-formed question. How about we start with questions that may not be quantitative but we can actually answer with real qualitative data. In what ways has life gotten harder for low income people? In what ways has it gotten better? Inequality is not susceptible to measurement, because we have to ask the next question which is “inequality of what?” We may not all agree on what should be equal and what shouldn’t, but at least if we were explicite, we would actually no what was at issue. Hiding behind inflation adjustments obscures the issues.”

      +1

  3. I have a hard time viewing this work as purely empirical. I think of it as synthetic. The end data series is constructed from combining the underlying data based on the choices of the researcher and often involves theoretical assumptions about the economy, which may or may not be influenced by personal politics. In my opinion a good work will show the outcome as invariant to different common theoretical assumptions. For example, in Jaravel 2019 (which is the academic work that the Wimer, Collyer and Jaravel piece is based on), uses different types of indices: Tornqvist, Laspeyres, etc. and he fines it doesn’t change his results significantly. That being said, it’s still not a purely empirical result, it is still based on theory and is contingent on that theory being a good representation of reality. It also involves hand matching of the data while stitching several data sets together (see pages 5-6 of the working paper below). There is room for the judgment of the researcher in the end result. That is not a point to denigrate the researcher (I like the use of the phrase “garden of forking paths” for example), but it at least influences how I think about the work.

    Another clearer example is the current dust up over the constructed tax progressivity measures of Zucam and Saez, where theoretical assumptions more clearly drive the results. That is not inherently problematic as long as you clearly state your assumptions and then defend them. They may or may not be accept, so be it. In fact, I am not really concerned if your assumptions are motivated by your politics as long as your assumptions are clear. This is a nice point that Michel De Vroey makes in his recent History of Macroeconomics where he critiques Friedman for his supposed lack of normative judgments in his positive conclusions but praises Lucas for recognizing, “the ideological dimension could not be dispensed with, it had to be tamed.” (De Vroey,2016, pg 200). In the present discussion, the problem is treating the resulting synthetic result that relies on theoretical assumptions as an empirical fact, especially in works facing audiences who do not know any better.

    I would be remiss if I didn’t mention that I think inflation heterogeneity is an important topic. You can debate whether people should be concerned with income inequality as much as they are, but the fact is they do care about it. Policy will depend on it. More specifically policy will depend on measures of real income, which depend on inflation. Better “measures” of it will allow for better policy.

    https://economics.harvard.edu/files/economics/files/ms27380.pdf

    • I like the point about how we should make our assumptions explicit and not give a gloss of some kind of pure objectivity. But I don’t think there *exists* a “purely empirical” anything.

      Even in something like Physics where you might assume that you could do something like measure the position of a ball and just say what its trajectory is… you can look deeply at this question and ask things like whether the speed of light is a constant through time, so that the definition of the meter does or does not correspond to different lengths at different times, or whether your chronograph accurately oscillates at a fixed interval or has some variation. This may seem like utter nonsense for normal video of a baseball pitch, but wouldn’t be at all nonsense for asking questions about the formation of galaxies, or for tracking bullets using ultra-high-speed cameras where the inter-frame time interval is both small and perhaps not perfectly stable.

      Interpreting the world relies on some theory, the question is just how much this theory is accepted, and in Econ there is more room for argument over theory, so the theoretical underpinnings should be somewhat explicit in the analysis… Not so much for the definition of length when describing baseball pitches with well calibrated video cameras.

    • James said,

      “In the present discussion, the problem is treating the resulting synthetic result that relies on theoretical assumptions as an empirical fact, especially in works facing audiences who do not know any better.”

      +many

  4. My goodness! Step back and think about the basic economic proposition before wasting so much effort with Category 5 noise.

    The “poor” market is the mass market – the market with the strongest competition and therefore the lowest margins. So any increase in labor or material costs will have a significant impact on profit and therefore be transferred to consumers quickly.

    The “rich” market is less competitive because rich people aren’t as tight with their money. So there’s typically a higher margin in that market and more flexibility to absorb rises in labor / materials costs. OTOH, rich people don’t have to buy the coveted manganese-yittrium alloy screw drivers with “The Wealth of Nations” inscribed in microscopic print on the Polished Lunar Basalt handles. If the price of MYAPLB screwdrivers gets to high, they can downgrade to plain old gold. So, because rich people can easily downgrade to less expensive products, sellers have less flexibility to raise prices.

