The macroeconomics of doing nothing

Suppose there is an increase in desired saving, and the monetary and fiscal authorities do nothing. What happens?

That's the most important practical question in macroeconomics over the last few years. And it's also a really stupid question. And understanding why it's a really stupid question, and changing the way that question is asked -- not just in academia, but in the real world -- is the most important practical task of macroeconomic theory today.

Let's start with "saving". In macroeconomics, "saving" is defined as "not spending part of your disposable income on newly-produced consumption goods and services". It's a purely negative definition. And it invites the supplementary question: "OK, so if you are not spending it on newly-produced consumption goods and services, what are you doing with your disposable income instead?".

But few macroeconomists would ever ask that question. Which is unfortunate, because the answer to that question matters a lot. For example, suppose you were a Keynesian of some sort. Your immediate instinct would be to answer my original question by saying that an increase in desired saving, if the monetary and fiscal authorities do nothing, would cause a recession. But if you stopped and asked the supplementary question, and I answered "they decide to buy newly-produced investment goods instead of newly-produced consumption goods", you would stop talking about a recession.

Did you, the reader, stop and ask that supplementary question -- about what people wanted to do instead? I bet few of you did. But now I've asked it for you, what were you implicitly assuming about that supplementary question when you answered the original question? Were you implicitly assuming they wanted to buy antique furniture, or land, or bonds, or money, or what? And I bet most of you have to stop and think about what you were implicitly assuming, and stop and think whether it would matter whether it were antique furniture, or land, or bonds, or money, or what.

What this means is that I rigged the original question in order to elicit a particular answer from you. I think it's called "push-polling", when political pollsters do it. But it doesn't have to be deliberate. And it is something we can do to ourselves. We ask ourselves a question in a certain way, and the way we ask it affects the way we think about it and answer it.

For example, if you say that the average Canadian owes $45,000, you think one way. If you say the average Canadian is owed $45,000, you think another way. But in a closed economy, or one with no net foreign debt, debits equal credits, so the two are the same.

"Saving" is a "doing nothing" concept. When we save, we are doing nothing with our income. When desired saving increases, people want to do even more nothing. But there are 1,001 different ways of doing "nothing". And it matters which one you do.

Talking about monetary and fiscal authorities "doing nothing" is just as problematic. Does it mean holding a rate of interest fixed? Holding the money supply fixed? Holding the exchange rate fixed? Holding the price of gold fixed? Holding the inflation target fixed? Holding the NGDP target fixed? Or what? Different economists will make different assumptions (implicit or explicit) about what it means to hold monetary policy fixed. Because they have different views on what it is that monetary authorities do. So what one economist thinks of as the central bank doing nothing, another will think of as the central bank doing something.

It's like in those "trolley problems". The difference between "doing something" and "letting something happen" might not seem that important, but it changes the way most of us think about moral problems, and what we ought to do, or not do.

Now, if it were only economists who got confused about different ways of framing questions, it wouldn't matter that much. Eventually, after a lot of arguing at cross purposes, we would get it sorted out. We would become more precise about the question being asked. And perhaps go on to argue about whose implicit assumptions about what people want to do with their savings instead, and what it is that monetary and fiscal policies are holding constant, make most sense. But it's not that simple. Because it's not just economists whose answers depend on how the question is framed. It's real people too.

Expectations matter. I mean real people's expectations matter. Economics is about those real people. "What is the effect of X?" depends in part upon how X changes people's expectations, and that in turn will depend, not on how economists frame the question, but on how real people frame the question. How will real people think about the event X? What will real people be holding constant when they revise their expectation after learning X?

Economists, in part, need to be ethnomethodologists. We need to think about how people construct their own economic models, and how they think about "doing nothing". But, like anthropologists, we need to stand apart from the natives' own way of viewing their world. We don't have to believe it ourselves to recognise that they believe it. And sometimes we need to be missionaries, and suggest that other ways of viewing the world might lead to better outcomes.

If people think about monetary policy as setting an interest rate, then if they see that interest rate fall to zero they will think that monetary policy is powerless to loosen further. But if people think about monetary policy as setting the price of gold, and the price of gold is still finite, they will think that monetary policy can loosen further. And the effects of monetary policy depend in very large part upon what people expect those effects to be, because monetary policy is not just about what the central bank does today but what it is expected to do in the future.

If people think about monetary policy as setting an interest rate, then "promising a higher future price level" means the central bank will have to do something in the future. It will have to deliberately set future interest rates too low for too long. But if people think of monetary policy as setting the price of gold, then the central bank can simply raise the price of gold today, and hold it there. It need do nothing in the future.

Roosevelt could promise a higher future price level without having to promise to do something in the future. People automatically assumed he would do nothing. By "doing nothing" they understood "not lowering the price of gold back down again".

To misquote A J Ayer, an economist who says there is no confidence fairy is not an atheist. He is a fellow metaphysician with a rival fairy of his own.

It is probably not feasible today to attempt the social re-construction of the gold standard reality. And it wasn't a very good reality in any case. But it does show, by example, that today's social reality is not the only possible reality.

