« Economy Lab | Main | Should we teach supply and demand in French? »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

I disagree.

Let's assume there are a large number of agents, say a continuum.

There are many equilibria, start in an equilibrium where nobody ever tries to save. No bonds exist and output is at its maximum, say 100%.

Now suppose that every day 5% decide they'd like to save (a different 5% each day), they will offer a backscratch today but won't consume one. Instead they'll take a bond. First problem, does a non-saver, who expects to be a non-saver tomorrow accept the deal? Yes, tomorrow he pays back the backscratch and consumes by issuing his own debt to one of the savers. Today total output of backscratches is 95% of max.

Tomorrow arrives and yesterday's savers consume 2 backscratches and provide 1. Those who issued debt provide a backscratch to redeem the maturing bond and issue debt to the current savers to consume.

It's easy to see that we know have an equilibrium in which output is 95% of the max forever. Yet there is no hoarding of bonds and bonds are not the medium of exchange. Most transactions are bilateral exchange of service.

So, a 5% increase in aggregate saving decreases income by 5%. That's not a paradox of thrift?

PS: and of course *aggregate* saving remained zero. In aggregate the reduction in consumption from 100 to 95 was not offset by aggregate consumption of 105 tomorrow. Aggregate saving was still zero, some individuals manage to save some periods but the net result is just a fall in income.

Basically the idea is it is like an endowment economy where I don't let the real interest change (it's fixed at zero) and I don't let the aggregate price level change (because there's no money) so people end up throwing away part of the endowment. The things that adjust to make us, in aggregate, happy to consume the aggregate endowment can't adjust.

"Fed hiring economists with new money is both fiscal and monetary."

In what way?

Wouldn't Fed purchases of the S&P 500 also be considered fiscal and monetary?

Nick,

I thought so.

The problem I have is that your thrift with hoarding and thrift without hoarding examples have the same macro result. There is no difference in the amount of macro income and macro saving in those two examples. The purchase of antiques is a balance sheet transaction. All it does is swap the ownership of money and antiques, without affecting anything else. In particular, it leaves the level of thrift and hoarding the same at the macro level. This is the problem I saw from the beginning. The paradox of thrift is a macro result corresponding to macro income and saving effects. I don’t think you can construct a paradox of thrift example where thrift isn’t equal to hoarding at the macro level.

Winslow R.,

Hiring Fed economists is effectively part of fiscal policy.

It’s an expense that affects the Fed’s net income, which is remitted to the Treasury, which affects the budget balance, which is fiscal.

That may be a fairly narrow accounting view of it.

But it’s also a fact.

Andy,

Are you an economist?

The reason I ask is that most economists (and certainly me) are focused on the flow of output--using resources to produce goods and services so that people can best achieve their goals.

Economists interested in monetary theory and banking constantly run into people whose focus suggests more of a focus on financial markets. For example, if people fear that stock prices may fall a lot, then maybe they won't buy stocks. My view of that is.. so?

I am a bit puzzled as to why we would be focused on assets that have a constant return forever. What?

Adam: Now we are getting to the *real* issue. This isn't "procrustian economics" (I can't spell "procrustian" either) of trying to force everything into a quasi-monetarist perspective. This isn't about "re-naming" the paradox of thrift, or just re-assigning "blame" for the problem. This is a real substantive disagreement. Your last comments show I'm either wrong, or am saying something that matters.

So, in your example, 5% of the backscratchers are idle. They have unsold backscratching services in period one ("today"). Why don't that 5% all agree to scratch each others' backs? If we assume there are gains from trade (that the marginal utility of a second back scratch is greater than the marginal disutility of scratching the first back) they would agree to that trade. (And if the MU of a second backscratch is worth less than the M disutility of providing the first, they would never have accepted the bond in the first place.)

Again, in a barter economy, there cannot be a deficiency of aggregate demand. Keynesian macro makes sense in a monetary exchange economy. Keynesian macro makes no sense in a barter economy. Keynesians need to understand that. They need to make their implicit assumption explicit. In a barter economy, the idle car workers and the idle farmers can just produce cars and food and swap them. But in a monetary exchange economy, the car workers won't produce cars unless they can get money from the farmers, and the farmers won't produce food unless they can get money from the car workers.

Nick,

I’m trying to understand this.

In a barter economy, why is it assumed that idle workers in industry 1 and idle workers in industry 2 would be willing to swap each other’s products if they’re employed instead? In the event they have no incentive to do so, why must they be employed?

Is it that auctioneer fellow?

If so, why is it assumed that the auctioneer is so smart?

Isn’t that rather wishful thinking?

But Nick, that's the thing. The savers want to trade one less backscratch today (leaving them with zero today) for an extra one tomorrow (so they'll have two tomorrow). They succeed so they are maximizing.

Everyone else wants 1 today and 1 tomorrow. They also succeed so they are maximizing.

There are no gains from trade left on the table. I think you're wrong.

Now, if rates could be reduced (say by policy) so that the savers don't want to save anymore then we can bring output back up to 100. But at the prevailing real rate (zero in this case) the exogenous increase in aggregate savings leads to a drop in aggregate income *without* leaving unrealized gains from trade. Given the prices they face everyone is maximizing.

And all without any money, not even hoarding of bonds. No hoarding of the medium of exchange (since there is none).

I’m having a mind block trying to resolve the back scratch GDP progression.

If yesterday’s savers consume 2 today, and today’s savers consume 0, why is today’s GDP not 100 rather than 95?

I just created a blog. My first post is about the paradox of thrift. I actually wrote this a while back along with a counterargument much along the lines adopted by people here. I intend to rewrite the counterargument and post it soon, because I want to make the argument that, overall, central banks actually help cause the paradox of thrift.

http://philosophyandeconomicsblog.blogspot.com/

yes, I guess in period two income is back to 100.

Nonetheless the example still works, the increase aggregate saving causes income to fall in the first period but it is never greater than 100 in subsequent periods. Aggregate saving is still zero, you still have an example of the paradox of thrift without any money.

There is something analogous to a paradox of thrift in a non-monetary economy.

Suppose that you are an apple farmer. You spend x resources to produce y apples, so your income is y apples. You do not want 1000 apples, so you try and trade some for meat, bread, and whatever other goods you like. But suppose that you save 200 apples and place them in cupboard to consume 3 months from now. This is obviously not a wise thing to do, because apples depreciate rapidly and will be nothing but waste in 3 months. Saving apples in this manner is clearly destructive. A better way to "save" apples would be to lend them to someone who consumes them immediately and promises to return the same quantity of fresh apples in 3 months time.

