Hi Noah, this is excellent analysis. Big ask but I'll definitely love to read one on Nigeria which I fear is toeing a similarly destructive path via high external borrowing made worse by its flailing manufacturing sector and poor revenue inflows. If you do ever get to do that piece I'm certain it would go very viral as Nigeria is approaching election season. Do consider this please.
Well your explanation of how Turkey did well and where it went wrong is all quite correct. To what went well you should add that Erdogan began his leadership career almost 20 years ago surrounding himself with non-nonsense economists and making a concerted and successful effort to tame the high inflation that was the scourge of his predecessors (and helped bring him to power). And to what went wrong you should add that the construction sector (together with the banking business financing it) became not only his and his family's pot of corruption, it became also the dominant patronage platform through which he corrupted the country's business elite and built a de facto one-party state (outside Istanbul). I dropped into Istanbul for mostly macro meetings a few times five to ten years ago, and basically everything the local finance guys were fretting at the time would go wrong has. It's very sad.
Could the inflation get worse? Yes, but it's not likely to get nearly as much worse as you're saying. I will give you that one of the most common preconditions for hyperinflation is in place: a single person in total control of both fiscal and monetary policy. Hyperinflation could in theory occur: it's completely up to Erdogan whether or not he wants to have it. All he has to do is decide to very quickly ramp up nominal public spending using central bank financing. But generally dictators only do that in the midst of a very severe economic crisis. It's essentially a despot's act of desperation, an attempt to seize as much as possible of a rapidly shrinking pie. I suppose it might be barely possible to generate hyperinflation through construction lending on an uber-astronomical scale, but that of course has never happened. A really bad case of unanchored expectations certainly can give you more than 50% inflation per year. It will not give you more than 50% inflation per month.
I could tell something was wrong with your conception of hyperinflation by your odd comments about it being a possible result of overly low rates and unanchored expectations. Does academia not teach hyperinflation at all? To get hyperinflation the authorities have to be doing something very much more aggressive than overly low rates. 50% annual is just very high inflation, in this case accentuated by the herd-of-chickens game of selling high interest local currency bonds to foreigners that might pay well or might devalue to losses. But hey at least you're not like one of those crypto peddlers claiming >5% annual is hyperinflation.
No, at this point your analysis has run badly off the rails. Unanchored expectations are the only way you can get hyperinflation. The question is what causes unanchored expectations. But tbh, unanchored expectations are probably required for 50% annualized inflation, too.
Sorry, you obviously have never studied even one case of hyperinflation and have no idea what you're talking about. Neither the private sector nor the general population have power to cause hyperinflation. There must be an authority or colluding group of authorities with unchecked power to create and spend money who must make a conscious decision to do so on an extreme scale. Period. End of story. If you believe otherwise you've been miseducated. Get back to me when you find any other kind of case. Of course that's impossible because there are none, unless you count examples where money's value suddenly shrinks at similar rates to hyperinflation through extinction of authority, e.g. the Confederacy or Hussein's Iraq.
Because I know what I'm talking about here, and telling me I'm "off the rails" because what I'm saying conflicts with one's own weak understanding of macro is way over the line.
Put another way, in all fields of scholarship there is an equivalent of the Hippocratic rule, to above all do no harm. And what that means in practice is not to pretend to know things that one doesn't know. And when people who do know, or who know that nobody knows, encounter a quack who's pretending to know, it's our duty to publicly expose and denounce that quack, lest others make the mistake of believing he or she actually knows. Noah isn't a quack, he's an academic-turned-economics-commentator, who more often than others of that type gets the big picture right. But he too often attempts to teach people things that he doesn't really know himself. And when he does that he should be scolded.
I think what's going on here is that Noah learned a primitive, closed-economy model of inflation in which the central bank accommodates a privately led credit expansion and focuses on the spiraling effects of rising inflation, falling savings rates and hoarding of goods. These are useful things to think about but very high inflation is usually fiscal (Turkey's is a mix of fiscal and pseudo-fiscal via patronage lending), and hyperinflation is always and everywhere fiscal.
