Skip to content

Big trouble coming with the 2020 Census

OK, first things first. For readers of this blog who live in the United States: Don’t forget to fill out your census. They’re doing in online, and you should’ve received a letter in the mail last month telling you how to do it.

And now the news. Dr. Z points us to this post by Diana Elliott and Robert Santos, “Unpredictable Residency during the COVID-19 Pandemic Spells Trouble for the 2020 Census Count.” Elliott and Santos write:

Just before lockdowns were implemented across the country, there was tremendous movement and migration of people relocating to different residences to shelter in place. This makes sense for the people involved but could be disastrous for the communities they fled and the final 2020 Census counts.

The 2020 Census, like most data collected by the US Census Bureau, is residence based. . . . Most residences across America have already received their 2020 Census invitation. Whether completed online, by paper, by phone, or in person, the first official question on the 2020 Census questionnaire is “How many people were living or staying in this house, apartment, or mobile home on April 1, 2020?” Households are expected to answer this based on the concept of “usual residence,” or the place where a person lives and sleeps most of the time.

Despite written guidance provided on the 2020 Census on how to answer this question, doing so may be wrought with complexities and nuance from the pandemic.

First, research reveals that respondents do not often read questionnaire instructions; they dive in and start answering. With many people scrambling to other counties, cities, and states to hunker down for the long haul with loved ones, this will lead to incorrect counts when people are counted at temporary addresses.

Second, for many, the concept of “usual residence” has little relevance in the uncertainty unfolding during the COVID-19 pandemic. What if your temporary address becomes your permanent address? What does “usual residence” mean during a global epidemic that could stretch for 18 months or more? And perhaps more importantly, what should it mean? . . .

The US Census Bureau must act. It will need more processing time to identify and remove duplicates in the returns—a phenomenon that occurs regardless of a pandemic—and will need to flag potential population spikes in certain communities. . . .

Unfortunately, communities face a zero-sum game for decennial population counts. Communities that gain population in 2020 because of the pandemic will reap the benefits of better funding and representation for the next decade. Communities with population loss will receive less than they deserve.

Every census count brings new challenges, some of which lead to miscounts that get brought to and battled over in court. Without a proactive approach to the 2020 Census that addresses these residency questions, the COVID-19 pandemic may be unintentionally inviting communities to wage contentious court battles over the accuracy of the count for years to come.

A few weeks earlier, Santos and Elliott had written a post, Is It Time to Postpone the 2020 Census?:

We know that COVID-19 testing in the US has proven inadequate, and community spread has now taken hold. The virus has spread to 45 of 50 states as of March 12, 2020, and it’s reported to be 10 times more lethal than influenza and much more contagious. . . . Although the decennial census is mandated by the Constitution, the extreme challenges raised by the pandemic may warrant an unprecedented delay to protect the census’s accuracy. These challenges include:

Difficulty finding and retaining enumerators . . .

Making hard-to-count populations even harder to count . . .

Lacking planning or protocols for conducting the census during a pandemic . . .

Should the census be postponed? Extended? Canceled? . . . Regardless of the option, it is hard to imagine that the 2020 Census could simply go on as scheduled. Some hard decisions face the US Census Bureau. The health of our democracy may be at stake.

We should’ve listened to them back on March 13th.

P.S. A commenter suggests they change the census form to clarify responses for people who are temporarily housed elsewhere because of coronavirus.

The problem is that the census form was already written, and I guess they don’t want to change the form in the middle of the Census. At this point they’d have to redo the whole thing, which I guess is what they should do, but that would be expensive. Also there are winners and losers from every change, and the winners from the current system might want to keep things as is. Finally, there are people outside and inside the government (but presumably not in the Census Bureau itself) who want to “drown the government in the bathtub” etc., and for them I guess it’s a plus if the census is a failure, as it will reduce legitimacy of future governmental actions. This is related to the War on Data that Palko and I wrote about a few years ago and which Palko reblogged recently.


  1. Zhou Fang says:

    In any reasonable research study, it would be absolutely insane to have the study go ahead in this current situation.

    • jim says:

      The Census isn’t a typical research study. It’s a massive constitutionally mandated count that’s in preparation for years in advance, and a lot of people including me have already completed it.

      Is there that much potential for “massive” effects? Who is moving all over to shelter in place? The biggest movers will be college students, but many maintain residency with their parents anyway.

      We completed the forms on-line, so it shouldn’t be that big of a deal to update the forms with some simple explanation and otherwise live with the consequences.

      • I hear the effects are actually significant and noticeable that people are fleeing big city areas. Not sure as a percentage of the cities, but as a percentage of the small towns they’re heading to it could be an issue.

      • N says:

        “Many [college students] maintain residency with their parents anyway.”

        Non-commuter college students are supposed to be counted where they attend university (the Census counts where you live the majority of the year, not where you maintain residency administratively speaking) so this is a bigger issue than your comment implies.

        • jim says:

          “Non-commuter college students are supposed to be counted where they attend university”

          Well you can see how that’s a potential point of confusion anyway, because states go out of their way to prevent out-of-state students from establishing local residency, so it might seem surprising to students that while they’re getting screwed out of resident tuition and can’t vote locally, they’re getting used for federal funding. :)

          • digithead says:

            Prisons do the same for communities but at least the college kids are actually in the community spending money and using services.