"But central banks really do, do, do set a rate of interest" is said by people who have swallowed the pill proffered and swallowed by the central bankers themselves, and by the rest of the population. That's how the natives at the central bank describe what they are doing, and most economic anthropologists have decided it's simpler to take the natives' self-description as a statement about concrete reality. It's not that there's a blue and red pill, one false and one true; but there are many pills, and each can re-create its own reality. But you can do different things in each of those different realities, because each reality has its own definition of "doing nothing".

Or, switching metaphors to my generation's, the pills that maternal central banks have given you now can't do anything at all. So maybe they should switch pills.

This was supposed to be a post on David Beckworth and Paul Krugman on balance sheet recessions. Half the population is in debt to the other half. Then something happens and the debtors can't continue to borrow and dissave. Nobody will lend to them, so they are borrowing-constrained. So they consume less. Aggregate savings rises, and the central bank has already pushed the rate of interest down to 0%, so can do nothing. So there's a recession, unless the fiscal authorities do something.

"The debtors stop borrowing from the creditors". "The creditors stop lending to the debtors". Those are two different ways of framing the same event. The first sounds like it would lead to an increase in aggregate savings, as the dissavers stop borrowing. The second sounds like it would lead to a decrease in aggregate savings, as the savers stop lending. What are the savers going to do instead, if they can't save by lending to the debtors? They can't just do nothing?

What can monetary policy do to solve the problem, if there is one? Has it already done as much as it can do, and can do nothing more? Would we believe it if it promised to do something in future? If people thought of monetary policy in a different way, then there could be something it could do now, without having to promise to do something in the future. Roosevelt did it. And the only difference between then and now is how people think about what central banks do, and how they do nothing. The change in perceptions about what "doing nothing" means has meant that Roosevelt's concrete policy lever now looks like a confidence fairy.

Comments

[Marcus Nunes]

Nick: Like firms that feel compelled to "reinvent" themselves, Central Banks may also feel obliged. But they are afraid to "lose their core compentencies".

[Lord]

Since people see debt as the problem, the other way should be to provide the means to reduce debt. Give sufficient money to everyone and debtors can pay their debts and creditors can collect them. Keep giving enough until debt is no longer a concern. Instead of giving preference to the increase of leverage, allow the decrease of it. This would be more egalitarian than benefiting just debtors and creditors. It could be done now rather than in the future and could be done sufficiently that it could be ended in the near future. Some would be concerned about inflation but better that than deflation and depression.

[Scott Sumner]

Over in my blog a commenter named W.Peden left this comment:

"The revolution will be complete when (to paraphrase Nick Rowe) we think in these terms:

Monetary policy: targets NGDP and therefore the control of inflation in the long run.

Fiscal policy: targets the capital structure through taxes, public spending.

Labour market policy: targets unemployment in the long run.

This is as opposed to the bad old days of Neo-Keynesianism in the 1960s and 1970s-

Monetary policy: targets the capital structure by adjusting interest rates.

Fiscal policy: targets unemployment through expansion.

Labour market policy: targets inflation through incomes policies."

This is perfect. If you did a post on this, please do it again. Someday we'll have NGDP futures targeting, and AD will be taken off the table. Macro will focus on labor market problems and financial market problems.

There is no way that more saving should reduce AD, unless the central bank is completely incompetent. Unfortunately . . . .

As I started to read the post I kept thinking "but how do you define 'doing nothing'," at which point you hit the nail right on the head.

In America we are currently searching around desperately for measures to boost AD, even as our central bank refuses to do so, and yet most of the criticism of our central bank is that they've boosted NGDP far too much!!! If Canada wasn't so cold I'd move up there tomorrow.

mentioned an earlier post of yoyurs that I missed. i don't recall how he described it exactly, but he

[Scott Sumner]

Oops, ignore my last garbled line

[anon]

Scott [quoting W. Peden]:

"Monetary policy: targets NGDP and therefore the control of inflation in the long run.
Fiscal policy: targets the capital structure through taxes, public spending.
Labour market policy: targets unemployment in the long run."

My understanding is that, given inflation-targeting central banks, we are in fact quite close to this vision because "the central bank always moves last" and will offset shifts in fiscal policy in either direction. Of course there will be differences, in that aggregate supply shocks in either direction will be reinforced because the CB's policy reaction will cause a shift in aggregate demand--but this seems beside the point.

As an aside: can incomes policies work by keeping monopolistic markups down and boosting AS? Nick suggested this as a theoretical possibility in a previous post.

[Lee Kelly]

Suppose the government passes a law forbidding people to withdraw base money from banks; all hand-to-hand currency is replaced with privately issued banknotes. While withdrawals of base money drain bank reserves, banknote issuance merely changes the composition of bank liabilities. Base money is now used solely for interbank clearing transactions.

Also, suppose the Fed decides to expand its interest-on-excess-reserves policy to include all reserves and will begin charging negative interest rates under particular circumstances.

Done supposing? Good.

