Our fiscal expenditures and future unfunded fiscal obligations act as a corrosive and destabilizing effect on the economies of the major trading nations.
We've abused the privilege of being the global reserve currency.
Or cut spending to shrink deficits. After all, the 2024 and after people who would pay the consumption taxes are not the 2023 and before people who racked up the debt. Maybe taxing pensions at a higher rate would address that generational injustice?
This is what I was thinking. I looked at gross govt spending from 2017 to 2023. In 2017, we spent $4 trillion. During Covid, it blew up to just over $6 trillion, and in 2021 it was almost $7 trillion (I am rounding up on all trillions) In 2023, it was closer to $6 trillion.....correspondingly, inflation has risen and eased with the spending of the government. I know that correlation is not causation and I will leave it to academics to prove it on a chalkboard, but it certainly is something that seems correct.
To another point in this post. I traded interest rate futures in Chicago and we used to look at the Money Supply numbers closely until the early 90s, we didn't.
Nigeria is currently undergoing monetary policy reforms towing the conventional line of interest-rate hiking in a bid to reduce inflation. Unfortunately, the required accompanying fiscal reforms are playing catchup.
The equation that purports to give the relation between disinflation and the change in expectations of the infinite sum of the discounted future primary surpluses (scaled by the outstanding period t-1 bond principal) is in contradiction to equation (2.7) of Chapter 2 of the pre-press publication of The Fiscal Theory of the Price Level (J. H. Cochrane: Nov. 12, 2021). Equation (2.7) in the pre-press version of The Fiscal Theory of the Price Level shows that the price level, Pₜ , is inversely proportional to the infinite sum of the discounted future primary surpluses.
Because the primay surplus, sₜ₊ₙ , n = j = 0, 1, 2, ..., k-1, k, k+1, ..., ∞, lies on the real number line from –∞ to +∞, linearization or log-linearization is not practicable.
With respect to the graphs taken from Smut and Wouters (2007), it should noted that those graphs illustrate the impulse-response functions of variations of the variables from their steady-state values. Smut and Wouters state that quite clearly in their paper. That is the reason why the graphs show the variables returning to zero as time progresses. The mathematical macro-economic model that Smuts and Wouter study is far from the simple two-equation new-Keynesian DSGE model used for discussion in the blog post and in the author's simulations from which his charts are drawn.
Very stimulating post. Allow me one question for my education.
When describing how monetary policy affects inflation in a world where the central banks sets the interest rate but does not control the money supply, you say (CAPITALISATION ADDED):
How does an interest rate rise affect expected inflation? At the old expected inflation, the higher nominal interest rate is a high real interest rate. That offers an opportunity to save, consume less today, make a good return, and then consume more in the future. When people do that, they lower demand today and raise demand in the future. With flexible prices, output stays the same, so LESS DEMAND LOWERS THE PRICE LEVEL TODAY and pushes up the price level in the future. Expected inflation — future price level over today’s price level — rises, pushing the real rate back to where it was.
I read the lowering of the price level today as lower inflation immediately (at time t in the notation).
Later in the post however you argue that (CAPITALISATION added)
[higher interest rates] do NOT produce a lower inflation EITHER IMMEDIATELY OR in the future
Is not there a contradiction between the two statements?
In the former context, just looking at i = r+Epi, higher interest rates at t raise inflation from t to t+1 (pt+1-pt) so they could lower inflation from t-1 to t. But perhaps all of the demand goes to future p and none to lowering current p. In the latter quote, we're looking at a full model, that pins down whether p goes down or p+1 goes up. With no change in fiscal policy, it's all pt+1 goes up.
Thank you for the clarification. If understand correctly, therefore, in the 'full model', inflation in time t (pt) is pinned down by the central bank reaction function to deviations in inflation from its target level (where, in your preferred reading, the decisive factor is that the unexpected deflation at time t is matched by a increase in the present value of primary surpluses from time to infinity).