            Take Montana State Prison in Deer Lodge which is in Powell County. Both Deer Lodge and Powell County get federal money because they have a significant amount of minority residents according to the census. The problem is that the majority of those minority residents are incarcerated. None of the federal money goes to the minority inmates.

      • Andrew says:


        I think there is potential for big effects, for reasons discussed at the linked post. I agree that updating the forms could make sense. But that’s a decision that would have to be made; it won’t happen by itself.

        • jim says:

          Problems I can believe.

          “The health of our democracy may be at stake”

          This I doubt. If the census isn’t right we’ll live through it without falling into tyranny, and if we do fall into tyranny, it won’t be because of the census.

          • Andrew says:


            Every little bit hurts. But, sure, I agree that “the health of our democracy” is an overstatement.

            • jim says:

              People claiming that every imperfection in direct representation is a “threat to democracy” is the real threat to democracy. It trivializes the real threats.

              Here’s Stiglitz mixing the trivial with the serious. “Specious arguments”, gerrymandering, various voting laws, and money in politics may or many not seem fair, but they’ve been a part of democracy since the day it was born. And surely voting is easier today than it was for most of the history of this country. These things are not a “threat to democracy”.

              OTOH, eroding the separation of powers is a very serious threat to democracy, as is the integrity of elections, both of which are legitimate concerns at the moment.

              • Rahul says:

                Jim: Exactly right.

                The problem is that each one of us loves to over-emphasize the little sphere of life they are involved in.

                There is as much a “threat to national hygine” by keeping hairdressers closed for a month.

                It’s all about perspective!

              • jim says:

                Rahul says:

                “There is as much a “threat to national hygine” by keeping hairdressers closed for a month. ”

                Ha! Yeah, and man I do a haircut bad! :)

              • Martha (Smith) says:

                Jim said, “Ha! Yeah, and man I do a haircut bad! :)”

                My last haircut (a month or so ago?) was the first I’d done myself in years (decades?). I looked at some You Tube videos before attacking the problem. The results weren’t great (my part was a bit off), but it was adequate. (And I’ve got to admit that “shoulder length” is a lot easier than short.)

              • Zhou Fang says:

                I mean, sure, if your baseline for democracy is the 1800s, but I think the possibility of reverting to that situation will look like alot like a threat to democracy for a lot of people, especially people who don’t look like you.

              • jim says:

                “Sure, if your baseline for democracy is the 1800s… I think the possibility of reverting to that situation will look like alot like a threat to democracy for a lot of people”

                First the fact that some things have always been done doesn’t mean we should keep doing them. There’s plenty of room for improvement.

                Second, none of those things are restricted to the 1800s or even the 1900s. They’ve been happening on and off continuously throughout American history. Both parties use and always have used gerrymandering, specious arguments and voting laws (or lack thereof) to see to their own interests, right up to the current day.

                Third, a “threat to Democracy” isn’t the same as “a threat to (my or your) personal interests”. Democracy is the right to participate in the government, not the right to get what you want.

              • jim says:

                Zhou Fang:

                I’d love to see a system that takes the creation of state and national congressional district boundaries out of the hands of the parties. But by the same token it can’t be done by bureaucrats who aren’t directly accountable to the voters. It would have to be done entirely by some agreed-upon algorithm, which couldn’t be altered by anyone.

                But if history is any guide, this election will be a massive Democratic landslide, and then all the people screaming about threats to democracy will be using all the same tools they claimed were a threat to see to their own interests.

  2. MM says:

    Surely you can ameliorate the problem by explicitly accounting for COVID-19 in the questions. Something like,

    Where are you living during COVID-19? Did your place of residence change due to COVID-19 relocation? What was your previous address?

    It would make it obvious to people how to answer the questions, forcing them to think about the effect COVID-19 has on their circumstances. It would also allow you to get more data and decide how to treat it for better decision making.

  3. Just deemphasize the decennial census in the areas where it’s used for policy etc. We run the ACS every year, use it to correct the estimates. yes I know this becomes a political issue but pretending that the census is reality is also a political issue

    • N says:

      Demographers typically count on the Decennial to correct the ACS, so this is exactly backwards from a demographic perspective. The ACS is extremely unreliable for small areas, and the Census is used for policy in a lot more settings than you evidently imagine, many of them mandated by statute. There are no good or simple solutions here.

      • The right way to think about this is that they both provide evidence which should be pooled together and analyzed using known issues to produce a proper estimate through time.

        But Bayesian methods seem to be far from the norm in any of this stuff.

        • Dalton says:

          It’s a supposed to be a CENSUS. The whole point is that there isn’t any estimation. When you sample the whole population you don’t have a sample, you have the population.

          • Andrew says:


            Nonresponse and multiple response are inevitable, so there’s still some estimation going on in any case. I agree, though, that it’s important to work hard to try to reach everybody. According to the Census website, the response rate so far is under 50%, so I think we’re in trouble.

          • If you ask every person in a room full of 100 people what their net worth is do you think you’ll get the right answer? How many people know their net worth down the last penny? What about height? Do you know your height to the nearest millimeter? Does that quantity even exist as a well defined number? It changes when you breathe.

