For it’s Harvard this, an’ Harvard that, an’ “Give the debt the boot!”
But it’s “Academic kerfuffle,” when the guns begin to shoot.
— Rudyard Kipling, American Economic Review, May, 1890.
Remember the “Excel error”? This was the econ paper from 2010 by Reinhart and Rogoff, where it turned out they completely garbled their analysis by accidentally shifting a column in their Excel table, and then it took years for it to all come out? And this was no silly Psychological Science / PPNAS bit of NPR and Gladwell bait; it was a serious article with policy relevance.
At the time this story blew up, I had some sympathy for Reinhart and Rogoff. Nobody suggested that they’d garbled their data on purpose. Even aside from that, the data analysis does not seem to have been so great (see here for some discussion), but lots of social scientists are not so great with statistics, and even if you disagree with Reinhart and Rogoff’s policy recommendations, you have to give them credit for attacking a live research problem. In this post from 2013, I criticized Reinhart and Rogoff for not admitting they’d messed up (“I recommend they start by admitting their error and then going on from there. I think they should also thank Herndon, Ash, and Pollin for finding multiple errors in their paper. Admit it and move forward.”), while at the same time recognizing that researchers are not trained to admit error. I was disappointed with the behavior of the authors of that paper after they were confronted with their errors, but I was not surprised or very annoyed.
But then I read this post by Gary Smith and now I am kinda mad. Smith writes:
In 2010, two Harvard professors, Carmen Reinhart and Ken Rogoff, published a paper in the American Economic Review, one of the world’s most-respected economics journals, arguing that when the ratio of a nation’s federal debt to its GDP rises above a 90% tipping point, the nation is likely to slide into an economic recession. . . .
Reinhart/Rogoff had made a spreadsheet error that omitted five countries (Australia, Austria, Belgium, Canada, and Denmark). Three of these countries had experienced debt/GDP ratios above 90% and all three had positive growth rates during those years. In addition, some data for Australia (1946–50), Canada (1946–50), and New Zealand (1946–49) are available, but were inexplicably not included in the Reinhart/Rogoff calculations.
The New Zealand omission was particularly important because these were four of the five years when New Zealand’s debt/GDP ratio was above 90%. Looking at all five years, the average GDP growth rate was 2.6%. With four of the five years excluded, New Zealand’s growth rate during the remaining high-debt year was a calamitous -7.6%.
There was also unusual averaging. . . . The bottom line is that Reinhart and Rogoff reported that the overall average GDP growth rate in high-debt years was a recessionary -0.1% but if we fix the above problems, the average is 2.2%.
That part of the story I’d heard before. But then there was this:
In a 2013 New York Times opinion piece, Reinhart and Rogoff dismissed the criticism of their study as “academic kerfuffle.”
C’mon. You are two Harvard professors; you published an article in an academic journal, leveraged the reputation of academic economics to make policy recommendations to the U.S. congress, and then you talk about “academic kerfuffle”! If you don’t want “academic kerfuffle,” maybe you should just write op-eds, maybe start a radio call-in show, etc.
It’s Harvard this, an’ Harvard that, when all is going well. But then when some pesky students and faculty at faculty at the University of Massachusetts check your data and find that you screwed everything up, then it’s academic kerfuffle!
UMass, can you imagine? The nerve of those people!
So, yeah, now I’m annoyed at Reinhart and Rogoff. If you don’t like academic kerfuffle, get out of goddam academia already. For a pair of decorated Harvard professors to dismiss serious criticism as “kerfuffle”—that’s a disgrace. It was a disgrace in 2013 and it remains a disgrace until they apologize for this anti-scientific, anti-scholarly attitude.
P.S. Just for some perspective on the way that work had been hyped, here’s a NYT puff piece on Reinhart and Rogoff from 2010:
Like a pair of financial sleuths, Ms. Reinhart and her collaborator from Harvard, Kenneth S. Rogoff, have spent years investigating wreckage scattered across documents from nearly a millennium of economic crises and collapses. They have wandered the basements of rare-book libraries, riffled through monks’ yellowed journals and begged central banks worldwide for centuries-old debt records. And they have manually entered their findings, digit by digit, into one of the biggest spreadsheets you’ve ever seen.
OK, you can’t fault the Times for a puff that appeared nearly three years before the error was reported. Still, it’s kinda funny that, of all things, they were praising those researchers for . . . their spreadsheet!
I’m not sure. I tried to find the first R&R response in the Times but it is gated and I don’t have access. I did find a copy of their expanded response a day later (https://archive.nytimes.com/www.nytimes.com/interactive/2013/04/17/business/17economix-response.html) in which they appear to take the criticism seriously. So, perhaps the kerfuffle was a knee-jerk response. Lamentable but maybe not deserving to too much scorn. If someone can find the original response (from the prior day), I’d like to see that.