You are missing the demand for loans at the current discount rate. Filling that would require the ever-accelerating printing that will drive inflation higher and higher. For the government to stop that, it has to make the discount rate "fake" and pretty much shut down the financial system and replace it with a central command system where the government decides who gets the limited supply of insanely profitable loans (can you say "crippling corruption"? I knew you could!)
Low interest rates plus high inflation will lead rapidly to hyperinflation. Technically the Turkish government could prevent hyperinflation by not actually honoring the low interest rates, but in that case they don't really have low interest rates, they have high interest rates combined with insane levels of corruption.
- Turkey imports its fuels and has a current account deficit, so its inflation is powerfully influenced by exchange rates. Despite increasing interference the lira is still a fully convertible currency, so when the CB lowers rates to negative like it did, the lira is sold off hard. So anticipated inflation gets brought forward for all imported goods, and locals dump any lira savings for dollars and euros and gold further suppressing the lira. That makes the retrospective inflation rate very different from the market-anticipated inflation rate. In other words, if you're analyzing from a closed-market model of inflation, a retrospective rate of >50% would be assumed to mean >50% anticipated inflation going forward, and the current ~17% base rate being advertised by Turkish banks for lira mortgages worth c. $100k would look like an invitation to all Turkish homeowners to rush to their banks and plunder the last scraps of an apocalyptic economy. But in reality that rate is for ideal customers buying new builds, and the lira has already been sold down enough to make the real offers not so enticiing.
- Assets are mostly priced in dollars. So whereas you might be able to pay lira for a new apartment that's not completed yet from a builder who's got credit from a bank who's got credit from the CB, to buy an existing apartment or house, if you take a lira mortgage you would have to convert into and pay in dollars. Again that market has already equilibriated.
" in that case they don't really have low interest rates, they have high interest rates combined with insane levels of corruption"
Yes of course, they've been doing this for a while now. The Erdogan family central bank is no way going to issue negative rate loans to just any yahoos who show up asking.
As for recent Turkey, private sector and general population expectations certainly matter a lot, but "unanchored" isn't really the right word for what's going on. The authorities repeatedly lure savings out of dollars into lira with high lira rates, then betray owners of lira debts and deposits with a sudden lowering of rates, which is well understood to mean another big spree on the way of patronage construction lending at negative real rates, and so with quite rational and carefully mathematically calculated expectations the lira is sold off hard.
I'd make the same comment I made about Noah's Jamaica piece: you want to look at services exports as well as goods exports, because these are highly tourism-dependent economies. (I just spent a couple of months in Turkey and if Noah thinks it was cheap a few years ago... wow, it's really cheap now. If not for COVID this might even be one of the benefits of debasing the currency.)
On an unrelated point: I'm not sure Erdogan is facing as much extraparliamentary political risk as Noah suggests. The ruling party has been very successful at breaking the autonomous power of the armed forces, and a lot of people think the "failed coup" of 2016 was really a false-flag operation intended to finish the job. Even before that, many army officers and others have been convicted in unfair trials on very doubtful charges of plotting to overthrow the government.
Yes, Erdogan is to blame for much of the problem but to not mention the disaster in Syria feels like a big omission. The millions of refugees have definitely led to economic imbalances in Turkey and Lebanon (which went through the same debt driven hyperinflation). Western countries should be sending far more aid to these two countries given their culpability in the total collapse of Syria and its roots in the Iraq war.
This is an astute analysis. What I have to add is in the spirit of extending the analysis. 1 If my memory serves me well, there was a surplus of liquidity floating around on international markets after OPEC ramped up crude as prices. I believe it lad to currency crises in Latin America, notably Argentina, 2 The post-1990 period ushered in not only the collapse of Communist central planning largely worldwide, but also catch-up growth in a host of less developed nations. It would be useful to look at a data set that includes all of the “potential growers”, sorting out winners from losers. Turkey would lie in the middle of the spectrum Imwould think. One approach would be to take as a classifying variable the growth in the share of global world trade income. One suspects the correlation between sustained per capita income and the trade success variable would be high. That would be a useful starting point for a serious aggregate economic analysis.
What specifically caused the dramatic uptick in growth in the early 2000s? I’d imagine most of the groundwork for your Chang-Studwell model had already been laid much earlier.