            Even if there aren’t *sampling* errors (and there ARE always sampling errors in the census) there are all sorts of model/measurement errors. There are multiple people just here on this page who have said they don’t know how to answer the questions. And that’s among people largely with PhDs in quantitative science fields.

            Next we can discuss the problem of how using the census numbers taken in 2020 to decide on how school money is apportioned in say 2022 is hugely problematic. For example suppose there’s a massive hurricane in the gulf coast in 2021 and literally millions of people relocate…

            So, pretending that all that matters is sampling error and that there is no sampling error in a CENSUS is naive.

  4. Jonathan (another one) says:

    The online form is fairly clear about this for people who read carefully…. all ten of them.

    As it happens, I was in my summer place in the Berkshires on April 1, but I lied (twice) and said I was in my regular home in NY and that the Berkshire home is just a temporary residence. OK, maybe “lie” is too strong a term, because we were asked to project residence in March and I did, though the error bands were quite wide, and they didn’t ask for a maximum likelihood estimate… and they didn’t ask for the error band.

  5. Thanatos Savehn says:

    Of course communities “face a zero sum game.” That’s the whole point of the Constitutionally mandated census – to apportion representation for a representative form of government. Sorry to be grumpily cynical but to me it sounds like the usual decennial effort of the Urban Institute and like-minded organizations to lay the ground work for lawsuits over whether New York loses one congressional seat or two.

    • Phil says:

      I’m not sure whether you’re suggesting that the numbers won’t actually be different (wrong) because of the current situation; or that sure, they’ll be wrong, but who cares, people just complain no matter what and what difference does it make anyway. Or maybe a third option?

      • I think he’s suggesting that they won’t be “wrong” and they won’t be “right” they’ll just be different from what they would have been without COVID-19 and the question of whether they’re wrong or right is really a political issue about “what does ‘right’ mean”

        I more or less agree… One thing I think that’s wrong about the decennial census is that it’s every 10 years, and so it rapidly goes out of date… We have the ACS which can tell us about time trends in between the census. We should really be using all available evidence to update the administrative rules annually. Of course we don’t reapportion the representatives annually, but things like federal highway money or school district money or whatever shouldn’t be relying on 9 year out of date data for example.

        • Martha (Smith) says:

          Good points in second paragraph.

          • jrkrideau says:

            Ten years may have been fine in the 18th Century but not the 21st.

            Canada has a census every five years. It help and there are a surprising number of organizations that use that data down to your local Timmy’s.

        • Thanatos Savehn says:

          Exactly. Florida will be more than happy to count the New York diaspora as its own and New York’s representatives, who’d rather not decide which of them gets thrown out of the life raft, will claim shenanigans. The NE Dems will call for models (that will, given their experts, predict salvation) and the SW Reps will demand only actual noses get counted. Lawyers will make bank. The courts will either discover Mr. P in the penumbras of the Constitution or they’ll leave it to peripatetic voters (who increasingly vote with their feet). The vote will be 5-4 either way. If you don’t like it pull the other lever next time.

          Given that we’re apparently about to spam the countryside with RT-PCR machines in hopes of corralling COVID-19 maybe we should just consent to cheek swabs and do the GATTACA thing and get it over with.

          • Rahul says:

            One feature of this Covid crisis is the legitimization of state-level NIMBYism at a scale never seen before. Maybe I forget some history.

            To me that is very worrying.

          • Phil says:

            Sure, I understand that’s what will happen. What I’m trying to get at is: does it matter?

            The strong form of the ‘no’ answer is that there is no right or wrong answer anyway, it’s all just politics (to echo Daniel L) so if you complain about the way the census is done you’re just a whiner and nobody should pay attention to you.

            The weak form of the ‘no’ answer, perhaps being argued by jim above, is that there are many imperfections in our system and an inaccurate census, though undesirable, is just one shuffling step in the wrong direction, but as long as the general movement on historical time is in the right direction, the mis-allocation of a few billion dollars or a few congressional districts is in the noise.

            I suppose I won’t even try arguing that the census numbers actually matter (although that is my position, and I am genuinely surprised to find that it has little support in this discussion), since that appears to be a bridge too far. For now I’m just trying to understand where people are coming from.

            • I think my point is that the census numbers shouldn’t be so high stakes. We should do the census, but know that *every* time we *ever* did a census it *always* had problems. We should use the census, with its known problems, together with additional data such as the ACS, and indirectly data from other surveys like BLS surveys of expenditures and etc etc to get a continuously corrected time-series… if this census has a larger error due to temporary migration issues, then it should get lower weight than it might otherwise, and the trends in the ACS from 2018 and 2019 can be used to pool the estimates towards a more correct number.

              Treating the census as “ground truth” is the problem.

            • jim says:


              Your assessment of my position is mostly accurate, but I don’t say we should ignore steps in the wrong direction. What I say in this *particular case* is that, weighed against everything else, the likely degree of inaccuracy isn’t sufficient to warrant stopping the census.

              But in the “stop the census” argument, there is an overlooked assumption: that whatever movements have occurred in the last few months are temporary. Maybe they aren’t. Maybe some kids will just continue their education on-line. Maybe some people will stay near their parents and work long-distance over the web. I doubt this will have a huge impact, but we don’t know that, and we really don’t know the impact of people moving around either.

              And maybe most people will read and understand the census form and fill it out right anyway.