When there is monetary equilibrium, saving money allows resources to be used to produce furture goods and services. Although the composition of spending between consumption and producer goods changes, total spending remains constant -- no paradox of thrift. This is bit like lending fresh apples now so that someone can return fresh apples in 3 months. The problem arises when there is an excess demand for money, because then saving money makes idle resources, i.e. the financial institutions that normally redistributes spending stop functioning. Savings just sit like mouldy apples in a cupboard, unemployed and depreciating, squandering precious time when they could be used for productive ends.

Does any of this make sense? I haven't had much success trying to explain this in the past.

Adam: OK, I understand you now, I think. If you "need" a backscratch, it gives you MU of 1. If you don't need a backscratch, it gives you a MU of 0. Savers need 0 today, and 2 tomorrow. Everyone else needs 1 today and 1 tomorrow. So if there are 95% savers the optimal employment level is 95%. That's what the central planner would dictate.

The labour demand curve (VMPL curve) has a horizontal section at 1, then goes vertical at 95. It's sort of Inverted L-shaped.

So, just to check I've understood your model (and to answer JKH's question), then is we assume 5% savers today, and no savers thereafter (for example), GDP is 95 today, and 105 tomorrow? (people can supply two backscratches per day if it's needed).

OK, assuming I've understood you correctly, your model makes sense. But to my mind, that's not a Keynesian model; it's an RBC model. The output and employment fluctuations are efficient. We make hay when the sun shines. The Value Marginal Product of Labour always equals the marginal disutility of labour. It's just that the VMPL fluctuates over time. Now, just as in RBC models, the government could use fiscal policy to intervene, and buy backscratches during recessions, but it's not efficient for them to do so.

RBC guys call those output declines "recessions", but it's not what I think of as a recession. It's not a Keynesian recession.

JKH: the Keynesian assumption (which I share) is that unemployment in recessions is inefficient. If we could only get the unemployed back to work, the extra output that could produce would, in aggregate, be sufficient to compensate the unemployed for giving up their "leisure".

If we could get all the unemployed together in one room, and strike a deal where the car workers produce cars for the farmers, and the farmers produce food for the car workers, with both groups keeping their money in their pockets, everyone would be better off. But we can't get everyone together into one room (because it's not just cars and food it's millions of different things). And everyone wants to hang onto his money. So the car workers won't buy food for money, and the farmers won't buy cars for money.

Bill,

I am an economist, though I work in the financial industry, so perhaps that accounts for (or results from) my focus on financial markets. But my example doesn't necessarily involve "financial markets" as such. It's about assets, which aren't necessarily financial (and in a barter economy, I don't think the word "financial" would even be meaningful). A house is an asset, for example, but it's not financial. And surely asset price stability is important. Why build a factory if you don't know how much the factory will be worth once you build it? And if asset prices are highly sensitive to certain parameters, a lot of resources get diverted out of productive activities into estimating those parameters (and convincing people about what the parameter values are). I gave an example of an asset whose return grows at a constant rate forever just as an extreme case, but I don't think it's too far from reality: we've seen instability in stock and real estate prices; some people call that instability bubbles, but I call it excessive sensitivity to parameter values. And it seems pretty obvious that this instability has important real effects.

Nick,

"If government spending and/or taxes change, it's fiscal."

"Government spending" is an ambiguous phrase. I assume it includes unilateral transfers, but does it include asset purchases? And does it matter who (the Treasury or the central bank) does those asset purchases? (Seems pretty arbitrary.) If it doesn't include asset purchases, does that mean that the TARP was not fiscal policy? If it does include asset purchases, does it also include conditional asset purchases (government guarantees)? Or does a government guarantee suddenly become fiscal policy if the government has to make good on it? And if it does include asset purchases, does this mean that "credit easing" is fiscal policy?

Actually GDP is 95 today and only 100 tomorrow. Remember we've assumed a max output of 100.

What happens when tomorrow arrives is the 5% yesterday's savers consume 2, one from a bilateral exchange and one that is owed to them.

The 85% of yesterday's non-savers who are also non-savers today consume one in a bilateral exchange.

The 5% non-savers with debt provide a backscratch to yesterdays savers, thus redeeming the debt. They then consume 1 by issuing new debt to todays 5% cohort of savers.

Todays 5% cohort of savers consumes 0.

This does require that there be a group of 5% who know they never want to save so they can go forever issuing debt and rolling it over (remember the assumption that you can only give one backscratch per day, even though you can receive more than 1). Effectively this group are financial intermediaries.

Again though, the point is that the exogenous increase in the aggregate propensity to save resulted in a fall in income and no aggregate savings. And all without having any hoarding of the medium of exchange.

Adam: OK. One day a saving virus breaks out, and affects a different 5% of the population each day, but each saver fully recovers the next day, and continues forever. So GDP goes 100, 95, 100, 100, 100 etc. But again, it's an RBC model.

Andy: yes, distinguishing exactly between monetary and fiscal can get pretty arbitrary. Transfer payments are normally seen as negative taxes, rather than government spending, because the government is not buying newly-produced goods and services. (But that can be a judgment call).

If you buy the assets with new money, I would call it monetary. If the assets are newly-produced goods, bought with new money, it's both monetary and fiscal. If you swap one financial asset for another, it's...whatever. A portfolio operation?

How's this an RBC model? No productivity shock, nobody withdrawing their labour due to temporarily low real wages. Nothing driven by the supply side.

GDP falls due to a shortfall in aggregate demand, this is Keynesian. Entirely Keynesian.

It looks like the 'standard view' and MMT both have operationally imposed limits.

The 'standard view' says monetary policy rules because new money is created by the Fed. But the Fed can't buy beyond its operational surplus without political consequences.

The MMT view says fiscal policy rules because new money is created by the government deficit spending. But the government operationally must issue bonds to offset deficit spending.

We could have a showdown at the OK corral and see which institution is first to break its operational restriction and then retains strength enough to keep operating. I'd bet on the government.

Otherwise we could continue to operate within the established framework where the government deficit spends and the Fed buys the resulting bonds.

Nick wrote: "If you swap one financial asset for another, it's...whatever. A portfolio operation?

The one you are talking about? Corruption.

Adam: it's a preference shock, rather than a technology shock. The Marginal utility of consumption falls to zero when output exceeds 95. Backscratches are no longer a scarce good. There's satiation in backscratching services. People might as well sit idle. You couldn't give away backscratches for free. The central planner would replicate the competitive equilibrium. Recessions are efficient responses to changes in preferences.