The Fed now announces that it will set the interest rate it offers on reserve balances relative to the interest rate on other highly liquid and safe assets, (e.g. T-bills). For the sake of argument, suppose they set the interest rate on reserve balances 2% lower than some average of highly liquid and safe assets.

Notice how the zero nominal bound has just disappeared because we replaced one set of institutional rules for another.

Since the cost of reserve balances must be passed onto customers in lower (or even negative) rates of interest on chequing deposits, the opportunity cost of holding either base money (for banks) or chequing deposits (for individuals and firms) remains constant. For example, if the interest rate on T-bills falls to about 1% because of a sudden and unexpected rush to liquidity and safety, then the interest rate on reserve balances will be -1%, and the cost will be passed onto individuals and firms in lower or negative interests on chequing deposits.

Suppose that despite the negative interest rate on money balances, banks, firms, and individuals suddenly become so risk averse that, on aggregate, they desire greater money balances anyway. What happens? Without a corresponding increase in the money supply (to offset both the increased demand for reserves and the increased demand for money), then NGDP, aggregate demand, nominal expenditures, national income, or whatever, will fall.

Let's suppose the central bank just lets this fall in NGDP happen for a while, maybe until it has fell more than any other time since the Great Depression.

Eventually, the central bank begins to try and satisfy the (by now now astronomic) increase in money demand with expansionary policy. But what happens to interest rates? As nominal spending and income rises, previously risky and illiquid assets become less so, because firms are now finding it easier to sell goods and services. The rush into highly liquid and safe assets is slowly reversed and the demand for T-bills declines. Interest rates begin to rise until the interest rate offered on reserve balances is no longer negative.

So now for my point. In this story, contractionary monetary policy has nothing to do with any zero nominal bound, because none exists: a liquidity trap is institutionally impossible. On the other hand, expansionary monetary policy is identified with rising interest rates on non-money assets -- and not just as some long run effect but an immediate consequence. In other words, the seeming importance of interest rates is something like an institutional epiphenomenon.

I win, right? Please let me know if I have made any mistakes.

[axiom]

Nonsense.

The gold standard is just another rule for setting interest rates.

When did a central bank under a gold standard not have to worry about setting interest rates?

[Nick Rowe]

Marcus: I think that's a good analogy. But it can't be easy to change a social reality. Cue Monty Python, Holy Grail:

"Who are you?"
"I am Arthur, King of the Britons"
"What's a king? Who are the Britons?"

Arthur is the king iff people believe he is the king. Otherwise, he's just a guy who doesn't have shit all over him.

Lord: sounds like helicopter money. Or what Australia did.

Scott: That was originally from a Festschrift for David Laidler. I was supposed to be commenting on Michael Bordo's paper. But Michael Bordo had done such a very thorough detailed job of covering David's research that I could think of anything useful to add. So I decided to go off on a big-picture tangent, about how Monetarism had won the war after all, despite losing the k% battle.

I thought I had done a post specifically on this, but can only find odd snippets when I search.

Basically, it went like this:

The monetarist revolution was a revolution in the assignment of policy targets to policy instruments.

In the 1970's Keynesianism:

1. AD (understood as unemployment) was assigned to fiscal policy.
2. The composition of AD, between consumption and investment, (or, in an open economy, between domestic absorption and net exports) was assigned to monetary policy, via its influence on interest rates and exchange rates.
3. That left inflation without an instrument. So they hunted around, and cobbled together a rag-bag of policies termed "industrial policy", which meant controlling monopolies in both labour and output markets.

And Monetarism:
1. AD (understood as inflation) was assigned to monetary policy.
2. The composition of AD, between G vs C+I (or, in an open economy, between C+I+G vs NX) was assigned to fiscal policy.
3. That left unemployment without an instrument. So the industrial policy ragbag was assigned unemployment.

Canada, indeed most countries, look very monetarist from the perspective of a 1970's Keynesian.

anon: "My understanding is that, given inflation-targeting central banks, we are in fact quite close to this vision because "the central bank always moves last" and will offset shifts in fiscal policy in either direction."

Yes, I agree. But I think there may be a subtle difference between "the central bank always moves last" and "the central bank sees it as it's *responsibility* to always move last".

Theoretically, wage and/or price controls can permanently reduce unemployment in a macro model with imperfect competition. But it's a lot more practical to implement in a representative agent/firm model where the policymaker can actually see all the curves, and all of PSST is ignored.

Lee: Dunno. It's not obviously wrong. Reminds me, I think, of how Bill Woolsey views AD. It's not interest rates that matter; it's the spread between the interest rate on the medium of exchange and other interest rates.

[Nick Rowe]

axiom: so, what did Roosevelt have to do to interest rates in order to raise the price of gold?

[kharris]

Good thinkers do not, by definition, think in terms of "the central bank doing nothing" when that expression is ambiguous. That is, in fact, one of the crippling tendencies of not-so-good thinkers - and partisans, of course. How many debates have you waded through about "globalization" during which no party to the debate bothers to explain what they mean by globalization?