    So we should expect inflation and rising costs to hit the mass – “poor” – market first. So what? Poor people have a shitty deal. Who knew.

      • Hi Daniel!

        Sure. And that’s reasonable. I didn’t mean to imply people shouldn’t try. Apologies. But when noise wins, acknowledge and seek a new path.

        I’m irked today like Phil was the other day. I saw a headline in the paper this morning:

        X ***MAY*** cause Y. (emphasis mine)

        WTH? When I see “may” in this context I conclude – usually correctly – that the signal is less than the noise, but even though the value of the signal is meaningless – the authors or headline writers decided that, since it’s pointing in our direction, may as well make the most of it.

        “may” is a weasel word that people use when they can’t use something meaningful. It’s technically equivalent with “may not” but allows the writer to allude to a positive association that doesn’t actually exist.

        So I’m irked about noise mining today.

        • Life Pro Tip:

          Whenever I see a “May Cause” headline, I mentally rewrite it to the logically equivalent “May or May Not Cause.” (“Eating Carrots May or May Not Cause Ankle Cancer.”) I am then instantly seized by a profound indifference to the article and feel an overwhelming urge to not read it, thus saving myself valuable time.

  5. It is not enough to normalize for inflation, or even to take into account variation of exposure to inflation. One has to speak of prices relative to total income (in some appropriate way), rather than absolute prices, or this discussion is not really addressing inequality. The marginal cost to me of $100 is much greater than the marginal cost of $100 to super rich man X, and this difference likely swamps the differences coming from variations in our exposure to inflation. Put another way, it is completely reasonable to assume that the utility to a poor person of $100 is greater than the utility to a rich person of $100 and there are straightforward ways to incorporate this into the model, and it needs to be done, particularly when one is examining putative inequalities.

    • Dan F. says: “it is completely reasonable to assume that the utility to a poor person of $100 is greater than the utility to a rich person of $100 and there are straightforward ways to incorporate this into the model”

      It is also completely reasonable to assume that the utility to a poor person of $100 is less than the utility to a rich person. Many of the poor suffer disabilities. These disabilities could be physical, mental or even socio-economic/cultural. For example, the person in a wheel chair may need additional expenditures just to be able to use their apartment. So, the amount of money that would cover rent for the person in the wheel chair would not bring the same amount of utility as it would for a non-disabled person. So, just as we must assume the declining marginal utility of money or else utility theory doesn’t make sense (hint: it doesn’t make sense), we also have to acknowledge that individuals cannot all have the same marginal utility of money at any other level of income at all.

  6. This entire discussion seems bizarrely disconnected from what we talk about when we talk about inequality. Suppose society consists of two people. A makes $10,000 per year and B makes $1 billion. The poor guy pays $5000 per year in food and $5000 per year in housing. The rich guy pays $10,000 per year in food and has a house with a $100MM annual cost and saves the rest.

    Now the price of the rich guy’s house rises to $105MM per year. Even assuming the rich guy gets no more utility from his house, it seems bizarre to argue that income inequality has fallen because the rich guy saves a little less every year. But in fact, even if you justifies this decrease in inequality n some philosophical level, you’d have to be careful in practice, because in fact the rich guy might have gotten an extra $6MM per annum in housing services (ie more than he paid) in, say, a better view. In that case, even on a philosophical level inequality would have increased.

    Since there is no way to adjust for individualized utility of income, adjusting for some hypothetical market basket of goods which varies by income class alone seems just wrong. This is like those articles one sometimes sees that people who make $500,000 aren’t really rich because their private schools cost so much and their houses are so damn expensive. Raising the price of private schools does not reduce income inequality. If the private schools give a lot of need scholarships, it *does* reduce interfamilial utility differences to some extent (to the extent that we can measure them) but that seems like a very different point. This policy, which is formally equivalent to taxing private schools and subsidizing attendance on a need adjusted basis, is what makes measuring inequality net of government subsidies so difficult.