May ask you a further question on the central bank's reaction function? In footnote 1 it
the central bank's setting of the interest rate target is characterised as:
I*t = r + Ep*t+1
Possibly owing to the common use of the asterisk to denote the 'natural value' of a variable, I thought that r should be replaced by r* in the formula. This however would not allow deriving the conclusion that the central bank's way of pinning down inflation at time t consists of 'threatening' hyperinflation (or hyperdeflation). At the same time, if r does not coincide with its natural value r, how should it be read?
"Higher interest rates slowly lower spending, output, and hence employment over the course of several months or years. Lower output and employment slowly bring down inflation, over the course of additional months or years."
No. It depends on whether the inflation being reduced is greater than the instantaneous income maximizing inflation (minimum amount of inflation needed to facilitate adjustment f relative prices when some absolute prices are sticky downward). If done implausibly skillfully, we have full employment of all resources (not just labor) all the way back to target.
The key error of the quote is that is has the causation between inflation and output (real income) backward.
"If prices were flexible, monetary policy would result in costless inflation, instantly,"
????
If prices were flexible there would be no need for monetary policy to produce inflation to maintain maximum real income in the face of shocks. Supply shocks would affect real income, but all resources would remain fully employed.
If the velocity is influenced by i, and still hold MV(i)=PV, we still have a theory based on interest rate tgt: higher interest rates decrease V(i) and so P, in the case of flexible prices and an increase future inflation. Is the Fischer equation a consequence of this dynamic? Why we need the Fisher equation?
Monetarism is pretty simple. It has never been tried. Take for example, some people think Feb 27, 2007 started across the ocean. “On Feb. 28, Bernanke told the House Budget Committee he could see no single factor that caused the market’s pullback a day earlier”.
In fact, it was home grown. It reflected a large drop in legal reserves.
Mr. Cochrane, First, I really enjoy following you, so I would like to make a plea for you to thoroughly review your pieces before posting. There are many obvious typos, and some that are less obvious, which make the meaning of sentences difficult to comprehend. Second, isn't your FTPL similar to what Henry Hazlitt suggests about money and interest rates in "Economics in One Lesson?"
When you say that saving more today decreases the price level, how should we interpret this? When we save more today, we bid up the price of capital goods relative to consumption goods. Is there a sense by which aggregage demand and the price level refer only to a consumption basket? Or should we interpret "save" only as investment in government bonds/paper money?
In a full model, or just in general, to what does aggregate demand refer? Is it supposed to be everything except money/near-money? Would the Tesla stock you cited in your article be comprised in AD?
It's a squishy Keynesian concept in common use. You can only demand more strawberries by demanding less bananas. How can we demand more of everything? The only way I have been able to understand it (Walras law) is as the mirror image of the demand for government debt and money. We demand more of everything else when we want to hold less government debt and money.
Have you read Scott Sumner on "market monetarism", in which a policy that takes into account expectations has long-and-variable LEADS instead of lags? He discourages analyzing monetary policy in terms of interest rates, as it tends to confuse things. A booming economy will tend to have high interest rates as a result, but this doesn't mean an action the Fed takes to hike interest rates (which they can't do via a lever connected to interest rates, but instead just by declaring that they're aiming at it and will conduct OMOs to hit it) will cause the economy to boom.
> One reaction to this grand failure has been for many “freshwater” types to simply declare everything since 1970 a big mistake.
Don't you mean "saltwater"?
How does the fiscal theory of the price level explain how Reagan/Volcker brought down inflation even while increasing the deficit?
I second this comment. "Money" should be the fundamental basis for an analysis of monetary policy, not "interest rates". Cochrane has given up on money far, far too early. Cochrane notices the indeterminacy of the new Keynesian interest rate framework ... and then falls into the fiscal theory with basically a Sherlock Holmes "once you eliminate all other possibilities..." kind of clever logic.
But Cochrane already realizes that the monetarist explanation has the correct sign (higher interest rates -- from tighter money -- result in lower inflation), and that the "wrong sign" is a critical failure of the new Keynesian / interest rate model. But, despite that, he somehow can't overcome his obsession with interest rates. (To be fair: just like everybody else in the field.)