I’m not disagreeing with Andrew’s take on this. But I’ve started to become super-wary of any purported “factual” statements I read. Too often these are either made up or taken out of context. The example this week was the “manifesto” by the accused assassin in the UnitedHealth case. Some independent journalist released what he said was the contents of that document – and it very clearly shows the motive for the killing. Some media reported that, most did not. The police have not said much and certainly have not addressed that released document. I frankly don’t know whether it was real or not. You can read the actual document that the independent journalist released, but who knows if it is real. And it seems strange that the media are not reporting that release, even if they distance themselves from confirming it – they can always say it is unverified. I suppose this could be defended as responsible journalism (to not report it until it is officially confirmed), but it doesn’t prevent the rumors or misinformation from running rampant.
All of this is to say that I find it difficult to believe anything any more. I know that R&R made those mistakes, I know that they stood by their conclusions while acknowledging the errors, I know that the ultimate issue of whether there is a trigger point in the debt/income ratio is a matter of legitimate dispute, but I don’t know how much they trivialized the criticism given what I was able to find.
See “Debt, Growth, and the Austerity Debate” by Reinhart and Rogoff, April 25, 2013 in the NYT. The final sentence in that article is “This time is no different, and the latest academic kerfuffle should not divert our attention from that fact.”
That too was gated, but I found it elsewhere. They did indeed have that sentence. I’ll wait for Paul Alper to weigh in, but the definitions I’ve seen of “kerfuffle” do not include anything that suggests that these disputes are trivial, only that they are debates. In many ways, Reinhart and Rogoff seem to have taken the critiques seriously. They do show that nasty habit of economists to never back down on their initial results (at least they do try to defend them). Just for the record, I don’t find this kind of work compelling – international comparisons of macroeconomic aggregates involve too many confounding and endogenous features for my tastes. Having a handful of countries with high debt burdens entails a handful of countries with all sorts of different features aside from their debt burdens – it’s hard to “adjust” for them all and I think the results are not convincing.
Yeah, it’s super low N. And with a lot of correlated inputs–after all there’s only one timeline for the universe. That in addition to the confounders. It’s sort of similar to the Freakanomics boys with their abortion analysis.
The biggest problem I have with Economics and debt is that it seems to be a fringe idea that the national debt is the source of money in the economy. This is the key insight of so called “modern monetary theory” but is apparently considered a kind of fringe theory. It’s not even clear to me what the modern standard Economics take is on money. It appears that a lot of economists think that the government borrows money when it sells bonds to banks. Which is stupid.
This stuff is pretty much trivial, and yet somehow controversial? Money supply is approximately the sum of all checking accounts. When a bank buys a bond from the govt, it doesn’t reduce the checking accounts of any citizens. Then, when the government spends money its bond sale authorizes it to spend, it does increase the checking accounts of whoever it buys goods or services from. So money supply goes up. It goes up by exactly the face value of the bond, so debt = money (to first order, there’s some issues when individuals buy bonds or when currency is printed, or when currency is burned up in a fire or money goes overseas etc but these are a small fraction of the reality)
So from a theoretical perspective, when debt goes high as a fraction of GDP, and GDP is basically money supply times average velocity of money, then we have:
debt is basically money,
D = M and we also have
GDP = M * V and we claim under some circumstances D/GDP is large, so that’s saying M/GDP is large but solving the previous equation for M/GDP that’s equal to 1/V so V is small.
In order for GDP to grow at some fractional rate, we have
1/GDP * d GDP/dt = 1/(MV) * (dM/dt * V + dV/dt *M) = 1/M * dM/dt + 1/V dV/dt
So we can grow GDP by deficit spending at a high rate and increasing M rapidly, or by increasing transactional velocity. Since V is presumably small by earlier analysis, 1/V is large, and even small changes in velocity dV/dt could produce growth. It gets more complicated if you start talking about “real GDP” because that’s GDP/P where P is some price aggregate statistics, but the effect of deficit spending and creating M on P is quite dependent on what the government is spending its money on. If it’s on fighter jets and foreign wars, it might take a while before anything appears in consumer price indexes. So there’s a whole feedback mechanism which is probably best modeled as either a delay differential equation or a PDE.
Of course, M is controlled by two sources, deficit spending, and bank lending, both of which create new M. So another mechanism for GDP growth is for banks to print a bunch of money, which would likely increase both M and V.