The MRS between consumption and leisure falls to zero past C=95. It's formally equivalent to a labour supply fall, because people place zero marginal value on what they can consume with their real wages, past 95. The unemployed already have had their back scratched once, and don't want a second backscratch.

Nonetheless, is it a paradox of thrift or not?

I'm quite sure we can tweak the basic example to make the fall in income inefficient without needing money.

and anyway, what kind of shock is an increase in the demand for real balance then?

Nick wrote: "And "money" means medium of exchange."

Let's assume two mediums of exchange, currency and demand deposits created from currency denominated debt.

If loan losses exceed capital for the losses, does that mean demand deposit defaults? If so, can there be less supply of one medium of exchange instead of more demand for medium of exchange?

Nick wrote: "Fiscal deficits, if they work, will have future costs. The higher debt will mean very difficult future spending cuts or tax increases."

What if the fiscal deficit was in currency? Would that be present costs instead of future costs?

"If you buy the assets with new money, I would call it monetary. If the assets are newly-produced goods, bought with new money, it's both monetary and fiscal. If you swap one financial asset for another, it's...whatever. A portfolio operation?"

I thought the link I posted was arguing that the QE1, wasn't stimulus. I just throw up random comment's sometimes. I was referencing these earlier posts. April 09, August 2010 I checked the bond numbers and it did look like the gov of Canada had more on offer through the fall of 2008/2009. I honestly don't know anything about the U.S. QE1, but the post I linked to seemed to argue that it was a one time thing to make up for the mark-to-market error of accounting. And not as monetary stimulus per-se. Read through the Gorton papers and mark-to-market is apparently a big deal during a bank-run. Reason I posted it was that you and AdamP were arguing about buying stock. I'm just wondering how you want the Fed to increase money supply, and is it the currency money supply that needs to increase or some other.

But I don't really care either, I would much prefer that you, Adam, Andy, Bill get this sorted.

Is the price of a backscratch bond pegged at par, i.e. one scratch tomorrow for one scratch today? If so I'm pretty sure you can have a paradox of thrift. I really don't think you can have one without a rigged price, barter or no barter.

Nick, re your last reply to me: I haven't read Barro and Grossman 1971; tried once and got tired. I'll have a look later and see if it has improved.

yes Kevin, this was my point to Nick above. The model is Keynesian since the problem is that the real interest rate fails to adjust. I made this point to Nick somewhere above.

Although, any adjustment to the real rate would have to be for it to fall below zero. Since I've imposed a max output of 1 backscratch per "employed" labourer and no growth the real rate can't be positive.

For what it's worth I agree with edeast that the buying of MBS by the fed wasn't monetary stimulus per se, in the sense of changing the available rates of intertemporal substitution.

It was preventing a total collapse in the money supply when a lot of paper that was functioning as money ceased to. After all, for a while there the fed was a net seller of treasuries as they bought up the MBS.

Adam: your model is of a pure consumption economy, with no investment or growth. But you could still have a positive equilibrium rate of interest, if there's positive time preference.

No matter. Suppose there's a law that pegs the price of bonds too low (the real rate of interest too high, sort of like usury laws, only in reverse). So there's an excess demand for bonds in the market in which bonds are traded for backscratches. Remember that the short-side of the market in disequilibrium determines the quantity actually traded. People will want to sell backscratches for bonds, but won't be able to. But if this is a barter economy, where backscratches can be traded for backscratches, that market can still clear. If the cost of giving a backscratch is less than the benefit of receiving one, people will scratch each others' backs. Sure, they would prefer to get paid in bonds instead, but if nobody is willing to pay them in bonds (because of the excess demand for bonds at the pegged price), they will accept payment in current backscratches instead, as the next best option.

Kevin: yes, you do need some sort of price that is pegged too high to get Keynesian unemployment. But in a monetary exchange economy, you can get unemployment if the price of everything is pegged too high against money (i.e. the price of money is too low), even if all relative prices are at the right levels (i.e. the real wage W/P is at its equilibrium value). The "classics" said that an excess supply of labour meant that W/P was too high. Barro Grossman showed you could get an excess supply of labour in a monetary exchange economy even if W/P was right, but both W and P were too high relative to M.

Yes, Nick I'd implicitly assumed zero time preference.

Nonetheless, the question is still out there. Paradox of thrift or not?

Increase in the (average) individual propensity to save leads to lower aggregatge output (and some unemployment) and zero aggregate savings. Is that not the paradox of thrift?

Clearly no hoarding of the medium of exchange. No hoarding of anything in fact.

Adam: paradox of thrift or not?

What was paradoxical about the original paradox of thrift was that thrift could be good for the individual but bad for all individuals taken together. It makes the individual wealthier and better off (over time, of course, because there's a current cost of foregone consumption) but if everyone does it (under certain conditions) everyone is poorer and worse off. A law banning thrift could make everyone better off. In your model, thrift does not make everyone worse off. A law banning thrift would make everyone worse off, if some people really do need 0 backscratches today and 2 tomorrow.

So yes, there's a recession, and unemployment, but it's a good recession and good unemployment. Now you could tweak your model to change this. Introduce some positive externality to employment and output, for example, and you could make recessions and unemployment into bad things. But even so, unemployment wouldn't be an excess supply of labour. Again, you could tweak your model to add monopoly unions or efficiency wages, so you got an excess supply of labour because real wages were too high. And with the right tweaks, you could probably get real wages to rise relative to equilibrium real wages when time preference changed. But it's still not Keynesian.

I wish I could convey my intuition more clearly. But in a frictionless barter economy there are more markets than in a monetary exchange economy. You can trade labour directly for produced goods, for example. So when a price control in one market prevents mutually advantageous trades in that market, people can get around the blocked market, and it doesn't prevent them trading the correct amounts in other markets. But if we live in a hub-and-spoke world of monetary exchange, where all trade takes place via the hub, which is money, any problems with that hub will affect trade everywhere.

Nick: "What was paradoxical about the original paradox of thrift was that thrift could be good for the individual but bad for all individuals taken together. "

I don't think that's right. The paradox is that it doesn't turn out to be good for the individuals (in sum) because (in sum) they don't succeed in saving (even if a few individuals do succeed).

It's like the whole chinese control of the capital account thing. Whenever I ask people if they agree that China controls their capital account with the US (since they have capital controls and we don't) they always agree.