This is the critical error: "The Fed does not control money supply. The Fed sets interest rates." That's simply not correct. The Fed sets interest rate TARGETS. The Fed achieves those targets by manipulating the money supply, via OMOs. (At least, this was how monetary policy worked until 2008.)
Here's another error: "Monetarist theory is perfectly clear: The Fed must control the money supply. If it does not do so, and either targets interest rates or provides “an elastic currency” meeting demand, the theory does not work." Not correct. Monetarist theory says that changes in the money supply cause changes in the economy. Whether those changes are deliberate or conscious on the part of the Fed, is NOT important.
Here's my challenge to Cochrane: imagine that ONLY money supply changes mattered for monetary policy. (Scott Sumner says that the primary monetary policy transmission mechanism is the Hot Potato Effect, coming from changes in the quantity of money.) Imagine that changes in interest rates were a ONLY symptom of monetary policy changes, but NOT a cause. It doesn't matter whether the Fed believes they are "changing interest rates". The Fed is mistaken about why their policies work. It is the money supply changes that change the economy. It doesn't matter that the Fed doesn't explicitly target the money supply. What matters is that the money supply does in fact change.
If Cochrane took the (market) monetarist claims seriously, could he himself work out a much simpler and more predictive economic model? Why does he abandon the fundamental nature of money so quickly? It answers all his questions about "sign" and "indeterminancy"!
Suppose in 2022 the Fed lowered rates significantly on the advice of that amateur monetary economist Turkish President Recep Tayyip Erdogan. That would have set off a massive (nominal) asset price boom that started in the housing market and spread throughout the economy
Does a higher real interest rate really increase saving and reduce consumption? Japan is an obvious counterexample. It seems that the relationship is ambiguous and nuanced. One could argue that households have a long-term saving target, and higher real interest rates allow them to reach that target with less saving, allowing for more consumption.
I find it strange that in your discussion of inflation you never mention the supply (and demand) for money. Most central banks see their policy rate as the instrument for controlling M (setting it above or below the neutral rate reduces or increases the supply of base money).
The first part of the post discussed money, and why it does not apply. I disagree -- I haven't heard monetary aggregates, or controlling money supply, discussed anywhere at central banks in about two decades. They set interest rates, pay interest on reserves, and view the "transmission" of monetary policy entirely in terms of interest rate effects on "spending," not by inducing a scarcity of the medium of exchange.
Yes, that is how central banks generally think about it. Given that almost all that target inflation overshot their targets a couple of years ago, often by embarrassingly large amounts, perhaps a change of perspective is in order.
Central banks control the monetary base. Those with floating exchange rates in principle have complete freedom to freeze it, grow it, or shrink it at any rate they please -- not that every course of action is equally wise. Central bank interest rates are the price of renting the monetary base. By pushing the nominal interest rate sufficiently far above or below the rate of inflation, central banks can discourage or encourage expansion of the monetary base.
Banks need the monetary base as reserves for clearing. There is no rigid relationship between the monetary base and broader money supply measures that include bank deposits and other forms of credit. Despite a substantial degree of flexibility in adjusting their portfolios, though, if banks are persistently short of reserves, they need to contract credit to stay in business. Central banks do not control broad measures of the money supply, but they do influence them. Then the supply of broad money interacts with the demand to produce generally rising, stable, or falling prices (spending in excess of growth in goods, spending about in line with growth in goods, spending less than growth in goods -- in each case, abstracting from technological change that tends to make goods cheaper).
Something for you to consider: what if central banks are mistaken about why monetary policy works? They may not "discuss" controlling the money supply ... but that may, in fact, be the economic reason why monetary policy affects the economy.
And central banks don't "set" interest rates (at least, not important rates like the FFR). They target them. And achieve those targets by changing the money supply (monetary base).
Our fiscal expenditures and future unfunded fiscal obligations act as a corrosive and destabilizing effect on the economies of the major trading nations.
We've abused the privilege of being the global reserve currency.