If one means, "the central bank should (not) expand bank reserves at a 3.5% pace over the coming year", then one should say exactly that. If one thinks that the central bank ought to do "something" but one is willing to allow specialists to figure out what "something" ought to be, then say exactly that, and take your licks form the jerks who will make you suffer for your modesty and clear statement.

[Scott Sumner]

Nick, Thanks for that summary--I like that a lot. You are right about gold, and axiom is wrong. The policy instrument was the ratio of gold to the base. That was the only variable that central banks could control. Interest rates were endogenous, because it was a fixed exchange rate system.

Lee, I agree that interest rates are an epiphenomenon, indeed I often use that term in my blog. But I'm not sure I understand your assumption. Is the price of banknotes in terms of reserves assumed to be fixed at one?

[Nick Rowe]

Scott: out of curiosity though, what *did* happen to interest rates when Roosevelt announced he was increasing the price of gold? My hunch is that they did not fall. Or, rather, interest rates on *safe* government bonds did not fall. Interest rates on riskier assets might have fallen, as they became less risky.

[jsalvatier]

Clearly the point that peoples internal models of the world affect how that world works is an important an true point, especially in monetary economics. However, this business of 'alternate realities' is wrong (probably mostly an example of #7 here http://lesswrong.com/lw/od/37_ways_that_words_can_be_wrong/). 'Reality' has lots of connotations that you do not intend and because of that you are likely to make readers more confused than they need be and risk the possibility of confusing yourself, since your brain is not immune to reasoning via connotations (no one's is).

[Nick Rowe]

jsalvatier. Maybe. Here's what I have in mind:

In one "reality" every Briton believes that Arthur is King of the Britons, and so they obey him, and expect everyone else to obey him, and so he really is King of the Britons.

In a second "reality", no Briton believes Arthur is King of the Britons (because they don't have have the concept of "King", or because they believe someone else is King), and so nobody obeys him.

Very different things happen in those two realities (in one the Saxons are defeated, and in the other they aren't), even as described by an outside observer who didn't understand the concept of a "King", but those physical differences are the result of the Britons' own perceptions of their world.

Britons can believe Arthur is King, or not. Those beliefs are not about anything "real", but they do affect what is real.

We can believe central banks set interest rates, or not. Those beliefs are not about anything "real", but they do affect what is real.

[OGT]

I think this paragraph is, let's just say unpersuasive:

"The debtors stop borrowing from the creditors". "The creditors stop lending to the debtors". Those are two different ways of framing the same event. The first sounds like it would lead to an increase in aggregate savings, as the dissavers stop borrowing. The second sounds like it would lead to a decrease in aggregate savings, as the savers stop lending. What are the savers going to do instead, if they can't save by lending to the debtors? They can't just do nothing?

Let's say I owe Nick Rowe $200,000 on the mortgage on my house and the house has suddenly gone underwater. I have a debt of 200K and Nick has an asset of 200K. I am credit constrained and technically insolvent. Assumnig there's been an unexpected disinflation giving me less than expected nominal wage growth I am likely to further reduce expenditures if I want avoid delinquency (though I may not want to).

Our saver, Nick, on the other hand has received something of a windfall on his interest income from the unexpected disinflation. His 'asset,' on the other hand, is very likely worth considerably less, with the threat of further hanging over his head. So I am insolvent and Nick's net wealth is significantly reduced. In all likelihood he will not do nothing, he will increase his savings.

All of which is to say that the neo-classicists' favorite accounting identity is much less clever than they pretend to think it is. Irving Fischer was correct. Financial markets are not complete, dynamically or otherwise, and pretending they are makes your work and theories less useful than they could be.

[Nick Rowe]

OGT: break it up into bits. What caused the decline in total real wealth?
The house price decline.
The debt.
The nominal wage decline.
The default.
The credit constraint.
etc.

That's how analysis works. It avoids the hopeless jumble of what we are comparing things to.

But, this post is not about wealth effects.

[Too Much Fed]

Nick's post said: "Scott: out of curiosity though, what *did* happen to interest rates when Roosevelt announced he was increasing the price of gold? My hunch is that they did not fall. Or, rather, interest rates on *safe* government bonds did not fall. Interest rates on riskier assets might have fallen, as they became less risky."

I'd like to know the answer to that too. My hunch is that the fed funds rate was near zero(0) already.

From the beckworth post: "He backed up the message by devaluing the gold content of the dollar and not sterilizing gold inflows."

My hunch to that is that the gold holder's demand deposit account was marked up and the same bank's reserve account was marked up 1 to 1. And, not sterilizing gold inflows means there was no loan/bond issued to remove the excess reserves. The fed funds rate went to near zero(0) but was already there.

In other words, they created more currency/demand deposits with no loan/bond attached increasing the amount of medium of exchange.