              • and maybe filling it out “right” will result in the *wrong* answer, because the migrations become more permanent… which all goes to my point that we want a continuously updated timeseries not an error prone snapshot in time that lasts for 10 years.

              • Phil says:

                Oh, I’m not arguing for (or against) delaying the census or changing it or whatever. I’m just trying to understand what I perceive as a sense that some people here seem to have, that the accuracy of the census doesn’t really matter.

                I agree with jim’s statement that an unusually large number people don’t know where their ‘permanent residence’ is at the moment, or even what that would mean exactly. And I agree with the point, made by several, that there’s not an unambiguous definition of what the ‘right’ census count would be, for any place in the country.

                I definitely agree with Daniel’s point that the idea of trying to get an absolutely perfect census every ten years is a pipe dream, unattainable with any reasonable level of expense or effort, and that we should (in a closer-to-ideal world) acknowledge that and just try to get the best picture we can. But we live in the world we live in, and even seemingly innocuous statistical adjustment for undercount has been rejected by the courts. Until there’s a constitutional amendment, the census is going to continue to be performed and is going to be the sole dataset for allocating congressional seats and the electoral college. I think that’s a fact.

                But just because there’s no single ‘right answer’, that doesn’t mean every answer is equally wrong. If the total number of people counted does not equal the number of people who live in the U.S. then the count is wrong, and the more those numbers differ the wronger they are.

                If the number of people counted in a congressional district is less than the number of people who live there by _any_ reasonable definition of who lives there, then the count there is wrong, and the more those numbers differ the wronger it is.

                And so on.

                In short, to the extent that anyone here is arguing that it doesn’t matter how the census is conducted, I disagree.

            • Thanatos Savehn says:

              It certainly matters to the politicians and to the state governments that rely on federal largesse (especially so for states like NY with enormous legacy obligations that they can’t hope to meet or avert because they are in thrall to public unions).

              If it were up to me we’d run it like a law firm. Recognize the noise and smooth out the otherwise bumpy ride by averaging over time. However, the Constitution requires an actual enumeration so to change it there’ll need to be an Amendment. On the other hand, very few states run the sort of “be counted” program that NY runs. If we supercharged the other 49 states to count every last nose within their borders I suspect NY would only fall faster. Thus, NY’s slow demise is probably helping it retain power that it couldn’t under almost any other scenario.

              • Phil says:

                Thanatos, you say “It certainly matters to the politicians and to the state governments that rely on federal largesse”, and who could disagree, but do you really think those are the only groups to whom it matters?

  6. Peter Dorman says:

    I won’t call what follows exactly a prediction, since I hope it’s unlikely; let’s just say it’s a possibility.

    The Republicans mobilized intensively for the 2010 state elections, realizing the benefit of controlling the decennial redistricting process. As a result, they achieved a substantial increase in their effective share of votes (compared to raw vote counts). They have an incentive to put forward the idea in 2020 that this year’s census is so flawed, due to relocation issues and the breakdown in door-to-door followup, that further redistricting must be postponed, perhaps until the 2030 census. This would be subject to court challenge, which might play out differently at the state level. (or resolved by a 5-4 at the US Supreme Court.) Disputes over redistricting would add to the growing delegitimation of electoral processes and outcomes in the US.

  7. Kenny says:

    I tried to fill out the census online but felt stymied by exactly what you describe. I moved to Brooklyn in the middle of March. (I hadn’t been keeping up with my feeds and I don’t follow more mainstream news at all.) I’m nominally sharing an apartment with three other people but two of them had left before I moved-in and the third left a day or two after – is our apartment still their ‘usual residence’? Given the stakes of the census, I didn’t feel comfortable answering at all.

    • Bill Spight says:


      Answer the census by phone. I did so last night. There was one question where the options did not seem to fit my situation. I quickly explained this to the representative, who suggested the closest fit. The whole call lasted less than 15 minutes, including getting through the automated system to a representative.

  8. Leon says:

    An upside is that the census will be a great data source for the many natural experiments underway.

  9. Rahul says:

    Isn’t it bizaree to even worry about a census in the midst of all this?

    Just postpone it and get on with the things that matter. Just divert the budget and resources to something more meaningful given the situation.

    • Andrew says:


      The difficulty is that the census is a big project. I agree with you, and with the authors of the above-linked article, that the census should be postponed. But the decision is not up to us!

      • Rahul says:


        What’s up to us is to lobby. Put pressure. Shout “postpone it” from the rooftops.

        The problem in such large government projects is always bureaucratic capture. There will be truckloads of men within the system with a vested interest NOT to postpone it just because they care more about their role and turf than the larger picture.

        We just do our part in counterbalancing that.

      • jim says:

        Stopping the census would turn a modest and manageable problem into a gigantic disaster.

        We have an election looming. That election really is at risk, and if that election isn’t clearly free and fair, our democracy really will be in trouble. So please let’s just fix the census as best we can and focus on a free and fair election.

        • Rahul says:

          What about postponing the election too?

          Compared to the rest of the world is the US seriously in jeopardy of not having had free and fair elections anytime recently?

          I feel that a lot of people when they don’t like the outcome love to blame things on not having had a fair election.