So Reinhart and Rogoff could have sat down and done this basic analysis, and then looked at how money supply, velocity of money, and price indexes all changed dynamically through time… or they could do some stupid linear regression over the wrong columns in their spreadsheet… but even if they’d done the right columns in their spreadsheet, they’d be really no closer to having a clue without dynamics.
I have no idea whatsoever why Dale Lehman mentioned me to weigh in–we have never met or spoken to each other although I suspect we live in the same midwest (currently freezing) metropolitan area–but, as it happens, I was already busy finding this:
https://www.merriam-webster.com/dictionary/kerfuffle
——————————————————————————————-
The Evolution of Kerfuffle
Fuffle is an old Scottish verb that means “to muss” or “to throw into disarray”—in other words, to (literally) ruffle someone’s (figurative) feathers. The addition of car-, possibly from a Scottish Gaelic word meaning “wrong” or “awkward,” didn’t change its meaning much. In the 19th century carfuffle, with its variant curfuffle, became a noun, which in the 20th century was embraced by a broader population of English speakers and standardized to kerfuffle, referring to a more figurative feather-ruffling. There is some kerfuffle among language historians over how the altered spelling came to be favored. One theory holds that it might have been influenced by onomatopoeic words like kerplunk that imitate the sound of a falling object hitting a surface.”
———————————————————————————————-
More to the point, I thought the Reinhart/Rogoff embarrassing debacle was settled quite a while ago even though Reinhart/Rogoff might want to suggest that despite any pesky/embarrassing computer errors, their assertions/recommendations are still valid.
Paul
We indeed live in the same area and have never met. But you are often the source of grammatical correctness or advice, so I defer to you regarding the proper usage of the word “kerfuffle.” From what you cite, it does not appear to trivialize anything – but the sound of the word might suggest that.
Daniel wrote:
> national debt is the source of money in the economy.
I’m not sure what that means. But, I think you are focusing too much on
money.
David, rather than continue here, as Andrew has requested I’ve responded to you offsite here: https://mastodon.sdf.org/@dlakelan/113658038140460047
” a lot of economists think that the government borrows money when it sells bonds to banks. Which is stupid.”
hmmm … the government accepts money in exchange for repayment later with interest … if only there was a word for renting money that way …
No, the bank doesn’t give the government money. because money is ‘stuff you can make final payment of goods and services with’, it’s, more or less, the sum of all the checking account balances. When there’s an exchange between a bank and the govt, the bank gets a license to request reserve from The Fed (aka a govt bond) and the govt gets “reserve” a quantity that isn’t usable for purchasing final goods and services (at least not usable by the bank, the govt is the only entity allowed to “spend reserve” if you want to think of it like that). The total quantity of money (sum of all checking accounts) doesn’t change AT ALL. What changes is bank liquidity, the ability of banks to handle imbalances of payments between them. So now if Wells Fargo bought the bond, Wells Fargo has an issue if suddenly its customers all want to buy something from a company that holds its money at Chase. If that occurs, the bank may go with its license to get new reserves, and sell the bond to The Fed. The Fed will just magically increase the banks reserve.
the purpose of the bond is really to launder the fact that the govt wants to spend money, so it just has The Fed make money out of thin air. It launders this process through the banks, and it pays interest to the banks for the privilege of this laundering.
Of course, if the bank doesn’t need more reserve, it can hold the bond, and for this very important work of doing absolutely nothing at all except keeping track of where this bond is… it will get interest payments which it can pay out to its shareholders thereby increasing the money supply by making the shareholders richer, for having done basically nothing.
What also doesn’t change when the bank buys the bond, is the lending power of the bank, because they can lend against that bond as much as they can lend against the “reserve” they gave up. The rules say they can more or less use the bond as lendable capital.
Once the bank has laundered some money flow from The Fed to its internal coffers, the rules of printing money have been satisfied, and the govt pays say a govt employee, or writes out a social security check or whatever, suddenly someone deposits that in their bank account, and at that point the money supply goes up. That’s “printing money”
So, deficit spending prints money, with this weird intermediary laundering process where the bank owners can either increase their private money balances by waiting for interest payments if they can just hold on to this magical pointless bond, or they can if they have trouble with liquidity exchange the bond for reserve at The Fed. If they can just wait for interest rates to drop, they can get even more reserve, which allows them to print even more money into existence via loans.
So, here’s how it works: when the govt spends money it prints money. When the govt taxes money, it destroys money. There’s a magical laundering procedure for printing money where there’s a delay between when the govt spends the money into existence, and when The Fed acknowledges that the govt spent money into existence. In the interim between this, there’s a token called a bond which keeps track of the fact that The Fed has to print some money.
That just shows the borrowing has a purpose substantially different from say everyday consumer lending. Literally every economist I’ve ever met knew this.