Also from the beckworth post, see the comments starting with Jazzbumpa about the amount of medium of exchange falling.

http://macromarketmusings.blogspot.com/2011/06/monetary-policy-efficacy-during-balance.html

[Too Much Fed]

Nick, do these two(2) charts help?

http://research.stlouisfed.org/fred2/series/AAA

http://research.stlouisfed.org/fred2/series/BAA

[Too Much Fed]

"And it invites the supplementary question: "OK, so if you are not spending it on newly-produced consumption goods and services, what are you doing with your disposable income instead?"."

And, "But if you stopped and asked the supplementary question, and I answered "they decide to buy newly-produced investment goods instead of newly-produced consumption goods", you would stop talking about a recession."

With their savings what if they are purchasing (or just keeping) the highest yielding asset that won't go down in value or be defaulted on and has a velocity of zero(0)? Currency, FDIC-insured demand deposits, and FDIC-insured CD's (need to include what happens to the interest but could be small compared to the amount of medium of exchange used to purchase it) are the first things I think of.

[Lee Kelly]

Nick, my comments were inspired by Bill Woolsey, because I think he is right.

Suppose the government fixes the price of apples but lets the market determine the price of oranges. Since apples and oranges are close substitutes, whether there is a surplus or shortage of apples at the fixed price will depend on the price of oranges.

Suppose we begin in equilibrium with bags of apples and oranges both priced at $5. If the price of oranges rises to $6, then some people will begin switching to apples instead. But the price of apples is fixed at $5; the increased demand for apples does not push up their price. In consequence, there is a shortage of apples when oranges are priced at $6. Likewise, when oranges are priced at $4, there is a surplus of apples.

The obvious way to bring the market for apples back into equilibrium would be to remove price controls, but that's not thinking like a central banker. Another more technocratic way would be to influence the supply and demand for oranges until there is no longer disequilibrium in the apple market. Perhaps some quasi-private independent agency set up by the government could buy and sell oranges to influence their price (OMOs: orange market operations).

The agency would look at data on the apple market and try to discern whether there is a surplus or a shortage, it would then set some target for the price of oranges that it expects will equilibrate the market for apples. Everyone would know that the agency controls equilibrium in the market for apples by influencing the price of oranges.

So apples are like money and oranges are like T-bills (in case you hadn't got that yet). Changing the price of T-bills can change the supply and demand for money (which has a fixed zero percent interest rate). Since the supply and demand for money determines nominal expenditures, it follows that changing the price of T-bills can change nominal expenditures.

The whole line of reasoning presupposes that apples (or money) have a fixed price (or interest rate). If that assumption is relaxed, then an increase in the price of oranges cause people to demand apples instead, which also pushes up the price of apples. The price of apples equilibrates its own market without the need to manipulate the supply and demand for oranges. The shortage or surplus of apples was never really because oranges were priced wrong, but because apples were priced wrong.

If base money is given an interest rate of its own which tends to move with other rates, then the strong relation between nominal expenditures and interest rates is severed. In other words, the sudden fall in nominal expenditures is never really because T-bills have the wrong interest rate, but because money has the wrong interest rate.

Of course, if you're a quasi-monetarist, you just don't care about the price of oranges at all. If the price of apples has to be fixed at $5, then the government agency should adjust the supply of apples (not the supply and demand for oranges) to that $5 is always the equilibrium price. This will sometimes mean making offsetting apple supply adjustments to changes in the price of oranges, but since other things effect the supply and demand for apples, concentrating on oranges only can be misleading.

Well, this is how I think about it. Economics is really all about apples and oranges.

[Phil Koop]

"If people think about monetary policy as setting an interest rate, then "promising a higher future price level" means the central bank will have to do something in the future. It will have to deliberately set future interest rates too low for too long. But if people think of monetary policy as setting the price of gold, then the central bank can simply raise the price of gold today, and hold it there. It need do nothing in the future."

It's not quite so simple. As it happens, Mike Konczal (relatively) recently posted some excerpts of Roosevelt's fireside speech (http://rortybomb.wordpress.com/2011/06/21/president-sets-a-price-level-target-in-a-depression-fdr-1932-edition/) in which he highlights passages wherein Roosevelt explicitly states that the price of gold is not an end in itself; he intends to keep making gold more expensive until general commodity prices are restored to previous levels. This is perhaps the clearest passage:

"Some people are putting the cart before the horse. They want a permanent revaluation of the dollar first. It is the Government’s policy to restore the price level first. I would not know, and no one else could tell, just what the permanent valuation of the dollar will be. To guess at a permanent gold valuation now would certainly require later changes caused by later facts."

In other words, he is engaged in quantitative easing, but with a defined target rather than a defined quantity. It's not quite NGDP targeting, but it's in the same spirit. In any case, he is clearly trading on expectations: the expectation that he will continue to raise the price of gold indefinitely until his objective is achieved. It follows that the difference between interest rates and gold price as policy levers does not turn on expectations. It is just that the gold price has no upper bound. In the symmetrical problem of reducing inflation, interest rates would have the corresponding advantage.

[OGT]

NR:"But, this post is not about wealth effects."