          • Knowing what I do about computer security (which is a fair amount, probably in the 80%th percentile), I honestly would put about a 50% chance on Trump’s election was carefully arranged by foreign powers hacking non-auditable voting machine results in swing states.

            There’s textual evidence that I thought that back in 2016 before the election in comments here on the blog. so it’s not just sour grapes.

            That being said, I’m not clear that we’ve done a great job improving that. Here in LA county the new machines seem good, they print out an auditable paper ballot and suck it into a ballot box. how true is that in Ohio or Florida or North Carolina or whatever? I don’t know.

          • jimI be says:

            “I feel that a lot of people when they don’t like the outcome love to blame things on not having had a fair election.”

            yes, I agree! I believe the 2016 election was free and fair.

            My concern is that, with all the controversy over the last election, this one must be, with out a shadow of a doubt, free, clear, and fair. I’m just not sure what would happen if there were questions about the election and Trump won by a small margin. I suspect a lot of people wouldn’t accept that and the consequences could be dire.

            And despite my belief that the previous election was fair, I put nothing past Trump and the people who support him. It’s really critical that this one be right.

            But I note that Trump is obviously worried about the economy and the impact it could have on the election. So he is worried about getting reelected.

  10. David Marcus says:

    The cost is irrelevant since the federal government can’t run out of money. (Yes, I know what happens if you print too much money, but that is not a problem we will be having.)

    • Rahul says:

      How do you know you won’t be having that problem? Just curious.

      • it seems likely that with the shutdowns, lots of people won’t have income to pay their loans. this could result in a collapse of the money supply similar to the 2008 housing crisis. the loan crisis this time is more likely to include the inflated corporate debt.

        when the money supply collapses, printing more of it is necessary if you don’t want commerce to lock up and the price of everything to collapse. if you print it and give it to census workers it will not be a large enough quantity to cause major problems.

        I argue we should instate a UBI. in normal times the UBI should redistribute money by taxing and refunding… but when monetary supply collapses the UBI is by far the best way to restabilize it. it ensures the goods demanded are the ones demanded by all members of society.

        • Rahul says:

          Historically, weren’t catastrophic events like the great wars a prelude to hyperinflation in many countries eg Germany / Greece etc? Why would this be different.

          • Dale Lehman says:

            If there are any macroeconomists out there (I am a microeconommist and consider macro to be astrology), perhaps they can shed some light further than my brief comments. But a few basics should be emphasized before people go too far in these directions. The only hyperinflations I am aware of resulted from the government literally running the printing presses (e.g., after WWI, Germany was forced to repay war debts but had no resources, so it printed currency – when the world refused to accept payment in this now worthless currency, all those bank notes stayed in Germany – with more cash to spend and precious little goods to buy, prices skyrocketed). Other hyperinflations had the benefit of understanding the German experience, but when you have no better short term options, then you repeat that sorry experience (e.g., Zimbabwe “knew” better, but really didn’t have any better options, if they wanted to pay their cronies and remain in power).

            The “printing” of money we are talking about with COVID is not from the printing presses – it is from the government massively increasing the money supply through the financial system (mostly banks, but also other financial institutions). When the government buys a large amount of financial assets (often at over-valued prices), they are putting “money” into the economy (in the form of credit that financial institutions can then loan out). In theory, this expansion of the money supply can lead to inflationary pressure. It can also stimulate the economy (and that is its intent).

            There are several links in this theory that may make it work differently than predicted. You can make credit available but that doesn’t mean anybody will want to borrow it – in the current environment, having massive amounts of credit is not likely to lead to an equally massive upswing in borrowing. So, the credit may just sit there – perhaps only going to purchase government assets (e.g. T bills) – seems like a circular pattern of government running huge deficits to put money in the system only to see it used to buy the T bills that the government issues to to borrow what it needs to run the deficit – remember I’m not a macroeconomist, so I don’t really see how it will work differently than this). So, while monetary policy has been the favored macro policy for several decades now, it is hard for me to see how it can work any magic at present.

            Fiscal policy, on the other hand, has the government running massive deficits (also borrowing massively, at today’s particularly low interest rates) and spending that money directly on goods and services (including direct payments to individuals, such as the UBI, unemployment insurance, etc.). This seems totally appropriate in today’s situation. Will it lead to large inflationary pressure? It depends. If the goods people want to buy are not available (due to damaged supply chains that cannot be repaired), the perhaps prices will rise rapidly. That seems unlikely to me, but not to be dismissed. More likely (in my opinion), it will help prevent the worst from happening. It seems like sound short run policy. Of course, there are long run consequences: eventually the debt gets repaid – to whomever loaned the government the money. And, if interest rates rise from their current levels, financing that enormous debt will take a large amount of tax revenues. To the extent that this debt is held by US citizens, this will be a redistribution from taxpayers to those citizens. If it is held by foreigners, it will be a redistribution from US taxpayers to foreigners. Likely it will be both, since the US debt is being financed both domestically and from abroad.

            That’s the extent of my foray into astrology. Feel free to discount most of it. The only thing you should necessarily retain, is the distinction between hyperinflation due to “printing money” and they type of “printing money” that is currently being done.

            • This seems like a good summary of the basics.