  7. I’ve skimmed the 59 previous comments, and haven’t seen this mentioned, so I’ll add this.

    It’s entirely plausible that it’s the different time periods at work here. If we look at low-end goods, we might find they are made in a factory in Chicago in the 1950s, in rural Carolinas in the 1970s, in Japan and Taiwan in the 1980s, and then moving to China in the 1990s — following a search for cheap labor. So prices will rise less for low-end goods than high-end goods.

    But once manufacture has moved to the lowest cost place, and once that place starts seeing inflation in labor costs, the price of the low-end goods will rise faster than the high-end goods made in the same place across that span of time.

    To me, the whole issue here seems to be a misguided attempt to deny increasing income inequality. In some cases, it makes sense to look at differential baskets, because there’s little option. For example, there’s the argument that inflation is higher for older people in the US because health care prices are rising faster than other prices. People don’t have much option about many medical treatments. But it doesn’t make sense to look at differential baskets where the expenditures are optional — e.g. Jonathan (another one)’s example of private schools, or the price of yachts.

    • I’m not sure we ever find the absolute bottom on the search for the cheapest labor in the world, but the point is well taken.

      On your other point:
      “In some cases, it makes sense to look at differential baskets, because there’s little option.”

      But is there really little option? Why not just use the poor person’s market basket and be done with it, arguing that any improvements beyond the poor person’s market basket is a choice. This is just the flip side of the adage: “The law in its majesty forbids both the rich and poor from sleeping under bridges.” Money allows you a higher standard of living than a poor person. Inequality is quantified by the extra money you have beyond that of a poor person without any consideration of what you spend it on.

      • Jonathan (another one) says: “Why not just use the poor person’s market basket and be done with it, arguing that any improvements beyond the poor person’s market basket is a choice.”

        Which poor person’s basket? The actual goods and services people purchase changes over time. There have to be substitutions. People in the 80s, didn’t have smart phones. People ate different food. I need an index set and time t based on a basket of goods, and then I can compare it to the price of the same basket at time t2. When the goods and services in the basket have to be substituted because over time the available products and services change (in both their nature and quality), the decision of what counts and as a sufficiently similar basket is arbitrary. The researcher has total freedom to make that choice. There is no getting around that. You write that “Money allows you a higher standard of living than a poor person” But, the entire question being asked when we adjust for inflation is whether money has risen in value or fallen and by what amount. So, if I cannot make the adjustment, I cannot know whether one group has more money in terms of value and by how much more. Since I cannot make the adjustment in a non-arbitrary way, I cannot answer the question of whether inequality has risen or fallen. So, for example, I can make the assumption that smart phones bring more utility than flip phones because they have more functionality, then I won’t substitute smart phones for flip phones. That seems perfectly reasonable, but then I need to ask what are smart phones replacing in the old basket of goods. However, I make that decision will have wildly different effects on the direction and amount of inflation.

        • On the other hand, your assumption is that the purpose of the exercise is to determine whether “utility” has risen or fallen. If, instead, the purpose is to determine whether quantity of some “typical” set of goods that *can be* demanded has risen or fallen… then the exercise is non-problematic, we can stick to relatively long lived goods whose quality doesn’t vary over orders of magnitude. I mean, a $1000 computer in 1980 and $1000 computer is 2010 were VERY different things, but an apple in 1980 and an apple in 2010 weren’t *THAT* different, same for chicken breast, folgers coffee, brown rice, a cotton t-shirt, a gallon of gasoline, 100 sqft of cheap carpet, 12 glass tumblers, a 2 lb loaf of whole wheat bread, a prescription for doxycycline, 100 band-aids, a pair of OTC reading glasses, a bottle of laundry detergent, 12 cotton athletic socks, a plane ticket to travel from LA to NY economy class, a cup of plain drip coffee, a ticket to a monster truck rally, a dozen doughnuts, the cheapest way to get 12 roses for your anniversary, 4 folding chairs, 2 lbs of farmed salmon fillet, a gallon jug of Carlo Rossi Chablis, a large pack of toilet paper…

          Let’s not make the perfect be the enemy of the good.

        • In some sense, using the poor person’s basket, or even maybe the median person’s basket gives us a measure of *power* or *freedom* or *latitude*. If you have many multiples of cost of the median basket, you have the *power* to demand lots of things, whether you do or not.