I then ask if that implies that they control the current account with the US, most people answer no. However, you and I both know they're wrong. But then this puts the US in paradox of thrift land, since (X-M) will always be negative you have that Y < C + I + G, so either the private sector dissaves or the public sector dissaves (or both) and that's true no matter what the level of Y!

This is the paradox, if the US (public and private) attempts to stop dissaving by reducing expenditures they fail! They reduce C + I + G and the only thing that happens is that Y falls by an even greater amount. The attempted saving is not beneficial to the individuals in aggregate (perhpaps it benefits sum but their benefit comes at the expense of others and so in sum total the individuaals do *not* benefit).

Now, in my model the 5% who are "unemployed" and forced to finance their consumption by borrowing don't want to do this. If we say that they are even taking a risk becuase now if they catch the virus in the future they'll be unable to save then they are worse off. Yet they do it because the risk associated with going in to debt is still preffered to being unable to consume the backscratch today.

Thus, if you could reduce the real rate so the savers now willingly go back to 1 backscratch today and 1 tomorrow then you'd improve welfare.

The idea was never that if they don't get the 0 today, 2 tomorrow consumption path they die, it was that they strongly prefer this while at the same time the unemployed would rather go in to debt and consume today (for sure) then try to force the bilateral exchange of service and risk getting nothing. Thus, the savings occur but a lower interest rate could prevent it. The point though is that at the real rate of zero the savers get their way at the expense of the unemployed. If you reduce the real rate to make them not want to save anymore then they are no worse off but the otherwise unemployed are better off.

Adam: I like your China example. I think you are onto something. If some people succeed in saving, they force others to dissave (in some circumstances) is non-obvious, and therefore deserves to be called a paradox. And if you want to call it a paradox of thrift, I can't really object, since it is a paradox and it is about thrift.

I still think my interpretation ("What was paradoxical about the original paradox of thrift was that thrift could be good for the individual but bad for all individuals taken together. ") is the original, standard one, but that's by-the-by.

Two points.

1. I did an old post, maybe nearly a year back, essentially supporting what PK said about China forcing the US to dissave, but arguing that money was an essential part of this story, because it forced the Fed to print more money to avoid a recession, and that it was attempted saving in a monetary economy that forced others to dissave or accept unemployment. In a non-monetary economy the US would and could just refuse to borrow from China, if they didn't want to. I never did get my head fully around that subject, because it's harder in an international context.

2. "Now, in my model the 5% who are "unemployed" and forced to finance their consumption by borrowing don't want to do this." That's where I disagree. The unemployed have a choice: they can finance their current consumption by borrowing; or they can finance their current consumption by telling the potential lenders to get lost and trading backscratches between themselves in barter. If they prefer the latter, the savers fail to lend, and so fail to save.

Adam: Here's another way of looking at it. If you want to save (and not invest) can you *force* someone else to dissave? You can't force someone to borrow from you if they don't want to borrow from you. But if you stop spending your money, and stick it under the mattress, nobody can stop you doing that, and it forces other people to hold less money (edit: or print more).

I wish I could clearly apply this insight to China.

Well I wasn't claiming that in real life the paradox doesn't manifest itself in money hoarding, certainly the problem with China is that they hoard our currency.

But you were making a theorectical point that I think is wrong, so I started thinking about a counterexample. It doesn't *have* to be about money hoarding because it can in fact happen in a world without money. The counterexample is stretched because I thought it up in a few minutes but I think it makes the point.

Even if the savers go to one back scratch tomorrow, won't they prefer to use their bonds rather then giving an exchange, dropping gdp back to 95. Still force people to save.

Adam: "..certainly the problem with China is that they hoard our currency."

That's what I assumed, in my old post, then commenters (Scott and others) pointed out that they hold US Tbills, not currency. But I still think, in a monetary exchange economy, the point still stands. Others can, in effect, force you to borrow from them, even if you don't want to.

"If you want to save (and not invest) can you *force* someone else to dissave?"

That's the point I was making way back.

That's exactly how the paradox of thrift manifests itself in a service economy.

Actual micro saving forces other micro dissaving - in order for macro saving to be zero.

The paradox of thrift isn't just an ex ante condition - it forces out actual economic events and the accounting events that measure them.

JKH: yes, but does *desired* micro saving manifest itself as actual micro saving and force others into actual micro dissaving? If the saving is hording, then yes. If not, then no.

yes, one of the reasons I didn't comment on Nick's post right away was that JKH and Andy were doing a fine job (I thought) in explaining why it's wrong.

"does *desired* micro saving manifest itself as actual micro saving"

Yes, I've asked myself that exact question.

The answer must be yes. You derive this by asking how would you attempt to construct an actual economic event in a service economy that could be described as the live consequence of the "paradox of thrift" in action without actual micro saving taking place. You can't do it. Somewhere, somebody must save at what turns out to be an excess in its macro consequence in order to get the ball rolling - i.e. to pass the tipping point for some economic consequence actually coming out of the paradox of thrift.

China gets paid in money (dollars) for its current account surplus. In that sense, it’s necessary for China to hoard dollars in order to save (CA surplus) and force dissaving (CA deficit) on the US.

The Chinese CA surplus equates to a component of Chinese GDP and income. It’s hoarding money in the sense of the flow of income that is paid in dollars and the corresponding saving that first appears as dollars.

But it swaps that money for bonds. That’s an asset or balance sheet transaction, subsequent to the income and saving event.

It requires hoarding of money in the sense the money is required to execute the subsequent balance sheet transaction.

But it does not require hoarding of money as a stock beyond that transaction. In fact, it requires the opposite (at the "micro China" level relative to the global macro).

On Beckworth's blog you said.... JKH: If a bank loan is paid down, and so loans and deposits both fall (and stay down, let's assume), and if those deposits were (say) chequable demand deposits, and so a medium of exchange, then the supply of money falls. If the demand for money stays the same, we now have an excess demand for money.

Do deposits necessarily fall when a bank loan is paid down? Banks receive a prepayment, but they don't send their depositor's money back. They buy 2-year t-bills or a FRE Hybrid arm.... hence, if anything, the private sector's negative demand for debt (bank loans) means that we will still have an increase in demand for more-money-like securities. or am I crazy?

In the "standard model", an individual who *desires* to save more can always do so, because in the standard model "save more" means "buy less consumption" and you can always buy less consumption if you want to. You don't have to get the sellers' agreement to buy less. But the standard model ducks the question "so, what does he do with his monetary income instead?". If the answer is "buy more antiques", then it's problematic, because you do have to get the sellers' permission -- he has to desire to sell more. If the answer is "hold more money" the answer is unproblematic.