              I’ll just add the following bit of theory (warning, i’m not an economist at all, I just play one on the internet)

              Here is some discussion on the various money supply measures…

              Suppose we take MZM as the current broadest definition. That’s

              1) Notes and coins in circulation
              2) Notes and coins in vaults
              3) Travelers Checks
              4) Demand deposits (checking accounts)
              5) Other checkable deposits at banks
              6) Savings Deposits
              7) Money market funds

              Now, suppose you have a bank that loaned out a lot of money to some companies. It did so by buying their bonds and putting them in money market funds. Now, the company runs a chain of restaurants, and it has no business, so it can’t pay its bonds… There are hundreds of such companies… The value of the bonds declines, maybe even to zero. The money market fund collapses… So the amount of money available in the economy is cut by say 20% in just 30 days.

              That’s a bit of a disaster. What happens if the government literally just “prints” money by writing out checks to every man woman and child, who deposit the checks in their bank accounts. The quantity of such checks is 20% of the previous MZM quantity.

              Now the total money supply is the same as it was 30 days ago, but its distribution is different. Instead of being concentrated in the hands of people who owned a lot of bonds through money market accounts, it’s distributed in the hands of every person in the country equally. The government doesn’t have to “borrow” this money, it can literally “print” it. However, it doesn’t print coins and bills, it just “manufactures” money in people’s checking accounts, either through electronic direct deposits, or through paper checks. It can literally “by fiat” force banks to accept those checks without any “gold” backing it or any of that nonsense.

              Now, clearly if you manufacture 10x the amount of money lost, you’d expect to see a lot of inflation, and foreign countries would stop accepting dollars… it’d be a disaster.

              but if you basically hold the money supply constant-ish by manufacturing more or less the amount of money that disappeared in a puff of smoke when companies went bankrupt and their bonds became worthless… the only “inflation” you’ll see is price pressure applied by changing *who* is demanding goods, and hence the demand for some things will go up, other things down, and in the medium term production shifts to the things that “everyone” wants instead of just the things that finance industry people who control big finance funds want.

              That would in my opinion be a good thing. We’ve seen plenty of what it looks like to hand the finance industry cash… you get significant financial asset inflation, and lending/spending on highly speculative ideas like companies that make apps to deliver frozen yogurt by automated drone or whatever.

              • Dale Lehman says:

                I agree with your policy direction and conclusion. However, I do think the government must “borrow” this money and cannot just “print” it. If it is possible to put this money into people’s accounts without borrowing it, then the government should simply do this all the time. Aside from inflation, there would be little downside. But the fact is that the checks the government would be writing are not derived from any revenues, so eventually the checks will be cashed (either by the immediate recipients or subsequent people, such as their landlords) and then the government will need to borrow what they haven’t collected through revenues. I don’t think it makes any difference that the government checks are replacing part of the money supply that was there 30 days earlier – there has been a real cost to the economy, and the government is now trying to address it. Like I said, I agree with your policy prescription, but I don’t see how this is done without government borrowing.

              • Rahul says:


                I agree. The crux is ” if you basically hold the money supply constant-ish”. And that does take huge discipline once you get that going. So I think it’s good, but we need to do it cautiously.

                To me that’s the crucial part. In old times they actually printed money like Dale says and now we have more sophisticated mechanisms for doing it. The fundamentals remain the same.

              • Dale, see my description of quantitative easing below… The government literally “poofs” money into existence by buying bonds with “nothing” to the tune of 3T over the last decade.

                At one point I dug through the Fed’s website to get clarity on how this works. It literally just manufactures electronic balances in computers to buy bonds from banks. Why do this rather than manufacturing electronic balances in everyday people’s checking accounts, and let them buy goods and services?

                As long as you’re not increasing the overall money supply dramatically, this is absolutely the way to go.

                Of course, in “regular” times when the money supply isn’t collapsing, in order to hold that supply constant, you can’t just manufacture money, you must tax it from one group and redistribute it to another, or you WILL have inflation.

                so, it depends a lot on circumstances as to whether you can “just print” it or whether you must tax it or borrow it.

            • Rahul says:



              I don’t see the distinction you draw: Even if we don’t use “printing presses” literally, the inflationary pressures are the same, fundamentally. You are injecting more “money” into the system. The only distinction is between good monetary policy and bad.

              I agree that there’s a role for monetary policy. I just think we should be cautious: I fear complacency if we adopt the “but that is not a problem we will be having” dictum.

            • Consider what the government does when it did “Quantitative Easing” back in 2010-2018 or whatever.

              It has communications with banks (the “discount window”), who it offers to buy government bonds from. It buys these at high prices, much higher than the prevailing market prices. A bond purchased at a high price is equivalent to saying it pays out a low interest rate. So when they say they’re controlling interest rates, what they really mean is they’re overinflating the value of government bonds held by banks by buying them at outrageously high prices.

              How does it accomplish this? Once it gets the bond, it literally just increments the electronic records of how much money is in the bank’s account at the fed. Poof, more money comes into existence by pure fiat. Once the bank has more money in its Fed account, it can lend more out…

              Why we should use this mechanism rather than the mechanism of UBI with “poof” money is a question we should all ask ourselves. Does it make sense when the finance industry runs the economy into the ground to just “poof” money into existence and *give it to finance people*?