This is a bit unfair to Neo-Fisherism, which makes a much, much narrower claim: a widely-anticipated, slow and steady interest rate increase, with stable and neutral fiscal policy, will over time lead to a rise in the inflation rate. All known (or, at least, widely known) macroeconomic models that are long-run stable -- that is, where small monetary policy misses don't lead to inflationary or deflationary spirals -- have Neo-Fisher properties. And that's basically the only thing Neo-Fisherians we're pointing out, as part of a program to further explore all the properties of models in an ongoing effort to improve them.
But no one serious in that group of people (they I know of) actually advocates using short-term nominal interest rate decreases to reduce inflation. Actually I don't think many of them advocate any policy at all, it is a pretty academic program. I guess you can blame them for that, but not everyone can be working on policy all the time, some people need to work on the nuts and bolts of these things.
Ah, BUT, if agents know that the near-dictatorial president-for-life is going to exert a consistent downward influence on rates, then according to Neo-Fisherism that should be disinflationary even if there are transitory upward rate spikes. As seen on the graph above.
In other words, yes, Neo-Fisherism really is failing here. Now, maybe Neo-Fisherians can patch things up by adding capital inflows and an international sector to the model. But the basic idea looks dead.
Well, you would still see inflation spike initially. But in any case that's really only true if fiscal policy were stable and neutral. See Japan. But I think there is zero chance that is what is happening in Turkey. Sure, the official budget is stable, but let's be real, all the money that the central bank is printing is getting funneled to cronies. Just because you cut out the middleman of the official budget account doesn't make the transfer go away. That said, if that weren't happening, I also wouldn't be shocked if velocity crashed, with all the attendant human suffering that would entail, and then inflation fell as real rates equilibrated again.
Don't get me wrong, I think Neo-Fisherism's predictions are about as useful for public policy on a reasonable time frame as string theory predictions that there are 11 dimensions. It's something that pops out of pretty much every non-crazy macroeconomic model, and so it is pretty important to get to the bottom of it. Not because it helps immediate monetary policy, but because if it is wrong then something is fundamentally wrong with basically all of the macroeconomic models, whether they be monetarist or neo-Keynesian.
In any case, my point was there aren't really any Neo-Fisherians cheering Turkey on right now, because they all know that even if he was acting in accordance with the assumptions, and even if they are right, the adjustment mechanism that gets you there is awful. No one is advocating for what he's doing.
Excellent piece. A return to orthodoxy however would entail a wrenching economic adjustment which might topple Erdogan absurd regime. Don’t expect him to initiate it.
[offtopic] A while back you posted some responses to typical lazy econ critiques, and pointing to the supposedly 'ideologically charged' nature of economics you said that it was a problem but that could be overcome and that practitioners should try to be non-ideological. You pointed out that to that response there was this: "But dude, being anti-ideological is an ideology too!". And you followed that with: "Mmm, an insightful and trenchant observation. Now go take another bong hit and leave me alone.". I was wondering what was your personal opinions on similar positions held by serious names in the philosophy of economics. For example, Julian Reiss author of a 'textbook' introduction to the philosophy of economics and editor of the Routledge handbook of the philosophy of econ has defended in "Fact-value entanglement in positive economics" and elsewhere that economics is value-laden in important ways. Do you think the thesis is wrong, or irrelevant? I would be cool to know what a person completely embedded in econ culture has to say about that, thanks for the blog, it is a pleasure to read.
Hi Noah, this is excellent analysis. Big ask but I'll definitely love to read one on Nigeria which I fear is toeing a similarly destructive path via high external borrowing made worse by its flailing manufacturing sector and poor revenue inflows. If you do ever get to do that piece I'm certain it would go very viral as Nigeria is approaching election season. Do consider this please.
Love,
Stephen
I have a few on Nigeria:
https://yawboadu.substack.com/p/inequality-within-west-africa-and
https://yawboadu.substack.com/p/nigerias-economy-part-three-1960
Can you do this again, but for Brexit? Bad choices turn out to be bad..