No, it's supposed to be about how 'real people' behave and think. I used secured mortgage debt because it's more 'real world' and corresponds to the Kityoki-Moore model (link below). But I think it would largely correspond to debtor-saver behavior in any significant financial shock. To expect 'savers' to begin spending when there savings are imperiled doesn't seem to be a intuitive expectation. In a full blown panic this isn't in any way revolutionary, its why we have a lenders of last resort.

In a sense I don't think I disagree with what I take to be your primary point that we're in a multiple equilibria world and (lord help us) the CB's are the one's charged with stabilizing expectations at the socially optimal level.

http://dss.ucsd.edu/~grondina/pdfs/Econ211_KiyotakiMoore1997.pdf

Lastly, I am not sure we should be so quick to 'blithely' compare FDR's Gold Clauses, which involved the committed campaigning of Roosevelt throughout his campaign and in office as well as an act of congress that abrogated the gold clauses of all private US debt contracts, and Bernanke setting a level target. The latter involving a speech and perhaps a paper.

[Winslow R.]

Nick wrote: "For example, suppose you were a Keynesian of some sort. Your immediate instinct would be to answer my original question by saying that an increase in desired saving, if the monetary and fiscal authorities do nothing, would cause a recession. But if you stopped and asked the supplementary question, and I answered "they decide to buy newly-produced investment goods instead of newly-produced consumption goods", you would stop talking about a recession."

...and wonder why we have such different interpretations of the words 'not spending'.

Nick wrote:"In macroeconomics, "saving" is defined as "not spending part of your disposable income on newly-produced consumption goods and services"."

Do these macro-economists include Keynesians of some sort?

At least some sorts of Keynesians believe savings of disposable income is 'not spending' disposable income.

[rogue]

"OK, so if you are not spending it on newly-produced consumption goods and services, what are you doing with your disposable income instead?".

Nick, how would it affect your prescriptions if the answer to that is "I'm not spending or investing until I see that the economy is getting better. By this, I mean, other people (not me) start to consume and invest first."

What if all people who are not part of those who owe (meaning they are those who are owed) feel this way? What if the indicator that they are looking for is consumption and investment by other people, not the cost of money?

[Nick Rowe]

TMF: Thanks for those links to interest rate data. They do help. Pity the time-scale isn't more fine-grained. The sense I get (helped also by Mark Sadowski in comments on Mark Thoma's blog) is that: safe short nominal interest rates stayed about the same, but rates on risky assets fell (presumably because the risk fell?)

Lee: it seems there are two variables that could be changed to get to equilibrium: change the rate of interest paid by the medium of exchange; change the real quantity of the medium of exchange.

Phil: good comment. I had oversimplified.

OGT: "Lastly, I am not sure we should be so quick to 'blithely' compare FDR's Gold Clauses, which involved the committed campaigning of Roosevelt throughout his campaign and in office as well as an act of congress that abrogated the gold clauses of all private US debt contracts, and Bernanke setting a level target. The latter involving a speech and perhaps a paper."

Fair enough. I can't push that analogy too far. But it still serves my purpose, by showing an example of a world where the central bank can do something other than set an interest rate.

"To expect 'savers' to begin spending when there savings are imperiled doesn't seem to be a intuitive expectation."

OK, but remember, they must, necessarily, by an accounting identity, plan to do *something* with their disposable income. The danger, to my mind, is that they use it to hoard the medium of exchange. Which brings me back to my old stuff about the paradox of thrift vs the paradox of hoarding. I see only hoarding (saving in the form of the medium of exchange) as causing a decline in AD and a recession. (I was going to talk about this in the post, but it would have broken my main chain of argument, so I left it unsaid).

Winslow: "Do these macro-economists include Keynesians of some sort?"

Very definitely yes. Old Keynesians, New Keynesians, Neo-Keynesians. The definition of "saving" as "income minus taxes minus spending on newly-produced consumption goods" was, I believe, introduced to macro by Keynesians, and became universal/standard/textbook. You have to go back to pre-Keynesian writings, AFAIK, to get different conceptions of savings, at least in macro.

[jsalvatier]

I agree that people's models of the world affect how the world works; you don't have to convince me. But 'most people hold a model like X' is a statement about reality. It would be much more accurate, clearer and less off-putting if you said something like 'if the Fed changes its language to talk about Y instead of X, it will change people's models of how the world works, in particular their models of how the Fed affects the world and this affects how the world actually works'. If you don't want to claim that people have internal models of how the world works, then say something like 'changes their behavioral properties'.

It's probably OK to use this 'reality' language once in a while, but doing so frequently is misleading and gets you written off as a crank.

[Nick Rowe]

rogue: "Nick, how would it affect your prescriptions if the answer to that is "I'm not spending or investing until I see that the economy is getting better. By this, I mean, other people (not me) start to consume and invest first." "

OK, so what *are* they doing with their disposable income? As I said to OGT, this goes back to my old posts about Says Law and buying antique furniture, etc. To my mind, what matters is this: are they planning to hold more medium of exchange? If they are, then that's a problem.