Notice how absolutely nothing changes about this system if the Govt just were allowed to “sell the bonds to the fed” directly… EXCEPT the bank account balances of the bankers.
Since the system would work as well without them, the function of the bank is purely to collect rent on the license they have to be the only ones who can sell bonds to the fed.
It’s pure bullshit. The banks serve no economic purpose in this particular path.
“published a paper in the American Economic Review, one of the world’s most-respected economics journals”
It was published in Papers & Proceedings, which is not peer reviewed. Economics has a tradition of sharing works in progress, and it’s not sensible to hold those to the same standards as published papers. It was pretty clear that policymakers were not persuaded by this, but simply leaned on the findings to support the policies they had already favored.
https://www.aeaweb.org/journals/pandp/editorial-policy
[quote]
The AEA Papers and Proceedings (P&P) includes thirty-five sessions selected from the AEA program at the ASSA Annual Meeting.
Twenty-eight are chosen by the president-elect of the AEA in consultation with the AEA Program Committee and are intended to illuminate the current state of research and thinking in economics. Selections always include a mix of invited and submitted sessions as well as the AEA Distinguished Lecture.
Seven additional sessions are selected by other AEA committees from the sessions they organize: the Committee on Economic Education, the Committee on Economic Statistics, and the Committee on the Status of LGBTQ+ Individuals in the Economics Profession (one session each) and the Committee on the Status of Women in the Economics Profession and the Committee on the Status of Minority Groups in the Economics Profession (two sessions each).
The papers are edited for logic and clarity but are not subjected to a formal refereeing process. In rare cases, the editors have refused to publish papers that fail to meet AEA standards.
[/quote]
Here is some language from Gary Smith’s column: “On April 5, 2011, Reinhart and Rogoff testified before 40 US Senators and urged an “adjustment” to cut government debt. When Senator Johnny Isakson asked, “Do we need to act this year? Is it better to act quickly?,” Rogoff responded, “Absolutely. Not acting moves the risk closer.… You have very few levers at this point.””
John:
B-b-b-but . . . Rogoff used to be a very good chess player, as we kept being informed. No way he could’ve been wrong about that, right?
Just too bad we don’t have Bobby Fischer around anymore to be giving us policy prescriptions.
Economist:
OK, but . . . here’s a public-facing article from 2010 by Reinhart and Rogoff, where they write:
And here’s the reference at the end of that article:
Nothing about “Papers and Proceedings,” which is fine with me—I recently published a paper in Nature, it was in the news section not the peer-reviewed section but it was still in Nature, and anyway I think peer review is overrated. My only point here is that this was being reported even by the authors as an American Economic Review paper so I think it does inherit, fairly or not, some of the reputation of the journal more generally.
I think sex is overrated too.
https://www.youtube.com/watch?v=byEGjLU2egA
Oh wait. I mean peer review. And uh…perfessors ;-)
But seriously, I think it’s good to have some editorial filter. I see a lot of crap from the conspiratorial right (and I’m sympathetic since I’m from that group too) that is totally unreadable. Like they don’t even use citations. In that vein, even a PNAS non-reviewed paper is useful in that at least it’s cleaned up. Also, in that it’s archived and citable (as opposed to blog posts that can disappear).
My recollection is that the Excel error made the smallest contribution out of the three differences with the Amherst paper. The larger ones were choices about which data to include and how to bin it. But the Excel bit was the easiest to explain to others, so that’s what people mistakenly remember as important. As Simon Wren-Lewis put it “The Reinhart and Rogoff spreadsheet error had a relatively minor impact on its own – the really important issues lay elsewhere.”
Your own link says “The Microsoft Excel mistakes commanded a lot of attention in how news media reported the failure to replicate their AER findings, but the case exclusion decisions are the bigger deal here.”
I would actually say that the biggest issue is something only some folks like Miles Kimball have discussed: their approach could only have to establish correlation rather than causality, and there was very good reason to believe backwards-causality applied from poor growth to debt.
This also has relevance to the replication crisis:
One formidable obstacle to retesting many economic theories is that we may have to wait several years for new data to accumulate. The Reinhart/Rogoff study used data through 2009. We now have 14 years of additional data. In the US, the debt/GDP ratio went above 90% in 2010 and is currently 122. The overall ratio of central government debt to GDP for 38 OECD countries crossed 90% in 2010 and kept going, never again falling below 90%. According to the tipping-point argument, we should have had a worldwide economic recession since 2010. We haven’t, because the tipping-point argument is bogus.