JKH: this is right back at our old "accounting vs economics" argument. Quantity demanded is not the same as quantity bought. A change in demand can have an effect, and is real, even if it makes no change in quantity bought.

Mr Rearden: The person has $100 in his chequing account, which he uses to pay down his loan. So there's $100 less in demand deposits and $100 less in loans.

Nick, seems to me that JKH is just imnposing market clearing.

“Do deposits necessarily fall when a bank loan is paid down?

Banking system deposits typically* fall when a bank loan is paid down. The money used to pay down the loan will typically be sourced from a positive balance in a bank deposit account at some bank.

If the loan and the deposit are at the same bank, it’s a balance sheet wash for that bank – assets down; liabilities down.

If the loan and deposit are at different banks, the bank with the prior loan will gain reserves at the central bank and the bank with the prior deposit will lose reserves at the central bank. The transfer of reserves mirrors the interbank payment that’s required to settle the transaction between the two banks. The typical response to the reserve change at each bank is a money market transaction. E.g. the bank long reserves may buy treasury bills; the bank short reserves may sell them; the net system effect in this case would be a wash.

The bank with the prior loan will also free up capital for deployment somewhere else. It may make another loan, at which time it might sell the treasury bills to net out the reserve effect of making a new loan. In making the loan, the proceeds will “land” in a new deposit somewhere in the banking system.

Central bank reserves are a matter of operational transaction accounting. Capital deployment is a matter of internal risk accounting for a given level of capital.

* Not always, I suppose. The borrower could sell a treasury bill or some other asset to the banking system for example. But I think the natural base case for adjustment is through asset-liability balancing rather than asset-asset balancing.

"JKH: this is right back at our old "accounting vs economics" argument. Quantity demanded is not the same as quantity bought. A change in demand can have an effect, and is real, even if it makes no change in quantity bought."

Yes. That seemed to be what you were doing in your post example, and I got a bit lost as to why you would pursue that. It seemed like you were simulating the paradox of thrift as a risk that didn't materialize for the reasons you gave - as opposed to the paradox of thrift as a risk that was present and that also materialized. I'm saying the materialization of the risk of the paradox of thrift requires micro saving by some unit that then throws the overall thing into a downward spiral.

I'm just not sure why you would deal with the paradox of thrift as a risk that doesn't materialize. The paradox contends actual follow through in the form of GDP contraction.

"It may make another loan, at which time it might sell the treasury bills to net out the reserve effect of making a new loan. In making the loan, the proceeds will “land” in a new deposit somewhere in the banking system."

Sorry, that's wrong. If it sells the bills to a non-bank, the money won't end up as an additional banking system deposit. The new loan will just be replaced by bills. If it makes the loan and does nothing else, the money will end up as an additional banking system deposit.

permutations, permutations ...

JKH: "I'm just not sure why you would deal with the paradox of thrift as a risk that doesn't materialize. The paradox contends actual follow through in the form of GDP contraction."

Here's another way of looking at it. The standard model says it does materialise (in a recession). I'm saying that it can only materialise if it's hoarding, so the standard model implicitly assumes hoarding.

JKH: Okay that makes sense, but like you said: it's like a poor assumption to think that debt is always paid down with savings. It seems like incomes are currently being disproportionately allocated to debt repayment in which case, incomes are deposited. A higher fraction of which is used to pay down debt. Thus, net-net, the banks' liabilities drop LESS than its assets. Doesn't this create more demand closer-to-money substitutes?

*Doesn't this create more demand for closer-to-money (2 year t-notes rather than a construction loan) substitutes (assets)?

"Here's another way of looking at it. The standard model says it does materialise (in a recession). I'm saying that it can only materialise if it's hoarding, so the standard model implicitly assumes hoarding."

I think that's OK (maybe except for Adam's model, but I think you're assuming a monetary economy).

I know you object to being accused of merely renaming, since you have a point of substance. But maybe the paradox of thrift does implicitly assume it is money that is hoarded from income - without being explicit about what it is that it is being implicit about :) Again, my simple interpretation is that income in a monetary economy is paid in money; therefore saving from income is saved in money, at least initially; therefore the paradox of thrift is about the paradox of saving money, or hoarding money. And I say initially, because saving money from income (as a flow) doesn't preclude using that money subsequently to acquire bonds or other assets. But maybe you disagree with that last part.

MrRearden,

At the micro level, I think there’s no doubt that some people are using saving to pay down loans.

From a balance sheet perspective, saving increases the personal or corporate equity position and the cash asset position (e.g. bank deposit). They then use the cash to pay down the loan.

Alternatively, they can use the cash to acquire other assets.

The logical type of asset to acquire given the environment is a low risk one – e.g. your two year note. This is consistent with the risk aversion on the liability side of paying down loans.

So the result is some risk-averse combination of shedding liabilities and increasing low risk assets, which I think is consistent with what you’re saying.

Nick, how close are you to the all time comment record?

Frances: about 50 short, I think. A couple of old posts on economics and accounting, and MMT, I think, got past 200. I'm really running out of puff on this one though!

Nick,

Off topic - this thing from Krugman reminds me of some post you did, I think.

http://krugman.blogs.nytimes.com/2010/10/04/a-note-on-currency-wars/

Nick:

No comment on the BoJ's announcement that it is looking into buying ETFs and REITs??

http://www.boj.or.jp/en/type/release/adhoc10/k101005.pdf

I had a look at the Barro-Grossman paper which is here (1.6 Mb PDF file).

If I understand Nick correctly, his point can be seen by looking at Figure 1, page 86. For anyone who has difficulty opening the file, this diagram is just a standard labour market diagram with supply and demand curves for labour, both ‘notional’ in this context. The curves cross at point A. To the left of that point is point B (same real wage, lower output), which represents a situation where “failure of the price level to adjust to clear the commodity market leads to excess supply in the labor market.” So B is off both curves.

Nick’s point, I take it, is that firms in this situation could hire workers on a barter basis, giving them a share of the extra output instead of money and making additional profit, also in the form of output. Obviously this workaround is fine for breweries, less so for makers of insecticide. But it’s true as far as it goes and it’s a fair criticism of Patinkin, whose model (according to Barro and Grossman; I haven’t done any fact-checking) predicts such a [dis]equilibrium. It doesn’t lay a glove on Keynes, whose model predicts that the price level will fall, driving the real wage up until it hits point D on the demand curve, directly above B.