          If on the other hand, you have only a moderate fraction of the median person’s basket, you have little power to make choices, you are highly constrained, there is less latitude, things fail in a brittle manner for you.

          As long as you don’t pretend you’re measuring utility, and instead simply measure this other quantity, that quantity is still a useful and valid measure of what *could be done* with your income.

        • “your assumption is that the purpose of the exercise is to determine whether “utility” has risen or fallen.” Not mine, that how the CPI is calculated. I am for doing it a different way. Your proposals sound fine. However, there is always going to be a point when the baskets just become too different for a relevant comparison. Quality varies, but also the entire lifestyle around the good or service vary. If you doubt it, you can order a expensive meal at a French restaurant that would have been considered peasant food a century ago. Or take social media, it has become a virtual necessity to participate in the today’s economy/society. I have lived on the outskirts of a rain-forest. How do you compare the poor in such a situation to our urban poor. I could make a pretty good argument that being poor in the Bronx is worse. In a rain-forest food is abundant and pretty much free. If your homeless, just throw up a shack from materials laying around the forest. On the other hand, the poorest of the poor in the Bronx have at least access to healthcare and emergency services. I think trying to compare the two situations by reducing it to a single number is not helpful. Maybe you agree with that.

        • I agree with what you’re saying here, but I also think that the perfect is the enemy of the good.

          Right now, today, is there such a dramatic difference between “living in LA” and “Living in Seattle” and “Living in Baltimore” that simply subsetting to 4 person households, calculating household income and dividing it by median cost to rent an apartment of at least 1500 sqft and buy food for 4 people is such a terribly off base calculation that it’s like comparing pancakes to tractors? No.

          If you can show me where the histogram of anything even close to this is calculated for every city in the US over 200k people every month for the last 40 years, I’d be very happy to go look at that.

        • To unpack this, here’s what’s a lot better about this than what we typically see:

          1) We’re calculating a dimensionless ratio of income to a fixed quantity of goods that we know people almost always demand, and whose quantity grows slowly with increasing income. few people with $400k/yr of income buy $45000/yr in apples and lettuce and raw chicken breasts. They also would rarely buy 4 times as much housing quantity as a $100k/yr family, quality is a different story, but since we’re using median prices for a fixed quantity… what we’re learning isn’t what people actually buy, but what multiple do they earn compared to what median people buy, a measure of ability to purchase multiples of what people who scrape by purchase.

          2) The fixed quantity of goods are relevant to typical quantities demanded by real families. We could scale different sized households by different sized apartments and food quantities.

          3) The focus is on the distribution of the quantity rather than some single statistic… Embrace variation! I’d suggest to simply publish the 1-9 deciles of this quantity, from which graphs like histograms could be made, or density estimates.

          4) The quantity so calculated results in a simple scale for which 1 is a distinguishing scale. The fraction of people below 1 is a decent proxy for “percent impoverished”. In fact, ideally you’d use income after taxes rather than a raw income. Furthermore you get something out of this better than just a percentage, you can calculate things like averages to figure out “how badly impoverished” people are for example.

          5) Spatial variation is important and calculating a histogram for all major cities would immediately show us an important aspect of variation across the country.

          6) The US is homogeneous enough that few people live in a situation similar to your rainforest where food falls off trees and a few tree branches is sufficient shelter. These goods and several others must be purchased by almost all of the US to be considered “out of poverty” in any meaningful way.

          7) The comparison is relevant to any “first world” or even 2nd world country, like Germany, England, Canada, Australia, France, Netherlands, etc… And is largely relevant to many mid-development countries, Brazil, China, India, Egypt, etc. The quantities, being dimensionless, are directly comparable across countries without any currency conversion hanky-panky

          And yet… if such a thing exists and is regularly calculated etc… I am completely unaware of it, and I have looked for something like it. I even spent several months producing something like it back in 2017, but got bogged down in just how hard it was to do given the data available… I also probably overdid it in terms of what I demanded from my model. Stan couldn’t sample my model reliably and I didn’t have the time to delve into debugging it.

          End result? We really know *very little* about some basic questions like “how hard is it for people to keep a roof over their head and food on their table” in the US… Seriously? In 2019?