There was a COVID-19 recession that began in February 2020 and was over by May 2020 in the United States and somewhat later in many other countries. The surge in the debt/GDP ratio certainly did not cause the COVID recession. On the contrary, the Great Lockdown coupled with government efforts to keep their economies afloat caused a sharp increase in the debt/GDP ratio—which then fell somewhat after the lockdown ended and government support programs waned.
The post-2009 data clearly disprove the Reinhart/Rogoff assertion. First, the worldwide GDP ratio has been above 90% for many years without causing a worldwide recession. Second, the large spike in the debt/GDP ratio in 2020 and 2021 demonstrates that causality often runs in the other direction: fluctuations in the debt/GDP ratio are not the cause of business cycles but the result of them.
Gary:
Reinhart and Rogoff might agree with you on that, at least regarding the covid recession. Here’s a news article from 2020 quoting Rogoff:
I’d like to know more about those monks keeping records of national debt. Surely that’s a matter for a court minister, not the kind of thing you’d outsource to a monastery? It’s even more baffling, because the R&R paper only goes back 200 years, by which time monastic manuscripts were an antiquity. I think maybe the NYT reporter was embellishing things.
Embellishments in the newspaper? Impossible! Anyway, as far as these monks are concerned, they probably did not keep track of the national debt. But that was not the claim stated in the article. In the Middle Ages, monasteries served as credit institutions for rulers. The aristocratic leaders of western and central Europe needed a lot of money to stay solvent – their personal finances and the government purse were usually not separated. Banks were an invention of the Renaissance, when people like the Fugger family came on the scene. Their financial power dwarfed that of the monasteries; IIRC, at the height of their power, the Fuggers had more wealth than all the middle European monasteries combined and then some. Capitalism and banking, my history book said, started with these guys*. Anyway, I imagine that economic historians would use these old records to estimate the level of government debt. It is far from an exact measurement, but I bet it is some of the best data from a time before tariffs and statistics offices.
*This is a somewhat controversial statement – I genuinely do not care about the starting date of capitalism.
I never felt sorry for them and I don’t now. They are incompetent if they can’t check the analysis of their spreadsheet or code and ensure the data are correct and the results include all the data as they are supposed to. All you have to do is count the number of rows or columns!!! Amazing. And they are sloppy *and* incompetent if they don’t even attempt to set up cross-checks in their results to ensure all the data are included.
Most academic research in the social sciences isn’t real research at all, it’s just a bunch of peacocks strutting around flaring their tails at one another and society in general – social signalling, top to bottom, from research design to the (usually) foregone conclusions.
In a way, the Reinhart and Rogoff debacle is benign. There was no intent to falsify the (spreadsheet) data. This is in marked contrast to what is going on in cancer research where “blots” are being manipulated via use of Photoshop.
On the contrary, the largest difference between R&R vs Amherst was NOT the accidental Excel error, but deliberate choices made in the analysis.
+1 to Dan Lakeland. Mainline economics seems, from the outside, to not have a satisfactory account of money and I’m given to believe that neoclassical economics basically treats it as a nuisance factor. I once had an Econ PhD try to explain M1, M2, M3 or whatever to me in the context of an argument about the definition and causes of inflation but it ended up being pretty tangential so I didn’t dig deeply into it. I just can’t take anything from neoclassical macroeconomics seriously at this point…
There’s insights from Steve Keen’s work: credit (defined as growth in commercial lending) is a factor in aggregate demand. Deficit spending is of course, also a factor in aggregate demand. Both processes add money to the economy, and generate velocity, deficits however, do not come with a debt obligation. There is no reason to pay it off. The federal “debt” is the money supply.
Commercial money, what we call loans, do come with a debt obligation, but there is another distinction from deficit spending, it also comes with an interest charge. Following the doubling rule, a mortgage of 7% means that a 10 year mortgage cost the borrower twice the price of the house, and a 30 year 4x. The money created by the loan, which is the principle is destroyed as a loan is paid, but the interest charge has to come from somewhere else.
That means either it comes from a continuous expansion of debt obligations in the economy, i.e. debt expands and/or the money supply expands (deficit spending). Because banks will curtail credit if economic prospects decline that automatically means that they can generate a debt deflation spiral because of their own caution. The buffer to this is deficit spending (if you read the circuit theory of money from Graziani, you’ll also pull out other little nuggets like deficit finance is cheaper because it doesn’t incur that interest charge, which IIRC is applied twice on inflation) But I think Graziani was at the forefront defining one leg of MMT, which is credit furnishes bank deposits. Lending comes first and its comes ex-Nihil.