Well your explanation of how Turkey did well and where it went wrong is all quite correct. To what went well you should add that Erdogan began his leadership career almost 20 years ago surrounding himself with non-nonsense economists and making a concerted and successful effort to tame the high inflation that was the scourge of his predecessors (and helped bring him to power). And to what went wrong you should add that the construction sector (together with the banking business financing it) became not only his and his family's pot of corruption, it became also the dominant patronage platform through which he corrupted the country's business elite and built a de facto one-party state (outside Istanbul). I dropped into Istanbul for mostly macro meetings a few times five to ten years ago, and basically everything the local finance guys were fretting at the time would go wrong has. It's very sad.
Could the inflation get worse? Yes, but it's not likely to get nearly as much worse as you're saying. I will give you that one of the most common preconditions for hyperinflation is in place: a single person in total control of both fiscal and monetary policy. Hyperinflation could in theory occur: it's completely up to Erdogan whether or not he wants to have it. All he has to do is decide to very quickly ramp up nominal public spending using central bank financing. But generally dictators only do that in the midst of a very severe economic crisis. It's essentially a despot's act of desperation, an attempt to seize as much as possible of a rapidly shrinking pie. I suppose it might be barely possible to generate hyperinflation through construction lending on an uber-astronomical scale, but that of course has never happened. A really bad case of unanchored expectations certainly can give you more than 50% inflation per year. It will not give you more than 50% inflation per month.
Great article! I would definitely love to read both books!
"50% inflation is definitely in the hyperinflationary range"
No. No, it is not.
Hyperinflation is typically defined as 50% or more *within a month*.
Okay now I'll read the rest.
Oh, huh. I had always thought it was one month at a 50% annual rate! I guess 50% m/m sets the bar a little higher.
I could tell something was wrong with your conception of hyperinflation by your odd comments about it being a possible result of overly low rates and unanchored expectations. Does academia not teach hyperinflation at all? To get hyperinflation the authorities have to be doing something very much more aggressive than overly low rates. 50% annual is just very high inflation, in this case accentuated by the herd-of-chickens game of selling high interest local currency bonds to foreigners that might pay well or might devalue to losses. But hey at least you're not like one of those crypto peddlers claiming >5% annual is hyperinflation.
No, at this point your analysis has run badly off the rails. Unanchored expectations are the only way you can get hyperinflation. The question is what causes unanchored expectations. But tbh, unanchored expectations are probably required for 50% annualized inflation, too.
Sorry, you obviously have never studied even one case of hyperinflation and have no idea what you're talking about. Neither the private sector nor the general population have power to cause hyperinflation. There must be an authority or colluding group of authorities with unchecked power to create and spend money who must make a conscious decision to do so on an extreme scale. Period. End of story. If you believe otherwise you've been miseducated. Get back to me when you find any other kind of case. Of course that's impossible because there are none, unless you count examples where money's value suddenly shrinks at similar rates to hyperinflation through extinction of authority, e.g. the Confederacy or Hussein's Iraq.
Why so aggressive Tom?
Because I know what I'm talking about here, and telling me I'm "off the rails" because what I'm saying conflicts with one's own weak understanding of macro is way over the line.
Put another way, in all fields of scholarship there is an equivalent of the Hippocratic rule, to above all do no harm. And what that means in practice is not to pretend to know things that one doesn't know. And when people who do know, or who know that nobody knows, encounter a quack who's pretending to know, it's our duty to publicly expose and denounce that quack, lest others make the mistake of believing he or she actually knows. Noah isn't a quack, he's an academic-turned-economics-commentator, who more often than others of that type gets the big picture right. But he too often attempts to teach people things that he doesn't really know himself. And when he does that he should be scolded.
I think what's going on here is that Noah learned a primitive, closed-economy model of inflation in which the central bank accommodates a privately led credit expansion and focuses on the spiraling effects of rising inflation, falling savings rates and hoarding of goods. These are useful things to think about but very high inflation is usually fiscal (Turkey's is a mix of fiscal and pseudo-fiscal via patronage lending), and hyperinflation is always and everywhere fiscal.
You are missing the demand for loans at the current discount rate. Filling that would require the ever-accelerating printing that will drive inflation higher and higher. For the government to stop that, it has to make the discount rate "fake" and pretty much shut down the financial system and replace it with a central command system where the government decides who gets the limited supply of insanely profitable loans (can you say "crippling corruption"? I knew you could!)