Keynes’s model has a problem too, since it predicts a countercyclical movement in the real wage. Barro and Grossman note that Keynes struggled with this criticism (see GT Appendix 3). But that’s an empirical matter, not a failure to make sense.

I find barter models distracting. I think the better approach is to assume away monetary disequilibrium, and consider what happens in a money economy.

First, suppose there is an increase in thrift. I don't think of this as a lower discount rate, but rather a shift in the supply of saving to the right. At the initial real interest rate, saving is greater than investment.

If the real interest rate falls, the quantity of investment demanded rises. Firms purchase more capital goods. Also, the quantity of saving supplied falls. Households purchase more consumer goods. The usual effect is a shift in the composition of demand away from consumer goods to capital goods. And so, a shift in the allocation of resources away from the production of consumer goods to the production of capital goods.

But wait.. Suppose that the decrease in the quantity of saving supplied and increase in the demand for current consumption includes an increase in the demand for leisure. Then this response involves a decrease in labor input and a decrease on current output and income. The increase in thrift has resulted in less output and income. Now, this isn't a "paradox of thrift" as it stands. The amount saved should increase. And, of course, there is no market failure involved in this process. And it is a type of RBC.

Since real output and real income are lower (in the short run,) so is real aggregate demand. Still, the "problem" wasn't caused by an inability to sell, but rather by an increase in the demand for leisure (and so decrease in the supply of labor.) Another way to say this is that some people were working to save, and with lower real returns on saving, choose to work less. (Further, if the initial increase in the supply of saving was at least partially funded by working more, then this effect would be partially or wholy, or more than, offset.)

Now, suppose that there is a price floor in credit markets so that the interest rate doesn't clear. This, of course, is an old Yeager thought experiment, and it is supposed to explain that "interest rates" are the wrong way to think about monetary policy.

Anyway, the interest rate is above the market clearing level, and so the amount saved is greater than the amount invested. Some of those who would like to save are frustrated. Now, if we assume that the frustrated savers just hold more money, then we have an excess demand for money, and reduced expenditures. This possibility will be assumed away. Perhaps the demand for money is unchanging.

Since some saving is frustrated and the actual amount saved will be limited to the amount invested at the interest rate floor, there is no choice but to consume. Sure, there are gains from trade that are frustrated. People would be willing to give up consumption today, free up resources to produce capital goods today, and obtain greater consumption in the future. But, the price floor makes it impossible.

But suppose to some degree the good consumed by frustrated savers is leisure. Frustrated savers don't buy consumer goods, they work less. There is less income, and less output. The price floor on the interest rate results in people choosing to work less. With real output and income lower, real aggregate demand is lower. Still, the problem isn't an inability to sell. It is that people don't want to work if they cannot save and obtain future consumption.

Now, let us drop saving for a minute, and consider a price ceiling on gasoline. This leads to frustrated buyers. Now, if the frustrated gasoline buyers just hold money, there is an excess demand for money, and lower expenditures on output. But, let's leave that out by assumption. And so, the frustrated buyers just purchase other consumer goods. The amount of gasoline purchased will be the quantity supplied. Now, suppose one of the goods people choose to buy instead of gasoline, is leisure. They work less, and output and income are lower. Real aggregate demand is lower. People were working to get gasoline, and since they can't get gasoline, they don't work as much. This is part of the disruption caused by the price ceiling. Is this a paradox of gasoline demand?

Now, back to the price floor on interest rates. Now the supply of saving increases. And so the price floor is more disruptive. Given the price ceiling, the amount actually saved equals the amount invested, which isn't changing. If some of the additional frustrated savers choose leisure rather than some other good, then there is less labor input, output, income, and real aggregate demand. The amount saved is actually unchanged (equal to the given amount invested.)

But the problem isn't a decrease in aggregate demand. The problem is that people don't want to work as much. They choose less current consumption, but if they can't get future consumption for it, they would rather have leisure. This preference ordering is possible.

One the margin, people want more future goods most, then more leisure, and then current consumption. If markets clear, then the price change could result in people working less, and the amount saved and invested rise. If the markets don't clear because of the price floor, the amount saved can't rise, and if people work less, then output falls.

I don't think this counts as the paradox of thrift. And I think this is what all that back scratching and bonds was all about.

I think that Adam P's approach is unable to distinguish between monetary disequilibrium and price floors or ceilings. I, at least, don't deny that market distortions like price controls will reduce real output and so real aggregate demand, even if there is no monetary disequilibrium. I think Hutt focuses on this sort of thing.

But I don't think a central bank creating or destroying money to manipulate credit market conditions to target interest rates is creating a price ceiling or floor. Nor do I think that the sort of changes in the willingness to work because of an inability to buy future consumption goods explains existing recessions.


Kevin: exactly. And that Barro Grossman point B (I can even remember it's called "B", without looking) shows that if there were no problem with barter we would go right to full employment point A. Even the insecticide workers could be paid in insecticide, and barter some of it for beer, with the brewery workers, and so on.

The problem with Keynes' point D, apart from the empirical problem, is that it's conceptually unclear whether the underlying problem is W/P being too high, or M/P being too low. In his verbal discussion, it seems to me that it's the latter interpretation he has in mind. But you can't really tell from his model.

BTW, in the second edition of Patinkin, IIRC, he did in fact conceptually discover the constrained demand curve for labour. He has priority over B&G. But Patinkin missed the constrained consumption demand curve (aka the Keynesian consumption function). B&G put them both together.

Bill: I agree with the whole thing. (But I have to keep reminding myself of your first line, that we are assuming it never spills over into monetary disequilibrium!) Who is Hutt? The name sounds vaguely familiar.

Nick: thanks, I’m glad that I now seem to be clearer as to what your concern is. I’m not persuaded that it’s really best expressed as a matter of money being a medium of exchange. The barter fix amounts to setting up a model of a well-functioning barter market inside a model of a dysfunctional monetary economy. But if the solution was that simple the bungling auctioneer, who causes the trouble in the Barro-Grossman story, would hardly have bungled in the first place. His problem isn’t so much that money thwarts him; it’s more to do with finding the right price for the stock of assets relative to the flow of goods and services. Really, what’s the correct ratio between the hourly wage of a semi-skilled worker and the price of an oil refinery, or the market capitalisation of IBM? In a Walrasian model the auctioneer gets that price right before anybody is contractually committed. It’s a lot to ask of the invisible hand.