Further, the larger the banking institution, the less their lending operations are constrained by “reserves” or flows. This is because the money tends to circulate endogenously within the bank’s own ledgers; consider this thought experiment: a nation with a single bank. When it creates money, the money simply goes from one account to another within its own ledger, and there is no external obligation created. It has unlimited lending potential, only constrained by ability of borrower to get money in their account to pay back their obligation.
I am a little rusty here, Keen’s other insight was when credit fades with respect to GDP, some ratio, that triggers a financial crisis, like GFC. So the significant factor is that it’s not so much the debt level it’s the plunge in credit, lending, that triggers a crisis. The problem with a glut of credit is its inevitable tapering.
He serialized his work on this topic in his sub stack for his forthcoming book where he covers all the math for this as well as the models and people can look it up there. It’s free to view.
* I get really salty about deficit hawks because they are never challenged with the question of where the money the government is supposedly borrowing comes from. The US the UK Japan, China, all issue and spend in their own money. The state is the only source of the currency, and its provision is the money supply. I want to see this challenge tested on a mainstreamer. Most of them are familiar with Milton Friedman statement about inflation being everywhere a monetary phenomenon. If they agree with that statement, they agree that the government originates the money supply.
Asserting that the government has to borrow the currency and issues is ridiculous as asserting that an apple farmer would have to borrow applesauce in order to grow apples.
> It appears that a lot of economists think that the government borrows money when it sells bonds to banks. Which is stupid.
> When a bank buys a bond from the govt, it doesn’t reduce the checking accounts of any citizens. Then, when the government spends money its bond sale authorizes it to spend, it does increase the checking accounts of whoever it buys goods or services from. So money supply goes up.
Do you think that is controversial? When an individual borrows money from a bank the banks reserves go down and that money goes into the money supply. The same thing happens when the US Treasury sells a bond to a bank, in a first step the “loan” doesn’t enter the money supply (the transfer goes from the bank’s reserves to the Us Treasury) but it does as soon as the government spends the money. Where’s the stupidity in considering that the government is spending borrowed money – just like any person or company would do if they had borrowed money from a bank?
Carlos, a bit tangential maybe but I often find hidden in these discussions a lack of agreement about things like the ‘loanable funds’ or ‘money multiplier’ models. To which I think this view from the Bank of England is a much more accurate description of money supply and creation:
https://www.bankofengland.co.uk/-/media/boe/files/quarterly-bulletin/2014/money-creation-in-the-modern-economy.pdf
I’m not sure how that article relates to my comment – if at all.
> Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower’s bank account, thereby creating new money.
Yes, I don’t think there is much controversy around that. When an individual borrows money from a bank the money supply increases.
When the US Treasury borrows money from a bank the money supply doesn’t increase immediately (money hold by the Treasury is not counted, just like banking system reserves are not counted). But when that money is spent – which is the reason why it was borrowed in the first place – it increases the money supply.
The point is that the US Treasury doesn’t “borrow” from banks. That’s the controversial bit.
What is it “borrowing”? Note, as I said before, if the US Treasury “sold it’s bond” direct to the fed the system would work basically the same except the banks wouldn’t be collecting interest.
What scarce resource or service is the bank providing? Nothing. You could say that the Treasury is “borrowing reserve” except reserve isn’t a scarce resource and can be and IS manufactured by the fed in unlimited quantities at no cost.
What the Treasury is doing, is laundering the creation of money. It’s able to pretend its borrowing things and that banks are providing a service. And it’s able to funnel wealth to the banking industry as pure rent. This has political value perhaps, but serves no economic purpose at all.
The Treasury “prints” every dollar it spends, it just does so through a convoluted process involving banks collecting rents on holding a token that says “some day, eventually go over to the fed window and acknowledge that we manufactured this much money”
Daniel
I’m not sure I follow. When the Treasury sells a bond and a bank buys it, money is transferred out of the bank’s account at the FED (at least that is my understanding). The usual textbook story is that this limits the bank’s ability to make loans, thereby contracting the money supply. It works in reverse when the Treasury buys bonds from banks. If the sale involved the FED rather than the banks, I don’t think it would work the same, since the FED does not make direct loans to the public. The money multiplier effect takes place through the bank’s ability to use fractional reserve banking.
Of course, that is the textbook version, and rarely (if ever) correct. I see two main problems with the textbook story. First, the neat mathematics of the money multiplier assumes that banks always operate at the minimum reserve ratio required by regulation. There is a long history (at least back to Keynes) of issues with that assumptions – which I have always found appealing. Bank lending depends on many things, especially the state of the economy and riskiness of potential borrowers. So, banks have reasons to not lend out all that they can, particularly when the economy is weak (which is when you would want to have the FED expand the money supply). “Animal spirits” play a major role in bank lending behavior, and economists have never had good explanations for animal spirits in my opinion.