[Nick Rowe]

jsalvatier: you may be right. But at the same time, if I talk about the CB's "communication strategy", hard-nosed critics will say "sure, but that's just talk; what are they actually going to *do* (with the rate of interest)? What physical levers are they going to pull?"

What those critics fail to see is that almost everything economists model is a socially constructed reality. We talk about prices and quantities exchanged. The things exchanged are property rights. Property rights aren't some physical aspect of reality that physicists can measure. Bernanke is just a guy with a beard. Arthur is just a guy without shit all over him. Neither has any real physical levers.

[Lee Kelly]

"are they planning to hold more medium of exchange? If they are, then that's a problem." - Nick

Unless the medium of exchange has a price of its own. That is, if the medium of account is something other than the medium of exchange, but something like a basket of goods of relatively stable value, ala Greenfield and Yeager ideas.

[Winslow R.]

Nick wrote: "Very definitely yes. Old Keynesians, New Keynesians, Neo-Keynesians. The definition of "saving" as "income minus taxes minus spending on newly-produced consumption goods" was, I believe, introduced to macro by Keynesians "

Post-Keynesians?

I've never learned to appreciate the economic distinction between consumption and very short-term investment except for tax purposes :)

[Nick Rowe]

Lee: Yep. I was implicitly assuming that the medium of exchange and account were the same.

Winslow: "Post-Keynesians?"

The ones I have read, yes.

*Some* (not all) MMTers *sometimes* implicitly use the word "saving" to mean S-I, as far as I can tell. But even then, I'm not sure whether they would include spending on used furniture as savings or not. Of course, non-economists (or self-taught economists who have not been indoctrinated into using the official language) can use "saving" in all sorts of ways.

[Too Much Fed]

Nick's post said: "TMF: Thanks for those links to interest rate data. They do help. Pity the time-scale isn't more fine-grained."

You're welcome. Let me see what I can do...
Try these:

http://research.stlouisfed.org/fredgraph.png?g=X6

http://research.stlouisfed.org/fredgraph.png?g=X5

And, "The sense I get (helped also by Mark Sadowski in comments on Mark Thoma's blog) is that: safe short nominal interest rates stayed about the same, but rates on risky assets fell (presumably because the risk fell?)"

I believe there was real GDP growth and some price inflation after about 1933 or 1934. That probably led to business profit growth and a better ability to pay the debt.

[Too Much Fed]

Nick's post said: "Lee: it seems there are two variables that could be changed to get to equilibrium: change the rate of interest paid by the medium of exchange; change the real quantity of the medium of exchange."

Technically, isn't the interest rate on the medium of exchange (currency or demand deposit[s]) always zero(0)?

As far as change the real quantity of the medium of exchange and I believe the way the system is set up now, the only way to create NEW medium of exchange is from the demand deposits from a loan (meaning debt and has to be borrowed into existence). Maybe someone can verify that.

[Too Much Fed]

"Roosevelt did it. And the only difference between then and now is how people think about what central banks do, and how they do nothing. The change in perceptions about what "doing nothing" means has meant that Roosevelt's concrete policy lever now looks like a confidence fairy."

I need to see some of the accounting of doing away with the gold standard to see if it was QE or QE-like or something more than QE or QE-like.

[Kevin Donoghue]

“But now I've asked it for you, what were you implicitly assuming about that supplementary question when you answered the original question?”

I was assuming that consumption decisions are made by households and decisions to invest (i.e., to buy newly-produced capital goods) are made by firms. The fact that households want to save more is no reason for firms to want to invest – quite the opposite.

“We need to think about how people construct their own economic models, and how they think about "doing nothing". But, like anthropologists, we need to stand apart from the natives' own way of viewing their world. We don't have to believe it ourselves to recognise that they believe it.”

But RE requires that the natives have the correct model. So you’re junking RE?

[Nick Rowe]

Kevin: OK. The standard story is that households consume and firms invest. (But we know that isn't quite right, because households do invest). But that still leaves open the question: what were you assuming (implicitly) that creditors did with their income if they stopped lending it to debtors?

"But RE requires that the natives have the correct model. So you’re junking RE?"

Good question. One I had mulled over when writing this post. No, is the short answer. A full answer would deserve a post. Here's a middling answer:

A "social institution" is a model that people believe. There is more than one possible social institution. So there is more than one correct model. "Rational expectations" is an adjective, not an adjective+noun. It's an adjective that applies to models, not to the world. "RE model" is adjective (RE) + noun (model). "Is this model an RE model?" makes sense in a way that "Are expectations rational?" does not.

[Kevin Donoghue]

I don’t think of creditors as trying to do anything in particular with their income when they are reluctant to lend. I think of them as trying to offload the IOUs they hold, thereby depressing the price (raising the yield). I’ve never really outgrown the first macro model I learned. There’s a portfolio decision which determines the price of bonds, which is 1/r because the bonds are perpetuities. Equivalently, you can think of that decision as determining the opportunity cost of holding money, which is r. You can focus on whichever you like because if the stock of money is willingly held so is the stock of bonds and vice versa. The other two markets are for labour (a Marshallian cross) and goods (a Keynesian cross). As a distinguished New Jersey economist is wont to say: you got a problem with that?