When bad stuff happens Theory yields to Practice
Agree. WE need to increase consumption taxes to shrink deficits until Deficits < Σ(expenditures with NPV>0)
From your lips to the central planners ears
Or cut spending to shrink deficits. After all, the 2024 and after people who would pay the consumption taxes are not the 2023 and before people who racked up the debt. Maybe taxing pensions at a higher rate would address that generational injustice?
This is what I was thinking. I looked at gross govt spending from 2017 to 2023. In 2017, we spent $4 trillion. During Covid, it blew up to just over $6 trillion, and in 2021 it was almost $7 trillion (I am rounding up on all trillions) In 2023, it was closer to $6 trillion.....correspondingly, inflation has risen and eased with the spending of the government. I know that correlation is not causation and I will leave it to academics to prove it on a chalkboard, but it certainly is something that seems correct.
To another point in this post. I traded interest rate futures in Chicago and we used to look at the Money Supply numbers closely until the early 90s, we didn't.
Great overview of the issues and excellent conclusion, thanks
Nigeria is currently undergoing monetary policy reforms towing the conventional line of interest-rate hiking in a bid to reduce inflation. Unfortunately, the required accompanying fiscal reforms are playing catchup.
The equation that purports to give the relation between disinflation and the change in expectations of the infinite sum of the discounted future primary surpluses (scaled by the outstanding period t-1 bond principal) is in contradiction to equation (2.7) of Chapter 2 of the pre-press publication of The Fiscal Theory of the Price Level (J. H. Cochrane: Nov. 12, 2021). Equation (2.7) in the pre-press version of The Fiscal Theory of the Price Level shows that the price level, Pₜ , is inversely proportional to the infinite sum of the discounted future primary surpluses.
Because the primay surplus, sₜ₊ₙ , n = j = 0, 1, 2, ..., k-1, k, k+1, ..., ∞, lies on the real number line from –∞ to +∞, linearization or log-linearization is not practicable.
With respect to the graphs taken from Smut and Wouters (2007), it should noted that those graphs illustrate the impulse-response functions of variations of the variables from their steady-state values. Smut and Wouters state that quite clearly in their paper. That is the reason why the graphs show the variables returning to zero as time progresses. The mathematical macro-economic model that Smuts and Wouter study is far from the simple two-equation new-Keynesian DSGE model used for discussion in the blog post and in the author's simulations from which his charts are drawn.
Very stimulating post. Allow me one question for my education.
When describing how monetary policy affects inflation in a world where the central banks sets the interest rate but does not control the money supply, you say (CAPITALISATION ADDED):
How does an interest rate rise affect expected inflation? At the old expected inflation, the higher nominal interest rate is a high real interest rate. That offers an opportunity to save, consume less today, make a good return, and then consume more in the future. When people do that, they lower demand today and raise demand in the future. With flexible prices, output stays the same, so LESS DEMAND LOWERS THE PRICE LEVEL TODAY and pushes up the price level in the future. Expected inflation — future price level over today’s price level — rises, pushing the real rate back to where it was.
I read the lowering of the price level today as lower inflation immediately (at time t in the notation).
Later in the post however you argue that (CAPITALISATION added)
[higher interest rates] do NOT produce a lower inflation EITHER IMMEDIATELY OR in the future
Is not there a contradiction between the two statements?
In the former context, just looking at i = r+Epi, higher interest rates at t raise inflation from t to t+1 (pt+1-pt) so they could lower inflation from t-1 to t. But perhaps all of the demand goes to future p and none to lowering current p. In the latter quote, we're looking at a full model, that pins down whether p goes down or p+1 goes up. With no change in fiscal policy, it's all pt+1 goes up.
Thank you for the clarification. If understand correctly, therefore, in the 'full model', inflation in time t (pt) is pinned down by the central bank reaction function to deviations in inflation from its target level (where, in your preferred reading, the decisive factor is that the unexpected deflation at time t is matched by a increase in the present value of primary surpluses from time to infinity).