          • Rahul, Dale, others… Here’s a blog post on the topic of the money supply and the Fed, and how money comes into existence in the modern world (complete with some not-too-crude language, I’ve been grumpy about this issue since ~ 2014 or so):

            • Dale Lehman says:

              The reason we care about the money supply is due to its impact on the real economy. Daniel, in your narrative example, the bank has loaned money to a business by buying the business’s bonds and then put that into a money market fund. Then, when the business (say a restaurant chain) goes bankrupt, the bonds decline in value (perhaps to zero), and the money market fund collapses. You then call this a decrease in the money supply.

              Several features of this story seem incomplete to me. First, the bank loan cannot just put the bonds into a money market fund – those bonds must be sold to another party, and the proceeds are put into the money market fund. The restaurant chain’s bankruptcy is a real event, and the collapse of that business may lead to a collapse in the value of the bonds which may lead to a collapse in whoever bought the bonds from the bank. The money market fund might collapse or it might be another fund that the purchaser of the bonds had invested in. The money supply itself declines because of the initial loss in the real business of the restaurant.

              The extent to which this chain of events can be countered by the Fed printing money depends on the extent to which the collapse in this chain of events is due to a collapse in confidence. If the lenders in this chain start believing they will not be repaid, then this can lead to recession – and the classic monetary response to expand the money supply can prevent that. However, there is nothing the Fed can do to erase the effect of the real collapse in the restaurant’s business. Otherwise, natural disasters would never impact the real economy – we would simply have the Fed counteract the real disaster and “poof” it away. Credit markets matter because they impact the real economy, and collapses in credit may be successfully mitigated by the Fed. But there is nothing the Fed can do to erase the real damage that occurred due to a virus, hurricane, etc.

              In the present circumstances, people are not paying their rent because they don’t have a paycheck. Then, their landlord stops paying their bills due to the lack of rent coming in. And so on, and we have a recession. But, is it caused by a collapse in the money supply? Or is the collapse in the money supply a symptom of the collapse in the real economy? I suspect it is partly both. To the extent that expectations and confidence have been shaken and will worsen the situation, the Fed “poofing” money into accounts can stem that tide. But, to the extent that the real economy has suffered a natural disaster, the Fed cannot “poof” away those costs.

              Of course, the issue of who suffers the costs and who does not matters, and policy will impact these. Those are mostly micro impacts, although they can have macro impacts (since our consumption-based economy relies on mass consumption, anything that spreads the costs widely while saving a concentrated segment of the economy, is likely to have negative macro consequences – hence, I would prefer the UBI to a bank bailout in the current circumstances).

              • Dale, I fully agree about the importance of the real economy. In my mind, the role that UBI plays in stabilizing the money supply also stabilizes the demand for critical goods to keep people alive, healthy, eating, housed etc. If people lose paychecks, and as a follow on they can’t pay their food bill, and as a follow on grocery stores can’t pay food suppliers and as a follow on food suppliers stop paying workers to pick produce, and as a follow on milk is dumped into sewers and vegetables rot in the fields… This is a MUCH bigger *real* impact than just the impact of the loss of restaurant-going-experience and restaurant food.

                (that’s already happening, see here: )

                The FED can’t prevent restaurants from losing business, it’s illegal and physically risky to sit in a restaurant or serve people food in groups at the moment. What CAN be done is to ensure that sufficient funds flow through the system that we don’t ALSO have a collapse in food production, housing occupancy, medical services, transportation of goods, etc.

                While it winds up being a transfer from producers of goods to consumers of goods without the consumers producing much in return, it is far preferable to spoilage of food, starvation of the poor, soap, shampoo, cooking utensils, and plumbing supplies sitting in warehouses, trucks sitting in parking lots, and 25 million bankruptcies to be processed in September alone.

                So, as I see it, the velocity of money is about to come to a near halt, and that will lead to money supply collapse as bonds become worthless, and all this will lead to death and suffering that is needless because we could just print the money and give it out in UBI to keep all the wheels spinning allowing demand to shift to delivery services, food production, hospital care, critical building maintenance, and etc etc.

              • Anoneuoid says:

                a collapse in food production, housing occupancy, medical services, transportation of goods

                food production -> obesity
                housing occupancy -> housing bubbles
                medical service -> drug pushers
                transportation of goods -> walmart et al driving local business bankrupt

              • Not quite.

                It’s more like:

                Government subsidized ranching and factory farming -> below market cost of processed and fast foods -> excess consumption of low quality diet -> obesity especially in poor populations.

                Restricted housing construction, NIMBYism, protection of entrenched powers -> growth of housing capacity below growth of population -> high prices for housing -> homelessness

                Government restricted competition in medical markets (patents, licenses, manufacturing rules, special subsidies for hospitals, insurance companies etc) -> Overprovisioning of certain medical treatments and underprovisioning of others -> Treat pain not causes -> painkiller prescriptions = overdoses

                Government subsidies of workers through special housing, food, and tax breaks subsidies, increased return on capital investment through automation at large scale -> workers marginal product reducing, but they can sort of survive on lower wages due to subsidies -> race to the bottom in wages, collapse of local economies

              • Anoneuoid says:

                Pretty much, every intervention has side effects that require further intervention then things get worse and worse until the thing collapses under its own incompetence.