The second issue is that this textbook version of money is quite narrow. Increasingly, people have a variety of ways to make purchases without using money actually sitting in bank accounts. So, if they are optimistic, they might increase their use of their many credit cards, running up balances that don’t get paid off at the end of the month. They have purchased things in the same way as if they borrowed from the bank through the textbook loan, but with relatively little control by the FED. They might even expand the money supply (broad version) like this when the FED may be trying to constrain it. There are now so many ways to access funds outside of the banking system that I have problems with the textbook definitions of the money supply – it is so narrow that it misses much economic activity.
I stand to be corrected on the above. While I was trained as an economist, I understand micro but have always considered macro the black arts. I even refused to teach introductory macro (and once was denied promotion for that) given the poor job I would have done. But I have trouble reconciling your explanation with the textbook version because I don’t see the FED playing the same role as the banks in the story.
My bad, I didn’t understand that the stupidity of thinking that the government borrows money when it sells bonds to banks was about the banks selling those bonds afterwards.
By that token people thinking that people taking mortgages are borrowing money becomes stupid when loans are packaged in securities and the fed buys them.
Carlos, being mindful of Andrews desire that the comment section not get too far into the weeds I’ll try to summarize my overall concern and then leave it at that. yes I do think there’s controversy. For one thing, the Bank of England paper Chris Wilson posted has generated a considerable amount of controversy, and from what I’ve read the “standard” economic take is that government borrows from banks, and that banks intermediate loans between people who want to lend and people who want to borrow, rather than manufactures money. There’s even a current paper from this year basically trying to refute the BofE take https://www.cato.org/sites/cato.org/files/2024-03/working-paper-80.pdf
I’m far from an expert on what the standard economic take is and I haven’t read a survey but I’ve read some of the controversies and I’ve seen the responses to the BofE paper by people who appear to have “standard” economic training. BofE/MMT view of how money comes about also disagrees with all the models described in Mankiw’s standard macroecon text. I’ve never seen a standard macroecon text that tells an accurate story of money, and if there were one I imagine people like Steve Keene would be pointing to it to bolster acceptance of their alternative takes. Also, money used to work differently back on the golf standard… Sort of.
So, how do I bring all this back to the original topic?
Apparently, the field doesn’t have a clear picture of what government debt even is, of the role of banking in the money supply, or of how GDP comes about, or have a description of the dynamics of money, debt, and inflation which could describe how a 90% “boundary layer” which is basically what R&R were claiming would work or come about, or a mechanism for why it would occur at 90% or at any other specific value of debt/GDP. Furthermore it seems impossible that such a thing could be explained because debt/GDP is not dimensionless, it has dimensions of time. Whatever the real cause of such a boundary layer is, I’m pretty sure it’s not dependent on for example the cesium atomic clock, so there’s a missing factor in the determinant. I suppose you might argue that the seasonal cycles of the year are fundamental to this boundary layer but if so that should be explicit.
Given the depth of the problems, How can I take this entire topic seriously?
On the other hand, Sure, maybe “most economists” don’t agree with R&R, great, so then why are we still talking about them? why are they not laughing stocks and why does economics still rely on this “style” of research in other fields? How is it that they get the seat at the table of govt decision makers? Is it because the purpose of Macroeconomics is politically to launder the preferred choices of elites through a veil of scientificness? That’s my view having seen sooo many ways in which macroecon appears to fit that take.
My current favorite person to read about economics is Blair Fix, because not only does he go full heterodox but he offers his analyses almost completely transparently online and they seem vastly more self consistent than anything I’ve read in a macroecon text. He’s got a PhD in econ but he’s rejected the standard explanations in his field and has alternative ideas.
https://economicsfromthetopdown.com/2024/08/22/from-commodity-to-asset-the-truth-behind-rising-house-prices/
https://economicsfromthetopdown.com/2023/02/04/do-high-interest-rates-reduce-inflation-a-test-of-monetary-faith/
https://economicsfromthetopdown.com/archives/
Almost surely I’ll be in moderation for the links so hopefully this post appears eventually
Carlos, selling to an entity that manufactures a fictional quantity at will (the fed) is not the same as selling to an entity that gives up something finite it holds in exchange. Unless you think borrowing from an infinite pool whose quantity is unchanged by the loan is the same as borrowing from the cash in my pocket.