May ask you a further question on the central bank's reaction function? In footnote 1 it
the central bank's setting of the interest rate target is characterised as:
I*t = r + Ep*t+1
Possibly owing to the common use of the asterisk to denote the 'natural value' of a variable, I thought that r should be replaced by r* in the formula. This however would not allow deriving the conclusion that the central bank's way of pinning down inflation at time t consists of 'threatening' hyperinflation (or hyperdeflation). At the same time, if r does not coincide with its natural value r, how should it be read?
"Higher interest rates slowly lower spending, output, and hence employment over the course of several months or years. Lower output and employment slowly bring down inflation, over the course of additional months or years."
No. It depends on whether the inflation being reduced is greater than the instantaneous income maximizing inflation (minimum amount of inflation needed to facilitate adjustment f relative prices when some absolute prices are sticky downward). If done implausibly skillfully, we have full employment of all resources (not just labor) all the way back to target.
The key error of the quote is that is has the causation between inflation and output (real income) backward.
"If prices were flexible, monetary policy would result in costless inflation, instantly,"
????
If prices were flexible there would be no need for monetary policy to produce inflation to maintain maximum real income in the face of shocks. Supply shocks would affect real income, but all resources would remain fully employed.
If the velocity is influenced by i, and still hold MV(i)=PV, we still have a theory based on interest rate tgt: higher interest rates decrease V(i) and so P, in the case of flexible prices and an increase future inflation. Is the Fischer equation a consequence of this dynamic? Why we need the Fisher equation?
not if M is endogenous. And there are also multiple equilibria, even if M is fixed. since i rises with inflation, M V(P_t+1/P_t) =P_t Y
Clear exposition on a controversial topic.
Monetarism is pretty simple. It has never been tried. Take for example, some people think Feb 27, 2007 started across the ocean. “On Feb. 28, Bernanke told the House Budget Committee he could see no single factor that caused the market’s pullback a day earlier”.
In fact, it was home grown. It reflected a large drop in legal reserves.
Mr. Cochrane, First, I really enjoy following you, so I would like to make a plea for you to thoroughly review your pieces before posting. There are many obvious typos, and some that are less obvious, which make the meaning of sentences difficult to comprehend. Second, isn't your FTPL similar to what Henry Hazlitt suggests about money and interest rates in "Economics in One Lesson?"
Please pass on typos, either here or by email. Believe it or not I do look before posting.
When you say that saving more today decreases the price level, how should we interpret this? When we save more today, we bid up the price of capital goods relative to consumption goods. Is there a sense by which aggregage demand and the price level refer only to a consumption basket? Or should we interpret "save" only as investment in government bonds/paper money?
This is a really simple model without capital.
In a full model, or just in general, to what does aggregate demand refer? Is it supposed to be everything except money/near-money? Would the Tesla stock you cited in your article be comprised in AD?
It's a squishy Keynesian concept in common use. You can only demand more strawberries by demanding less bananas. How can we demand more of everything? The only way I have been able to understand it (Walras law) is as the mirror image of the demand for government debt and money. We demand more of everything else when we want to hold less government debt and money.
Have you read Scott Sumner on "market monetarism", in which a policy that takes into account expectations has long-and-variable LEADS instead of lags? He discourages analyzing monetary policy in terms of interest rates, as it tends to confuse things. A booming economy will tend to have high interest rates as a result, but this doesn't mean an action the Fed takes to hike interest rates (which they can't do via a lever connected to interest rates, but instead just by declaring that they're aiming at it and will conduct OMOs to hit it) will cause the economy to boom.
> One reaction to this grand failure has been for many “freshwater” types to simply declare everything since 1970 a big mistake.
Don't you mean "saltwater"?
How does the fiscal theory of the price level explain how Reagan/Volcker brought down inflation even while increasing the deficit?
Scott has now responded to this blog post here:
https://scottsumner.substack.com/p/the-interest-rate-monoculture
I second this comment. "Money" should be the fundamental basis for an analysis of monetary policy, not "interest rates". Cochrane has given up on money far, far too early. Cochrane notices the indeterminacy of the new Keynesian interest rate framework ... and then falls into the fiscal theory with basically a Sherlock Holmes "once you eliminate all other possibilities..." kind of clever logic.