    • You have a problem here:

      “But once manufacture has moved to the lowest cost place, and once that place starts seeing inflation in labor costs, the price of the low-end goods will rise faster than the high-end goods made in the same place across that span of time.”

      Who’s benefiting from the rising labor costs? :) The very people who are paying the higher prices. So this doesn’t make inequality rise necessarily.

      Also the price of high end goods isn’t controlled by labor costs. It’s controlled by market demand, which is much more flexible in high end goods, since wealthy people can move down market when prices rise. Poor people are already at the bottom of the market, so the only way they can cap prices is to stop using things.

      • “Who’s benefiting from the rising labor costs? :) The very people who are paying the higher price”

        Not if the rising labor costs are in China, and the lower-income people buying the lower-end products are in the U.S.

        But, yes, part of being richer is having options. A big part. I had the option to buy 2 coffees at Starbucks this morning, but made a pot of Folgers and put it in travel mugs instead. I did NOT get lower utility (whatever utility may mean) from spending less than $1 instead of $5. I actually got more utility because at some level I feel like I’ve outsmarted the system in a small way.

  8. There’s a lot of confusion floating around this thread because of assumptions about the purpose of this (or similar) exercises.

    What is or at least should be the purpose of this exercise? I think it’s to determine whether the total quantity of goods that can be purchased by a poor person *of a fixed type* has gone up, down, or stayed the same…

    In the abstract, the quantity of goods is at least related to their income I divided by the price of those goods P.

    Now, if we just measure the price as a fraction of some typical quantity of poor person’s income… we can do S(t) = P(t)/I(t), and then Q(t) = 1/S(t) is the abstract quantity of goods you can buy for the amount of income this class of people is getting at any given time.

    Now, obviously, people don’t spend all of their income on these goods, but if they spend a relatively stable fraction of their income on these goods, then it’s just off by a constant, and if they spend a slowly changing fraction then it’s off by a slowly changing function which we can write as a constant plus some small quantity… K + small_and_slow(t)…

    So, if we measure price of goods as a fraction of say median income of the group of people who tend to buy that basket of goods (call this dimensionless price)… we can immediately determine whether the quantity people are *able to demand* goes up or down with time… by calculating 1 over the dimensionless price…

    We don’t have to make any assumptions about declining utility or anything, since we’re NOT MEASURING UTILITY, we’re just tracking whether “people can afford more or less of this basket”, not whether they *want to* or *choose to* or *need to* just, how much could they get?

    People seem to be jumping to utility right away… and I think this is missing the point.

    Now, in this paper, rather than looking at say median income or some such thing, they’re taking the position of trying to calculate an income which gives constant capacity to demand cheap goods… and using this as a “poverty threshold inflator” which ok, that’s one way to use the idea I guess… I personally think there’s nothing meaningful about a threshold, and it’s better really to do the following:

    Generate a few thousand abstract people with nominal incomes according to the observed nominal income distribution.

    Estimate a basket of goods typically demanded by each person using consumer expenditure survey or something, with some bayesian uncertainty.

    Calculate price history of the estimated basket.

    For each person, based on the basket price for that person’s basket: calculate I/P.

    Look at the Bayesian posterior over the frequency distribution of I/P through time, (in other words, at each time point, multiple samples of the histogram/density plot representing the probability distribution over the frequency distribution at that point in time)

    These samples of the histogram show what it’s reasonable to believe, the distribution of people’s ability to purchase the sorts of goods they tend to want to purchase has done through time.

    • Can’t reply to your post above where you say:

      “6) The US is homogeneous enough that few people live in a situation similar to your rainforest where food falls off trees and a few tree branches is sufficient shelter. These goods and several others must be purchased by almost all of the US to be considered “out of poverty” in any meaningful way.