              • Bill Spight says:

                A recession is defined as a consecutive two quarter drop in Gross Domestic Product. Usually the GDP is measured in currency. But we know that people are not working because of the pandemic and efforts to protect ourselves. We do not need to wait for the currency calculations. And if anybody thinks that this will not continue for more than one quarter, there is a governorship of Georgia they might be interested in.

            • jim says:

              Anoneuoid says:

              “transportation of goods -> walmart et al driving local business bankrupt”

              Why does Amazon get a pass on this all the time? It’s BIZARRE. My lib friends from school were all ready to storm wal*mart and wipe it off the planet, but they all think Amazon is like SOOO *kewel*!!!

              Wal*mart never sold below cost to run everyone out of business. Rather, they built a more efficient supply chain. Amazon, OTOH, sells AND delivers below cost, uses stock to pay execs and it’s cloud biz profit to cover it’s losses in retail! It’s like WMT on meth.

              • Anoneuoid says:

                I didn’t give amazon a pass, they are in the same category as walmart when it comes to destroying small local businesses.

              • IMHO the real evil of Amazon is how they refused to collect sales tax for about 15 years, allowing them to sell 5-10% below all local retailers for well over a decade, and then after fighting it until it was obvious they’d lose, they caved and started collecting the taxes. My personal preference is that we switch to flat income tax, UBI, and a moderate sales tax. I see flat income + sales tax as equivalent to reduced tax on capital gains, since capital gains are usually not used to purchase retail goods, but it’s much easier to administer and allows you to hike sales taxes on goods with externalities (like say gasoline, cigarettes, paint, and other pollution causing products)

                But I don’t see it as an OK thing for some companies to weasel out of taxes, while others are forced to collect them.

              • jim says:

                “But I don’t see it as an OK thing for some companies to weasel out of taxes, while others are forced to collect them.”

                Yup. I forgot about that.

              • Phil says:

                I may be the only person in America who refuses to buy from Amazon. So many of their business practices over the years have been horrible, and that’s still the case today. Plus I loathe Bezos. The remarkable thing to me is that I have many friends who agree about the horrible business practices and the horrible Bezos, but who still shop at Amazon. Certainly explains how they can continue to get away with it.

              • Rahul says:

                @Phil / Daniel etc.:

                Do you loathe Amazon alone or online stores in general?

                Is it a mom-n-pop store vs big-online-retailer debate? Or brick-n-mortar chain vs online-only-retailer? Or is it just about Amazon for you?

                I think there’s several nuances here which often get clubbed together.

                As an aside: Am in the midst of a 3 week lockdown here in India. And one that’s much more severely enforced than the western model. Amazon and other online retailers are my lifeline. They are doing a decently good job for now. Much better option than any of the local physical stores.

              • I don’t loath Amazon. quite the opposite. I think they provide a great service. I loath the fact that our politicians fail to create and enforce simple laws so that large companies can get away with the shenanigans like not collecting sales taxes. It’s politics that is loathsome.

              • Phil says:

                Most of my issues with Amazon are Amazon-specific. Or, rather, other companies may do the same things I object to, but I don’t think there’s any other major companies that does it all. Not everything on the ‘criticism of Amazon’ Wikipedia page is something I find objectionable, but there’s plenty of stuff that I do think is awful.

                Perhaps the most recent news, which I don’t even see on that Wikipedia page, is that Amazon initially offered their seasonal workers sick leave if they contracted coronavirus…but only if the public, or other employees, donated the money to pay for it. Check Snopes if you don’t believe it.

                I’m not trying to convince anyone reading this that Amazon is horrible, that’s not my point. My point is that I know many people who agree with me that they’re horrible, but who patronize them anyway.

      • jim says:

        Rahul says:

        “How do you know you won’t be having that problem? Just curious.”

        The big reason is the collapsing value of labor.

  11. Kaiser says:

    The line “respondents do not often read questionnaire instructions” triggered me. Too few understand the pain of data collection enough to have pity on us statisticians. I did a survey of students once in an exercise to estimate the class perception of price inflation. There are questions asking respondents to estimate the percent change in price of say milk in the last 3 months. Anticipating the nightmare of getting answers in all kinds of formats (10, 10%, 10 percent, 0.1, etc.), I wrote down, right next to the empty box, a note, something like “Enter whole numbers between 0 and 100; e.g. 10 for 10 percent.” Well, I got 10%, 0.1, and everything in between. If only they knew how angry this makes the data analyst!

    • Martha (Smith) says:

      A good survey needs to have a “trial run” to refine the questions to avoid problems like yours (Yes, you did anticipate a problem, but your solution also had a problem). In this case, I think it would have been better to phrase the question as something like,

      “Which option below best fits your estimate of the percent change in price of milk in the last 3 months?”, with options that cover a reasonable interval.

      (See for more discussion of wording questions and the work that needs to go into this.)

  12. jrkrideau says:

    Too few understand the pain of data collection

    My former organization reported that, back in the 1980s, we had a canoe sink during a survey in Canada but the staff were rescued and managed to complete their work.

  13. David says:

    U can’t get a good count with all the illegal Mexicans so why even try to do a census

    • Andrew says:


      The census misses a lot of people but they still have to try, partly for legal reasons and partly because it helps with estimates and imputations to try as hard as you can to count everyone.

  14. Dietmar says:

    Any thoughts on current developments? Why are statisticians not on the barricades?

Leave a Reply