But Cochrane already realizes that the monetarist explanation has the correct sign (higher interest rates -- from tighter money -- result in lower inflation), and that the "wrong sign" is a critical failure of the new Keynesian / interest rate model. But, despite that, he somehow can't overcome his obsession with interest rates. (To be fair: just like everybody else in the field.)
This is the critical error: "The Fed does not control money supply. The Fed sets interest rates." That's simply not correct. The Fed sets interest rate TARGETS. The Fed achieves those targets by manipulating the money supply, via OMOs. (At least, this was how monetary policy worked until 2008.)
Here's another error: "Monetarist theory is perfectly clear: The Fed must control the money supply. If it does not do so, and either targets interest rates or provides “an elastic currency” meeting demand, the theory does not work." Not correct. Monetarist theory says that changes in the money supply cause changes in the economy. Whether those changes are deliberate or conscious on the part of the Fed, is NOT important.
Here's my challenge to Cochrane: imagine that ONLY money supply changes mattered for monetary policy. (Scott Sumner says that the primary monetary policy transmission mechanism is the Hot Potato Effect, coming from changes in the quantity of money.) Imagine that changes in interest rates were a ONLY symptom of monetary policy changes, but NOT a cause. It doesn't matter whether the Fed believes they are "changing interest rates". The Fed is mistaken about why their policies work. It is the money supply changes that change the economy. It doesn't matter that the Fed doesn't explicitly target the money supply. What matters is that the money supply does in fact change.
If Cochrane took the (market) monetarist claims seriously, could he himself work out a much simpler and more predictive economic model? Why does he abandon the fundamental nature of money so quickly? It answers all his questions about "sign" and "indeterminancy"!
Suppose in 2022 the Fed lowered rates significantly on the advice of that amateur monetary economist Turkish President Recep Tayyip Erdogan. That would have set off a massive (nominal) asset price boom that started in the housing market and spread throughout the economy
Does a higher real interest rate really increase saving and reduce consumption? Japan is an obvious counterexample. It seems that the relationship is ambiguous and nuanced. One could argue that households have a long-term saving target, and higher real interest rates allow them to reach that target with less saving, allowing for more consumption.
I find it strange that in your discussion of inflation you never mention the supply (and demand) for money. Most central banks see their policy rate as the instrument for controlling M (setting it above or below the neutral rate reduces or increases the supply of base money).
The first part of the post discussed money, and why it does not apply. I disagree -- I haven't heard monetary aggregates, or controlling money supply, discussed anywhere at central banks in about two decades. They set interest rates, pay interest on reserves, and view the "transmission" of monetary policy entirely in terms of interest rate effects on "spending," not by inducing a scarcity of the medium of exchange.
Yes, that is how central banks generally think about it. Given that almost all that target inflation overshot their targets a couple of years ago, often by embarrassingly large amounts, perhaps a change of perspective is in order.
Central banks control the monetary base. Those with floating exchange rates in principle have complete freedom to freeze it, grow it, or shrink it at any rate they please -- not that every course of action is equally wise. Central bank interest rates are the price of renting the monetary base. By pushing the nominal interest rate sufficiently far above or below the rate of inflation, central banks can discourage or encourage expansion of the monetary base.
Banks need the monetary base as reserves for clearing. There is no rigid relationship between the monetary base and broader money supply measures that include bank deposits and other forms of credit. Despite a substantial degree of flexibility in adjusting their portfolios, though, if banks are persistently short of reserves, they need to contract credit to stay in business. Central banks do not control broad measures of the money supply, but they do influence them. Then the supply of broad money interacts with the demand to produce generally rising, stable, or falling prices (spending in excess of growth in goods, spending about in line with growth in goods, spending less than growth in goods -- in each case, abstracting from technological change that tends to make goods cheaper).
Something for you to consider: what if central banks are mistaken about why monetary policy works? They may not "discuss" controlling the money supply ... but that may, in fact, be the economic reason why monetary policy affects the economy.
And central banks don't "set" interest rates (at least, not important rates like the FFR). They target them. And achieve those targets by changing the money supply (monetary base).