Low interest rates plus high inflation will lead rapidly to hyperinflation. Technically the Turkish government could prevent hyperinflation by not actually honoring the low interest rates, but in that case they don't really have low interest rates, they have high interest rates combined with insane levels of corruption.
Some things to bear in mind:
- Turkey imports its fuels and has a current account deficit, so its inflation is powerfully influenced by exchange rates. Despite increasing interference the lira is still a fully convertible currency, so when the CB lowers rates to negative like it did, the lira is sold off hard. So anticipated inflation gets brought forward for all imported goods, and locals dump any lira savings for dollars and euros and gold further suppressing the lira. That makes the retrospective inflation rate very different from the market-anticipated inflation rate. In other words, if you're analyzing from a closed-market model of inflation, a retrospective rate of >50% would be assumed to mean >50% anticipated inflation going forward, and the current ~17% base rate being advertised by Turkish banks for lira mortgages worth c. $100k would look like an invitation to all Turkish homeowners to rush to their banks and plunder the last scraps of an apocalyptic economy. But in reality that rate is for ideal customers buying new builds, and the lira has already been sold down enough to make the real offers not so enticiing.
- Assets are mostly priced in dollars. So whereas you might be able to pay lira for a new apartment that's not completed yet from a builder who's got credit from a bank who's got credit from the CB, to buy an existing apartment or house, if you take a lira mortgage you would have to convert into and pay in dollars. Again that market has already equilibriated.
" in that case they don't really have low interest rates, they have high interest rates combined with insane levels of corruption"
Yes of course, they've been doing this for a while now. The Erdogan family central bank is no way going to issue negative rate loans to just any yahoos who show up asking.
As for recent Turkey, private sector and general population expectations certainly matter a lot, but "unanchored" isn't really the right word for what's going on. The authorities repeatedly lure savings out of dollars into lira with high lira rates, then betray owners of lira debts and deposits with a sudden lowering of rates, which is well understood to mean another big spree on the way of patronage construction lending at negative real rates, and so with quite rational and carefully mathematically calculated expectations the lira is sold off hard.
Though come to think of it, these days it's probably main the local banks and oligarchs buying those lira bonds out of euro/dollarized savings.
I'd make the same comment I made about Noah's Jamaica piece: you want to look at services exports as well as goods exports, because these are highly tourism-dependent economies. (I just spent a couple of months in Turkey and if Noah thinks it was cheap a few years ago... wow, it's really cheap now. If not for COVID this might even be one of the benefits of debasing the currency.)
On an unrelated point: I'm not sure Erdogan is facing as much extraparliamentary political risk as Noah suggests. The ruling party has been very successful at breaking the autonomous power of the armed forces, and a lot of people think the "failed coup" of 2016 was really a false-flag operation intended to finish the job. Even before that, many army officers and others have been convicted in unfair trials on very doubtful charges of plotting to overthrow the government.
https://en.wikipedia.org/wiki/Ergenekon_trials
Yes, Erdogan is to blame for much of the problem but to not mention the disaster in Syria feels like a big omission. The millions of refugees have definitely led to economic imbalances in Turkey and Lebanon (which went through the same debt driven hyperinflation). Western countries should be sending far more aid to these two countries given their culpability in the total collapse of Syria and its roots in the Iraq war.
I'm surprised there's no mention of https://www.bloomberg.com/news/articles/2021-12-21/how-erdogan-s-new-tool-of-fx-linked-deposits-works -- is the thinking that this doesn't come into play here? / that it hasn't been active long enough to become significant yet? / that it's big enough to be worth its own post?
This is an astute analysis. What I have to add is in the spirit of extending the analysis. 1 If my memory serves me well, there was a surplus of liquidity floating around on international markets after OPEC ramped up crude as prices. I believe it lad to currency crises in Latin America, notably Argentina, 2 The post-1990 period ushered in not only the collapse of Communist central planning largely worldwide, but also catch-up growth in a host of less developed nations. It would be useful to look at a data set that includes all of the “potential growers”, sorting out winners from losers. Turkey would lie in the middle of the spectrum Imwould think. One approach would be to take as a classifying variable the growth in the share of global world trade income. One suspects the correlation between sustained per capita income and the trade success variable would be high. That would be a useful starting point for a serious aggregate economic analysis.