Dale, when a bank gives “reserve” (which is not part of the money supply and thus shouldn’t be called “money”) to the treasury in exchange for a bond, its lending power is hardly affected at all. that bond has market value that it can use as part of its capitalization to argue that it should be allowed to lend. In fact, just recently The Fed bailed out a few banks that got stuck with too much govt bonds and when they raised the interest rates, the bonds market value collapsed. So The Fed bought those bonds at PAR value manufacturing reserve for them. That happened just in the last year or so. The rules of whether a bank is allowed to lend have gotten quite complicated, they involve “stress tests” but there is *no* reserve requirement anymore. That went to 0 in 2020.
Banks hold reserve because they need to deal with imbalances in transactions and exchange this reserve between the banks. As G says in a post elsewhere, if there were just 1 bank, it could lend infinitely. There’s a reason banks have been consolidating through time. Big whales face far less lending restrictions, particularly when they’re too big to fail.
And the point is that renting money for a period of time for a fee is the literal definition of “borrowing”.
That the aim of the fed is different from me charging my credit card to buy donuts is standard elementary knowledge, taught to literally every economist and finance student, and is on about the same level as F=ma.
You don’t have some secret knowledge that the “stupid” economists lack.
ok, enough on all this . . . we’re getting deep into what Phil calls “garbage time” here.
I haven’t eaten a donut since like 3rd grade, so had no idea how much they cost.
It came up in convo yesterday and I was shocked at the price. It was $10 plus taxes/fees/tip for a half-dozen donuts.
So like $2-3 per donut. Thats a gallon of gasoline per donut.
Andrew, agreed we’re deep in the weeds, apologies for having it go this far. anyone who wants to engage in questions of the details of money and debt can discuss further off-site here:
https://mastodon.sdf.org/@dlakelan/113646582500534284
Dale, you mentioned you created a Mastodon account. You can go to that link and click the reply arrow below one of the posts and type replies that way. In future if you just want to send public messages to me, just typing in my account ID in the message means I’ll see them. There are about 5-6 other regular readers on this blog who I’m connected to via Mastodon as well.
(1) The May issue of the AER is the Papers and Proceedings. If people write out the full citation correctly it is there. If not, then folks know the May issue is not refereed.
(2) The comments here illustrate a huge problem with social-science research that has little to do with statistics. Lots of people opining without having read the paper in question. Lots more saying “economics does” or “economics doesn’t” without the faintest idea of what, in fact, economists think or do.
(3) These comments all miss a different kind of arrogance: RR claimed to use the lessons of history to warn about the future. Yet they not only ignored the history, they ignored most of the published work by the people who work in that field. Some of their comments reflect an a priori contempt for work they never read. To me, that shows the nature of the exercise. The data were irrelevant, however handled; RR knew the answer going in. That’s why egregious error was just a “kerfuffle.”
Alfred:
That’s funny–I’d never heard of this “May issue” thing before. It’s a kind of insider code so that economists can signal to each other that the paper is not refereed, but outsiders will have no idea.
I altered the Kipling attribution at the top of the above post to account for this.
To be fair, the May issue has a different title from the other issues. But it is printed in the same format, so one would be forgiven for thinking they are all the same. If you really want to be careful, I would not say the AER, I would say “The Papers and Proceedings of the American Economic Association.”
By the way, some economists put May-issue papers on their vitas in a misleading way.
‘Lots more saying “economics does” or “economics doesn’t” without the faintest idea of what, in fact, economists think or do.’
Well given that there are dozens, or maybe dozens of dozens, of economics blogs writen by economists of many different flavors, at least what some economists think is a matter of record and pretty easy to find out. And the fact that there are dozens, or maybe even dozens of dozens, of economics blogs – often with comment threads filled by the comments of dozens, or maybe even dozens of dozens, of economists, gives a pretty fair hint about what economists (and some statisticians) spend their time doing.
So, Alfred, that claim is disputable.
OTOH, I agree that many people comment on blogs – and write and publish academic papers, even in peer-reviewed journals – without apparently knowing what they are talking about. But still that’s pretty easy to do, since there is so much BS floating around in the published social science literature – including the economics literature. To some extent, most of this work could be legitimately published under the banner of “highly speculative” – except that its usually not treated as highly speculative, and it’s often so poorly characterized in terms of the research design that – as you suggested – it can be shot down without much recourse to analyazing the statistical methods.
As Andrew always says: GIGO, but i gues the caveat is his definition of G is much narrower than it oughta be.
I don’t uindI have been reading the comments. Economics has a pretty clear theory of money that all but a handful of economists agree upon. The disagreements within the profession come from the next-step questions, such as how money interacts with fiscal policy. I agree that it would be easy to learn about that from reading blogs and comments on blogs. But these comments don’t reflect that effort. This is precisely my point.
I just arrived here after reading about the black plastic kitchenware post. Hey, it doesn’t matter if the sign changed, the conclusion can stay the same.