A full answer would deserve a post.

I look forward to it. Your “middling answer” leaves me wondering whether the agents in your model have different models which just happen to give them all the same solution. (If they get different solutions your “short answer” surely ought to be Yes.)

[Marcus Nunes]

Sentiment of "doing nothing" until Nov 2012:
Monetary and fiscal policies cannot be expected to turn this situation around. The US Federal Reserve will maintain its policy of keeping the overnight interest rate at near zero; but, given a fear of asset-price bubbles, it will not reverse its decision to end its policy of buying Treasury bonds – so-called “quantitative easing” – at the end of June.

Moreover, fiscal policy will actually be contractionary in the months ahead. The fiscal-stimulus program enacted in 2009 is coming to an end, with stimulus spending declining from $400 billion in 2010 to only $137 billion this year. And negotiations are under way to cut spending more and raise taxes in order to reduce further the fiscal deficits projected for 2011 and later years.

So the near-term outlook for the US economy is weak at best. Fundamental policy changes will probably have to wait until after the presidential and congressional elections in November 2012.
http://www.project-syndicate.org/commentary/feldstein37/English

[Patrick]

I'm wondering if maybe the lesson is that expectations need to be reinforced with reality. People learn. Expectations about tomorrow are formed by experience today. It's not enough for the CB to simply promise something. It has to actually do it.

The CB has to be ready to have its bluff called. Those betting against the CB have to get smacked. If some people are seen to make money shorting the CB, there goes their precious credibility.

Roosevelt didn't have to worry too much about a higher future price level. He had a mechanism to deliver a higher current price level. So doing nothing tomorrow was an option. The rest kinda looks after itself. I think the same is true of inflation in the '70's. The Fed raised interest rates, and shrunk the money supply. There's no arguing with or misinterpreting that, but if you didn't learn to set expectations accordingly, you got smacked and recalculated/learned for the next period. Once enough people got smacked, expectation fell into line.

[charles]

I actually asked myself the second question, but I don't have the same answer. For an individual, there are THREE types of investment goods :
1) investment goods that will produce wealth in the future (the classic one)
2) investment goods that won't produce wealth in the future ! Of course the investors don't know it when they take their investment decision, but the lure of getting in the current fad, simply sometimes for social pressure reasons ("everybody" buys real estate, golf memberships, etc, is very strong. It is actually consumption, but people don't realize it. It is becoming prevalent in "saving" Asia.
3) investment that reduce the ability of other economic agents to compete for scarce resources (commonly known as hoarding, but it is actually a defining feature of most property rights).

Type 3 is anathema for macro-economists and the political leaders who employ them, because their economical position is essentially a roughly constant share of the overall cake, hence the powerful incentive to grow the cake at all costs, the easiest way being by way of population growth. Nevertheless, it is a perfectly rational strategy for individuals or group of individuals. The more they are ressource constraints on the economy and the easier it is to implement hoarding (through superior repression technology), the more Type 3 investments appear attractive. This is where we are now.

[Richard H. Serlin]

“For example, if you say that the average Canadian owes $45,000, you think one way. If you say the average Canadian is owed $45,000, you think another way. But in a closed economy, or one with no net foreign debt, debits equal credits, so the two are the same.”

Yeah, but median could be a whole different thing. 90% of the people could owe an average of $45,000 to 10% who are owed an average of $405,000.

[Scott Sumner]

Nick, As I recall FDR's decision to raise gold prices had little impact on risk free rates, but lowered risky rates (as they generated rapid economic growth.) Indeed we can have a high degree of confidence that the risk spread fell, as there were many different policy signals that depreciated the dollar throughout 1933, the effect on asset prices was quite strong, and sign was pretty consistent.

[Morganwarstler]

"Half the population is in debt to the other half. Then something happens and the debtors can't continue to borrow and dissave. Nobody will lend to them, so they are borrowing-constrained. So they consume less. Aggregate savings rises, and the central bank has already pushed the rate of interest down to 0%, so can do nothing. So there's a recession, unless the fiscal authorities do something.

"The debtors stop borrowing from the creditors". "The creditors stop lending to the debtors". Those are two different ways of framing the same event. The first sounds like it would lead to an increase in aggregate savings, as the dissavers stop borrowing. The second sounds like it would lead to a decrease in aggregate savings, as the savers stop lending. What are the savers going to do instead, if they can't save by lending to the debtors? They can't just do nothing?"

This sounds smart but it isn't.

WHAT WE ARE CONCERNED ABOUT: the 50% of the savers - the ones who have hard assets to their name - have been playing musical chairs, and the music stopped, BUT the over-leveraged did not get guttedlikefish.

Instead the 50% who don't have anything were used as an excuse to save the over-leveraged "savers" (read bankers).

IF you are serious about getting the music going again, you HAVE TO FORM POLICY to gut the losers like fish.

if you won't do it, you aren't serious about starting the music.