“The problem with Keynes' point D, apart from the empirical problem, is that it's conceptually unclear whether the underlying problem is W/P being too high, or M/P being too low. In his verbal discussion, it seems to me that it's the latter interpretation he has in mind. But you can't really tell from his model.”

That’s a feature, not a bug. There are two exogenous nominal variables, M and W, and what matters for demand is the ratio between them, the money supply measured in wage units, M/W. (Since goods are heterogeneous in Keynes’s model while labour isn’t, he rarely brings P into it.) So formally, wage-cuts come to precisely the same thing as an expansion of the money supply, but for Keynes the formal model is just the starting-point for the discussion of policy options:

If, indeed, labour were always in a position to take action (and were to do so), whenever there was less than full employment, to reduce its money demands by concerted action to whatever point was required to make money so abundant relatively to the wage-unit that the rate of interest would fall to a level compatible with full employment, we should, in effect, have monetary management by the Trade Unions, aimed at full employment, instead of by the banking system.
Nevertheless while a flexible wage policy and a flexible money policy come, analytically, to the same thing, inasmuch as they are alternative means of changing the quantity of money in terms of wage-units, in other respects there is, of course, a world of difference between them. Let me briefly recall to the reader's mind the four outstanding considerations.

My paraphrase of the four reasons which follow: (1) we don’t live in a dictatorship, where W can be cut by decree; (2) many non-wage contracts will be in money too, and justice would require changing them also; (3) cutting W crushes debtors and (4) the prospect of deflation will deter investment.

Nick:

"3. I spend $1,000 on antique furniture. $1,000 thrift. Zero hoarding."

You and your counterparty exchanged assets: cash for *old* furniture. Assuming the counterparty hoards, (3) is equivalent to (2) -- no consumption/investment happened, GDP is unchanged.

Substitute existing gold for antique furniture with the same result.

Kevin: basically agreed, but:

If barter *were* frictionless, there would be a way around a disfunctional monetary system, but then we wouldn't have a monetary system in the first place. I introduce barter purely hypothetically just to show there must be some frictions, and it's those frictions that create a monetary exchange economy, and also show why monetary exchange matters. (You probably don't disagree on this).

Keynes: if Keynes had held P exogenous, instead of W, or in addition to W, he could also have shown that whatever P cuts could do, M increases could do just as well, or better.

(That bit about monetary policy by the Trade Unions always struck me as ironic, in the 1970's, when we had the Trade Unions targeting real wages, and the Bank of England targeting full employment, and the two targets weren't compatible, so we got an inflationary spiral.)

vjk: Sure. But will the counterparty *want* to sell antiques? And will he *want* to hoard money? Actual savings always equals actual investment, by accounting definition. But that's not what's at issue here. If *desired* savings is greater than *desired* investment, what is it that adjusts to bring them into equilibrium?

Apples bought always equals apples sold. But what adjusts, if quantity demanded is not equal to quantity supplied? The accounting conventions won't answer that question. They weren't designed to.

Adam P said: "Now suppose that every day 5% decide they'd like to save (a different 5% each day), they will offer a backscratch today but won't consume one. Instead they'll take a bond. First problem, does a non-saver, who expects to be a non-saver tomorrow accept the deal? Yes, tomorrow he pays back the backscratch and consumes by issuing his own debt to one of the savers. Today total output of backscratches is 95% of max.

Tomorrow arrives and yesterday's savers consume 2 backscratches and provide 1. Those who issued debt provide a backscratch to redeem the maturing bond and issue debt to the current savers to consume.

It's easy to see that we know have an equilibrium in which output is 95% of the max forever. Yet there is no hoarding of bonds and bonds are not the medium of exchange. Most transactions are bilateral exchange of service.

So, a 5% increase in aggregate saving decreases income by 5%. That's not a paradox of thrift?"

And, "PS: and of course *aggregate* saving remained zero. In aggregate the reduction in consumption from 100 to 95 was not offset by aggregate consumption of 105 tomorrow. Aggregate saving was still zero, some individuals manage to save some periods but the net result is just a fall in income."

What happens if you run this scenario with say 1,000 people and the SAME 2 people continually save?

Adam P said: "But then this puts the US in paradox of thrift land, since (X-M) will always be negative you have that Y < C + I + G, so either the private sector dissaves or the public sector dissaves (or both) and that's true no matter what the level of Y!"

current account deficit = gov't deficit plus private sector deficit?

Dissaving with currency denominated debt is the big problem with running a current account deficit even if it lowers price inflation. I would also say that selling financial assets or "dissaving" with currency can be a problem too.

"Keynesian macroeconomics makes no sense whatsoever in a barter economy. Unemployed workers want jobs, so they can buy goods? Firms won't hire them, because they can't sell goods? What's the problem? Why can't the firms just pay the workers in goods?"

This is just an artifact of your timeless production model.

Suppose you have farms that, in period 1 employ planters and in period 2 employ harvesters. You can pay harvesters with food, but you cannot pay planters with food, because they do not create food, they create the capacity to sell food in the next period.

The general gluts are not caused by lack of barter -- they became widespread and recurring with the industrial era when large amounts of long term capital investment were necessary, and these investments were funded by selling debt. Also a period when you saw the rapid growth of credit markets to fund that investment.

Keynes highlighted the volatility of long term investment and the inability of money market interest rates to ensure that a sufficient quantity of investment occurs to maintain aggregate demand.

All of that is "real", and this is close to observed behavior. I think the effort should try to be to create models in which the observed behavior is explained, rather than promote models that trivialize interest rates and investment and instead argue that people are hoarding the medium of exchange -- something that is not observed.

There is a lot of data as well as logic pointing out that investment demand is relatively inelastic to the overnight interest rate. The cost of capital includes an earnings growth component as well as a dividend component, and both are long term. Now perhaps, if development of a factory could be achieved overnight, then overnight rates might be more effective, and the CB control of these rates might be enough to maintain demand.

But look at what is happening -- the pension fund industry standard is still for a return of 8%. The economy is growing at 6%. The institutional investors have not lowered their return demands in response to central bank rate targeting. What about businesses? Their cost of capital has not fallen either. More than a year of zero interbank rates and still the non-financial business community does not face a materially lower cost of capital. Now if you combine the same cost of capital that market the secular earnings boom with reduced investment demand, you have a recipe for idle resources.