      7) The comparison is relevant to any “first world” or even 2nd world country, like Germany, England, Canada, Australia, France, Netherlands, etc… And is largely relevant to many mid-development countries, Brazil, China, India, Egypt, etc. The quantities, being dimensionless, are directly comparable across countries without any currency conversion hanky-panky”

      I am not as certain as you that even across the US, we are homogeneous enough to make these comparisons (LA and NYC, sure — Appalacia and the Bronx I doubt). But, I think that to complete your project of getting a better dimensionless measure, you ought to figure out some type of check on homogeneity. Let’s again take one of your examples: square footage. Asian cultures have adapted to living in much smaller spaces. Apartments in Hong Kong are a lot more efficiently organized and people are generally neater. So, there is really a sense in which they are getting more out of every single square foot than a person in suburban Chicago is. Maybe that is not a big enough deal to matter for making your comparison, maybe it is, but should you want some type of check on that issue to say that two places are homogeneous enough for the comparison to make sense. That’s what I mean when I say that all comparisons will break down at some point. I agree with you lots of times the kinds of differences I am pointing to won’t matter, but many other times they will, and the very fact that your intuitions about how often that happens differs from mine should make you want to build in some check to insure that we are near the breaking point. Sorry that I can’t offer more than that.

      • > So, there is really a sense in which they are getting more out of every single square foot than a person in suburban Chicago is

        sure, you could argue that square footage should have say a first order correction from country to country, and also maybe food should be scaled to the variation in calorie requirements for different populations… and maybe if you want to include some quantity of transportation, you should allow the mode to vary from place to place… I don’t find those objectionable. I’d argue we should also include bayesian uncertainty describing our model error… then we can be more realistic about how our “corrections” are not perfect as well…

  9. I think there’s a lot of confusion about this because no one wants to admit reality: the data we have can’t resolve it.

    My view is that scientists should recognize when the data are inadequate to solve the problem, and turn their energies toward generating applicable data.

    • I’m not so convinced of this. The archive of data above is rather substantial https://statmodeling.stat.columbia.edu/2019/11/06/have-prices-have-risen-more-quickly-for-people-at-the-bottom-of-the-income-distribution-than-for-those-at-the-top-lefty-window-breakers-wait-impatiently-while-economists-struggle-to-resolve-this-dis/#comment-1157403

      I don’t think it can completely answer all the questions posed here, but just asking better questions with the available data is a good start. The next step would be collecting more specific data, or processing the raw data collected by these surveys in a different way perhaps.

      • How do you assess the inflation in the cost of a telephone? A modern cell phone isn’t remotely comparable to a 1990 land line phone. It’s a computer with more power than the PentiumI I bought in 1993. It’s a phone. It’s a camera. It’s a world atlas at any scale. It’s an encyclopedia. It’s a conduit for advertising. I haven’t even started. Sure, it costs $500, but it does the work of what $5000 did 30 years ago.

        So how do you assess that? It’s fine to come up with a broad brush measure that works on a Q to Q or Y to Y basis, but beyond that inflation is meaningless except for a few commodities.

        • people always point to the latest tech gadgets that improve exponentially with time and say how silly it is that people might want to compare things over time.

          Look, everyone, based on their body size/type, has to eat some number between 1000 and 3000 calories a day to survive. If you don’t, you die. There’s not much choice about it.

          If you spend the night outside in North Dakota you die in less than 6 months, once the serious deep freeze occurs.

          If you sleep on the streets in New Orleans you drown in the next hurricane.

          Shelter, Food, clothing, medicine, hygiene products. There are approximate floors on the quantities you can demand while still remaining alive. Even the homeless only survive by coming up with some shelter work-arounds.

          So, calculate those floors, figure out the cost of those floors, and calculate Income/CostofFloor. To say this number is useless is obviously wrong. To say that it fails to tell us everything about the economic is obviously correct, because it’s not supposed to.

          Today our government still uses some back of the envelope calculation someone in the USDA did in 1960 with a pen to “figure out the poverty threshold”…

          They don’t even publish a histogram of the ratio of income to the poverty threshold number, much less calculate a meaningful poverty threshold.

          There’s a whole multi-million dollar effort to create a supplemental poverty measure that STILL doesn’t even publish histograms of Income/PovertyMeasure by major metropolitan area… https://www.census.gov/library/publications/2018/demo/p60-265.html

          FFS

  10. The idea that you would normalize currency by different baskets of goods for two people that might be living less than a mile apart and could even be siblings sounds preposterous to me, at least in the context of measuring income inequality.

Leave a Reply to Martha (Smith) Cancel reply

Your email address will not be published. Required fields are marked *