What specifically caused the dramatic uptick in growth in the early 2000s? I’d imagine most of the groundwork for your Chang-Studwell model had already been laid much earlier.
Great analysis! Thanks for your wonderful articles, I learn so much
This is a bit unfair to Neo-Fisherism, which makes a much, much narrower claim: a widely-anticipated, slow and steady interest rate increase, with stable and neutral fiscal policy, will over time lead to a rise in the inflation rate. All known (or, at least, widely known) macroeconomic models that are long-run stable -- that is, where small monetary policy misses don't lead to inflationary or deflationary spirals -- have Neo-Fisher properties. And that's basically the only thing Neo-Fisherians we're pointing out, as part of a program to further explore all the properties of models in an ongoing effort to improve them.
John Cochrane has a good, if extremely technical, explanation of the dynamics in response to Lars Svensson's paper on Sweden, in which Lars argues that Neo-Fisher effects don't seem to hold. https://johnhcochrane.blogspot.com/2020/11/a-neo-fisherian-challenge-and.html
But no one serious in that group of people (they I know of) actually advocates using short-term nominal interest rate decreases to reduce inflation. Actually I don't think many of them advocate any policy at all, it is a pretty academic program. I guess you can blame them for that, but not everyone can be working on policy all the time, some people need to work on the nuts and bolts of these things.
Ah, BUT, if agents know that the near-dictatorial president-for-life is going to exert a consistent downward influence on rates, then according to Neo-Fisherism that should be disinflationary even if there are transitory upward rate spikes. As seen on the graph above.
In other words, yes, Neo-Fisherism really is failing here. Now, maybe Neo-Fisherians can patch things up by adding capital inflows and an international sector to the model. But the basic idea looks dead.
Well, you would still see inflation spike initially. But in any case that's really only true if fiscal policy were stable and neutral. See Japan. But I think there is zero chance that is what is happening in Turkey. Sure, the official budget is stable, but let's be real, all the money that the central bank is printing is getting funneled to cronies. Just because you cut out the middleman of the official budget account doesn't make the transfer go away. That said, if that weren't happening, I also wouldn't be shocked if velocity crashed, with all the attendant human suffering that would entail, and then inflation fell as real rates equilibrated again.
Don't get me wrong, I think Neo-Fisherism's predictions are about as useful for public policy on a reasonable time frame as string theory predictions that there are 11 dimensions. It's something that pops out of pretty much every non-crazy macroeconomic model, and so it is pretty important to get to the bottom of it. Not because it helps immediate monetary policy, but because if it is wrong then something is fundamentally wrong with basically all of the macroeconomic models, whether they be monetarist or neo-Keynesian.
In any case, my point was there aren't really any Neo-Fisherians cheering Turkey on right now, because they all know that even if he was acting in accordance with the assumptions, and even if they are right, the adjustment mechanism that gets you there is awful. No one is advocating for what he's doing.
Excellent piece. A return to orthodoxy however would entail a wrenching economic adjustment which might topple Erdogan absurd regime. Don’t expect him to initiate it.
[offtopic] A while back you posted some responses to typical lazy econ critiques, and pointing to the supposedly 'ideologically charged' nature of economics you said that it was a problem but that could be overcome and that practitioners should try to be non-ideological. You pointed out that to that response there was this: "But dude, being anti-ideological is an ideology too!". And you followed that with: "Mmm, an insightful and trenchant observation. Now go take another bong hit and leave me alone.". I was wondering what was your personal opinions on similar positions held by serious names in the philosophy of economics. For example, Julian Reiss author of a 'textbook' introduction to the philosophy of economics and editor of the Routledge handbook of the philosophy of econ has defended in "Fact-value entanglement in positive economics" and elsewhere that economics is value-laden in important ways. Do you think the thesis is wrong, or irrelevant? I would be cool to know what a person completely embedded in econ culture has to say about that, thanks for the blog, it is a pleasure to read.