As an aside, Adobe's stock plunged recently after their earnings call. They met earnings, but they lowered their earnings growth rate guidance. They are laying off about 15% of their workforce now -- and this is the third such round of layoffs since the recession began, even though they consistently met their earnings targets throughout. When they lay off, they are not firing employees who are too expensive, but closing down business units and projects that fail to meet their 15% future profit growth needs. As collateral damage, labor is cut when the capital is liquidated.

Now if enough businesses face a cost of capital that is too high -- remember the 8% pension fund assumption -- then they will all individually cut investment in an attempt to climb back up their MPK curve, but the paradox of thrift comes in, so that in aggregate they will fail to grow their earnings at the required rate. All of this is "real" -- e.g. the earnings can be in terms of goods. Spot prices can be perfectly flexible, as long as the earnings growth requirements are sticky. None of this reflects a desire to hoard the medium of exchange.

RSJ: "the pension fund industry standard is still for a return of 8%"

This doesn't tell me that the cost of capital is high; this tells me that there's a pension fund crisis brewing.

Andy, it's been 8% for a very long time now.

We have had a secular earnings boom, in which the corporate profit share of GDP was steadily climbing, so that 8% was justified during that boom period. That 8% figure, or it's equivalent, has become embedded in firm management, investment advisors, the institutional guys -- basically everyone. Unfortunately, try telling people that this is unsustainable -- it's been working since 1980, after all -- for all of their professional lives.

No individual firm is going to be willing to accept lower growth rates going forward because the last 30 years were a debt driven fluke. They all think that they are special and need to grow at high rates, and when those rates don't materialize, they interpret that as a signal to cut investment in the existence projects and find something else that will deliver earnings growth. I've know many profitable projects that were cancelled because the growth rates weren't high enough.

Curiously enough, none of the non-financial actors look at the CB policy rate as guidance to their cost of capital. Why don't the economic models also make this distinction?

You make negative remarks about fiscal stimulus, writing: "Fiscal deficits, if they work, will have future costs. The higher debt will mean very difficult future spending cuts or tax increases." Which will it be: spending cuts or tax increases? That matters to a voter who (like me) prefers smaller government. He would like, let us say, a Federal government that spends only 13% of GDP. But the Federal government will tend to bump up against the upper limit of what is tolerable to the voters, which (let us say) is 26% of GDP; so our small-government voter is destined not to get his first preference. Turning to second-best, he would probably prefer a government that spent 13% on actual programs and 13% on interest on the national debt to a government that, because it had no debt, spent all 26% on programs. So he would be happy to see the government burdened with debt; that would constrict its activities by its need to pay interest. Debt cripples the government, which is what the small-government voter wants.

So long as future tax increases are perceived, in political terms, as even more "difficult" than future spending cuts--so long as there really is a practical upper limit on what taxpayers are willing to bear--increases in the debt are not "costly" at all (from the pro-small-government point of view). Fiscal stimulus: bring it on!

"A desire to buy antique furniture is a form of saving, because antique furniture is not newly-produced." This implies a truly bizarre definition of 'saving'. First, how new is "newly"? One minute ago? A decade ago? Something in between (but where, exactly)?

A piece of furniture is a "consumer durable," which means that its acquisition is partly consumption and partly saving/investment. It provides current services--for example, one can sit on it or gaze at it--and it also provides future services of the same sort, besides serving as a store of value through its potential to be resold. Acquiring and holding/using an antique chair is not pure saving; it is partly consumption. And the same can be said for acquiring and holding/using a newly produced chair; there is no difference with regard to consumption/saving.

Increased hoarding of money will cause a recession only if the fiat-money authority does not sufficiently increase the supply of money to accommodate the increased demand (i.e., the reduced velocity). If one views *keeping the money supply constant* as *doing nothing* one will consider that the only basic change in the economy has been the increased desire to hold money, so this must be the cause of the recession. But suppose there were in place an automatic mechanism for adjusting the supply of money to the demand. Then the authority would have to *do something* to thwart this mechanism in order to keep down the quantity of money when the demand to hold it increased. Then there would seem to be *two* causal factors in producing the recession: the increased desire to hold money *and the action by the monetary authority preventing the quantity of money from rising*. We would not want to ascribe the recession simply to increased hoarding, ignoring the monetary authority's (perverse) action. Indeed, if we viewed fluctuations in the demand to hold money as *normal*, we might even want to offer the monetary authority's extraordinary action as *the* cause of the recession.

Philo: My guess is that current fiscal deficits will imply *both* future spending cuts *and* future tax increases. Both are politically unpopular, and undesirable, so a rational government will equalise the marginal costs on both margins. And unless the marginal cost curve is perfectly elastic or inelastic on one of those two margins (which seems implausible), they will move up the MC curve on both.

You can measure the "size" of government as government spending/GDP, or government taxes/GDP. (Both measures are very flawed, but that's just an aside.) By the first measure, government will get smaller, by the second measure it will get larger. People who like "small government" are normally objecting to the latter. (As a thought-experiment, suppose someone else voluntarily gave our government aid so it could spend more without increasing taxes; how many people would dislike this increase in the size of government?)

Everybody wanting to buy an antique chair is *neither* desired consumption *nor* desired investment. Only newly-produced goods count, because it's the demand for newly-produced goods and services that creates the demand for the labour and other resources needed to produce them. We can argue over whether a demand for newly-produced chairs is desired consumption or investment (I say it's investment), but which one it is doesn't affect the standard paradox of thrift argument, which recognises that an increase in desired savings matched by an equal increase in desired investment will not cause a fall in desired aggregate expenditure or a recession.

I basically agree with your last comment, about the response of the money supply. (I was assuming it fixed, for simplicity).

Contra Brad DeLong and like-minded proponents of fiscal stimulus, you write: "The 'exit strategy' to a fiscal solution looks very ugly"--a remark that seems designed to inject some political realism into the discussion. But this may be unfair to DeLong. He is not much interested in whether gracefully exiting from running large deficits would be politically possible--or even whether running the deficits to begin with would be. His contention (as I interpret him) is that under certain political conditions--roughly, what would prevail *if most people had economic insights comparable to DeLong's*--large deficits would work to end the recession. And under those conditions, a graceful (by DeLong's lights) exit from deficit spending would also be possible--indeed, it would be guaranteed!

Policy recommendations almost always take this form: the proponent is saying that *if only most people agreed with his thinking on the subject*, such-and-such policy would work well. That usually makes them rather far removed from being "practical"; considerations of *realism* are inappropriate.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad