I just finished reading John Cochrane's paper, where he applies the fiscal theory of the price level to the current crisis. (Thanks Adam P!). OK, I admit I didn't really read it thoroughly; I skimmed it. It's full of good insights, and also bad mistakes. The best thing to do with a paper like that is to skim it for good ideas.
It would be easy to trash this paper for its mistakes, but pointless. Instead I'm going to mine it for what I see as the gold of its insights, and leave behind what I see as its mistakes.
Is this a good way to analyse what John Cochrane's paper is really saying? Of course not. But I don't really care about that. I just want the insights, and whether they are accurately portrayed is a side-issue. In fact, I might as well say now that they are not accurately portrayed. That's why I put the scare quotes around "John Cochrane's" in the title of this post.
OK, I admit, it's not really John Cochrane's argument for fiscal stimulus; it's an argument for fiscal stimulus that I got from reading John Cochrane's paper.
The problem for the last few months has been a very large increase in the demand for short-term government debt, where by "debt" I mean both base money and bonds. Interest rates on short-term government bonds have fallen to zero or near zero. Interest rates on longer term government bonds have fallen, but not to zero. Interest rates on commercial bonds (and yield rates on stocks) have increased. The prices of real assets, like houses, farmland, capital goods, gold, oil, metals, has either fallen or stayed the same. The demand for newly-produced goods has fallen, as shown by falling prices (or inflation rates) and output.
It certainly does not follow from any accounting identity that the fall in demand for newly-produced goods must have been caused by an increase in the demand for government debt. But this is not an unreasonable interpretation of the facts. Or, to be more precise, there has been a shift in demand away from newly-produced goods (which is what "aggregate demand" normally means) towards government debt. And people don't much care, at the margin, whether that debt is base money or short-term bonds.
So, the assumption is that at the previous level of prices, inflation rate, interest rates and output, there was an excess demand of government debt and an excess supply of newly-produced goods. How does this disequilibrium get resolved?
Nominal interest rates on short-term government debt fell to zero. That helped reduce the excess demand, but it wasn't enough. Falling expected inflation moreover increases the real yield on government debt (which is nearly all nominal debt) and offsets the benefits of lower nominal rates. Falling real output will eliminate the excess demand for debt, but we don't want that to happen. A fall in the price level will eventually increase the real supply of government debt, but would take too long, and we don't really want that to happen either.
There are four policies the government might take to eliminate the excess demand for short-term government debt.
First, a truly massive version of "operation twist" might help. Buy long-term government debt and sell short-term government debt. This would probably help some, but there also seems to have been an increased demand for long-term debt, as evidenced by falling long interest rates, so it would probably not help enough, if the overall demand for government debt, both short and long, increased. Plus, it might increase future "rollover risk" unacceptably.
Second, a truly massive government purchase of private financial and real assets might help. TARP, on a very large scale. If the private sector wants to sell private assets and buy government debt, then the government should do the opposite to restore equilibrium. I'm going to leave that one there.
Third, a truly massive current fiscal deficit, offset by equal future surpluses. It would have to be truly massive, because the excess demand for debt is an excess demand for a stock, and the new supply of government debt to finance a current deficit is a flow, and it takes a long time for an increased flow of water to fill a bathtub, unless the flow is very large. This would cause serious distortions in the time-path of government spending and tax rates. Lots of government spending on stuff we don't need now, and too little government spending on stuff we will need in future. Big cash transfers now, and high distorting tax rates in future.
Fourth (and this is where John Cochrane's fiscal theory of the price level comes in), instead of increasing the current supply of government debt, do something to reduce the current demand for government debt, by changing expected future fiscal policy.
The fiscal theory of the price level says that (M+B)/P = Present Value of real primary budget surpluses, and so given existing values for nominal stocks of base money M and nominal bonds B, the price level P is determined by the present value of government primary surpluses.
You get the fiscal theory of the price level if you consolidate the government and central bank budget constraints, forget that M pays no nominal interest, while most bonds normally do, forget the fact that money is a stable ponzi scheme and so the terminal value probably does not converge to zero in the limit, and assume that government surpluses are exogenous with respect to the existing debt but that deficits can nevertheless always be financed.
Yes, there are a lot of problems with the fiscal theory of the price level. It's what you would expect to get if a modern finance theorist made the mistake of reading Knapp's "The State Theory of Money". Money, even fiduciary money, is not always a creature of the state. And having a de jure or de facto monopoly on printing money is a profitable business, so the the surpluses "backing" the value of money are negative, yet money does not have a negative value. (You can resolve this paradox mathematically by recognising that the real interest rate on money is usually negative, so the present value of a permanent stream of negative surpluses, discounted at a negative interest rate, becomes positive once again). And the fiscal theory of the price level implies that an open market sale of bonds, which kept (M+B) constant, would leave the price level unchanged, even if M=0 (thanks Bill Woolsey).
Despite those problems (and they are not trivial) there is a kernel of truth in the fiscal theory of the price level. Future fiscal policy affects the current fundamental value of the government's liabilities, just as the future profitability of a firm affects the fundamental value of its liabilities. (The problem with the fiscal theory of the price level is that it wants this metaphor to run on four legs, while it only limps on two.)
I can identify five channels whereby expected future fiscal policy might affect the current demand for government debt (base money plus bonds).
First, if future deficits are expected to be bigger, there is a bigger chance that the government will seek increased seigniorage from the central bank to help finance the deficits. This will increase the expected future price level, and so increase current expected inflation, and so reduce the current real yield on government debt, and so reduce demand for government debt.
Second, if future deficits are expected to be larger, and the national debt bigger, there is a bigger chance that the government will want a higher price level in future to reduce the real value of its future debt obligations. This too will increase the expected future price level, and so increase current expected inflation, and so reduce the current real yield on government debt, and so reduce demand for government debt.
Third, if future deficits are expected to be larger, and the national debt bigger, there is a bigger chance that the government will default on its debt. Default could be a deliberate refusal to pay; it could be a refusal to pay caused by political gridlock preventing tax increases or spending cuts; it could be an inability to rollover maturing debt; or it could be an inability to pay except by printing money. This would increase the current risk premium on government debt, and so reduce demand for government debt.
Fourth, if future deficits are expected to be larger, and deficits increase future aggregate demand and the price level, that will increased expected inflation today, and so on..
Fifth, if future deficits are expected to be larger, and deficits increase future aggregate demand and real income (as they might if prices are sticky and we are talking about the near future when the economy might otherwise still be in recession), then that increase in expected future real income might reduce the current demand for government debt. The reason this would happen is that people would want to save less today, firms would want to invest more today with higher expected future sales, and private debt would be safer if people and firms had higher future income. All this would reduce the demand for government debt directly, and indirectly by increasing the supply and quality of private debt.
OK, you might say this is all starting to sound rather Keynesian. And it is in a way. But an important aspect of the Keynesian argument for fiscal stimulus is that we need it quick, because it is current fiscal policy that affects current aggregate demand. All my arguments above are about how future fiscal policy might work today. And, how it might work by reducing the current demand for government debt (base money plus bonds), and thus indirectly affecting current demand not just for newly-produced goods and services (aggregate demand), but also for private debt and other assets. It's not the same. Current fiscal policy matters only insofar as it affects expected future fiscal policy, in these arguments. Or, it's the whole stream of current and expected future fiscal policy that matters today.
Is it a good argument for fiscal policy? That depends. Would it work? Would the cure be worse than the disease? Would monetary policy be better? I have only looked at the first of those three questions.
I myself would draw only one conclusion: if we see an economic recovery begin, even before the fiscal stimulus has really begun, it might be the effect of expected future fiscal policy. Or it might be monetary policy, financial policy, or the economy just recovered by itself.
"It would be easy to trash this paper for its mistakes, but pointless. Instead I'm going to mine it for what I see as the gold of its insights..."
Boy, is this what happens when Canadians blog? Reasonableness on the interwebs? Doncha know you're supposed to flame, trash and snark?
Posted by: loomis | May 26, 2009 at 11:11 AM
“It would have to be truly massive, because the excess demand for debt is an excess demand for a stock, and the new supply of government debt to finance a current deficit is a flow, and it takes a long time for an increased flow of water to fill a bathtub, unless the flow is very large.”
Not really. The difference between “normal demand” and “excess demand” for a stock is conceptually equivalent to a type of flow. There’s no reason to think potential government supply can't match this flow.
Posted by: anon | May 26, 2009 at 11:37 AM
O loomis...doncha know that was not Nick The Venomous but "Nick" who has decided to dismiss John, worthy colleague and nearly flawless, for "John", isolated episode of former John who needs to vet more and drink less if he is to continue with NR thinking that 4 insights and 5 mistakes is a pretty good score.
Alright...I did not keep score, but is John, conduit of good and bad ideas, likely to be happy with the "John" treatment? If we do not clear the hurdle that 6' 10" Nick steps over so nimbly (Nick, 139 lbs...not Sumo Nick at 439 lbs) are we (possibly just bad conduits) likely to be not only offended, but further retarded in our pursuit of those good ideas?
Posted by: calmo | May 26, 2009 at 12:05 PM
what?
Posted by: loomis | May 26, 2009 at 12:33 PM
Nick,
The one of the whole points of the fiscal theory is that it pins down the price level with M=0!
Posted by: Adam P | May 26, 2009 at 12:35 PM
loomis: Reminds me of when I was new in Canada, many years ago, as a scruffy young grad student, I drove my old car into MacMaster University and asked the parking attendant, an older gentleman, where I could find the economics department. "Excuse me pointing, sir" he said, "it's over there."! Country's gone downhill since then.
anon: I would interpret what happened as an increased desire to save in the form of government debt, and a desire to switch portfolio towards government debt. The first is a flow excess demand. The second, if it happened instantly, would be a stock excess demand. I expect it didn't happen instantly, but it was fairly fast. So it would be a very large flow, if we modeled it as a flow.
calmo: welcome back! I think I almost understand your comment (it takes a lot of practice, loomis). Not sure how John Cochrane would react. I hope he would at least be happier than he is at the response his writings often engender. And understand an attempt to learn from him what I can. I think Paul Krugman recently wrote a post on reading Minsky's book. He said that apart from the one good idea, the rest of the book was really bad. But who remembers the bad stuff? Good ideas are very scarce. One good idea makes up for 10 mistakes, as long as there are people around who can tell the difference. Universities are supposed to be places for new ideas, even if most of them are hopelessly wrong. What's the saying? Even a mistake is useful, if it's clear, so we can learn from it.
Adam: So the fiscal theory of the price level was designed to apply in the limit as we approach a cashless, reserveless, economy? I didn't know that. It makes more sense of it.
My own (off the cuff) theory of the price level with competitive monies would be:
P = i - r + l + E[P(+1)] where i is the own rate of interest on money, r is the prevailing rate of interest in the economy, and l is a measure of the ease of use/liquidity premium on this particular type of money. I think I would divorce the theory of the value of money from the value of bonds.
Posted by: Nick Rowe | May 26, 2009 at 04:24 PM
For P read 1/P (I think).
Posted by: Nick Rowe | May 26, 2009 at 04:26 PM
Nick,
The fiscal theory applies with or without money, one of the advantage is that it applies in a cashless society while the quantity theory doesn't.
A couple more points related to what was said on Scott's blog:
You said that a government that never had any surpluses would have an infinite price level. Well, first of all you don't need actual surpluses so much as you need the ability to generate them if you need to. It's just like maintaing a fractional reserve gold standard, the notes are valued even if the government/central bank don't have enough gold so long as it is believed they have the resources to get more.
Secondly, it's pretty clear that a government that had absolutely no possiblity of ever generating a surplus could not in fact issue debt with any value, including currency. That's Germany in the 20s or Zimbabwe now. It would be like trying to maintain a gold standard without any gold or the possibility of getting gold.
Posted by: Adam P | May 26, 2009 at 05:07 PM
Nick, I think Adam is right about the surpluses, although I am not certain. It may not seem like deficits and surpluses balance in the long run, but the budget deficits we read about in the newspaper are nominal, and are often much larger than real deficits. Indeed in real terms the US runs surpluses whenever the national debt rises by less than the rate of inflation, which is surprisingly often.
Adam, You need a medium of account to have a price level. It seems to me that the quantity theory of money could be applied to any medium of account, whether it was a medium of exchange or not. I realize that it wouldn't be "money" in the sense most people describe the term, but the intuition behind the QTM would be the same. Unless I am mistaken, Hume's QTM applied to the medium of account (gold), not the media of exchange, which were not typically gold.)
Nick, Future fiscal deficits in the US have recently soared by trillions. Inflation expectations barely budged. I hope we don't try to double down as I am much more worried about future deficits and taxes than I am confident that future fiscal stimulus can boost AD. If it were to "work" by creating expectations that the US was becoming a banana republic where the Fed was the handmaiden of Congress, then I would regard it as a pyrrhic victory. (I understand you aren't endorsing it, but I thought I'd put my 2 cents in.)
Posted by: Scott Sumner | May 26, 2009 at 06:33 PM
Amusing to see the switch from "best arguments why fiscal stimulus doesn't work" to "worst arguments why fiscal stimulus does work" (I'm paraphrasing).
Interesting that this argument relies on decreasing demand for government debt rather than directly increasing demand for newly produced goods.
What do you think is the magnitude of the five effects that you list, compared to the direct effect on AD of government spending and of consumption from tax rebates?
Acknowledging the argument that fiscal stimulus may not really have kicked in yet, what do you think is the effect on AD of an expected future increase in government spending/consumption? My hypothesis would be that this increases private investment, in the expectation of higher demand in the near future. But again I don't know whether that would be a weak effect or a significant one.
Posted by: Leigh Caldwell | May 26, 2009 at 07:07 PM
Leigh
Surely there is some regional empirical evidence out there. Lets say the DOD decides to build a new base (or a state government decides to build a new jail) in some underdeveloped location - what is the announcement effect before they actually start employing people there - probably the initial effect is that banks become more friendly to local businesses.
Scott
Re effectiveness of expectations of future policy - one of the problems is here is that you really need to know two things here:
1. what the expectations of future policy are
2. how effective people think that policy will be.
Most people think that policy changes in response to experience, and that its effectiveness is variable.
Posted by: reason | May 27, 2009 at 08:32 AM
Adam and Scott:
Some definitions on surpluses:
The surplus that gets reported in the newspapers is:
"headline surplus" = T - G - iB where iB is the nominal interest on debt.
The headline surplus tells us the change in the nominal debt.
If we adjust this for inflation, both to convert G,T and B to constant dollars, and to recognise that part of the nominal interest on debt is just an inflation adjustment, we get:
"real surplus" = t - g -rb (where t = T/P etc.)
The real surplus tells us the change in the real debt.
John Cochrane expresses the Fiscal theory of the price level in terms of the primary surplus.
The "nominal primary surplus" = T-G
The "real primary surplus" = t-g
Forget money, and lets just think of it as the long run government budget constraint.
You can run headline deficits forever, provided nominal GDP is growing, and you at least pay the difference between i and the growth of nominal GDP on the debt.
But you cannot run a primary deficit forever, because it means you are not even paying the interest on the debt. You are running a Ponzi scheme. The debt will grow faster than nominal GDP. (Unless i is less than growth rate of nominal GDP, in which case you can run a Ponzi Scheme forever.
So ignoring M, the long-run government budget constraint is:
B = PV( primary surpluses), which is exactly what JC has, and is correct, provided i exceeds growth rate of nominal GDP, to rule out Ponzis.
So you MUST have a primary surplus sometime.
Now lets bring M back, and assume B and T is always zero. It is perfectly possible for a government with a printing press to finance some spending.
But the fiscal theory of the price level in this case says:
M/P = PV( real surpluses) = PV(-g) which looks negative.
So you get an apparent contradiction. The fiscal theory of the price level says you cannot finance any government spending just by printing money. But we know you can.
You resolve the contradiction by noting that if money pays no interest, the real rate of interest on money can be negative (so the PV formula is perverse), or else the real interest on money can be less than the growth rate of real GDP, so ponzi schemes are stable.
Posted by: Nick Rowe | May 27, 2009 at 10:10 AM
Scott:
"Nick, Future fiscal deficits in the US have recently soared by trillions. Inflation expectations barely budged. I hope we don't try to double down as I am much more worried about future deficits and taxes than I am confident that future fiscal stimulus can boost AD. If it were to "work" by creating expectations that the US was becoming a banana republic where the Fed was the handmaiden of Congress, then I would regard it as a pyrrhic victory. (I understand you aren't endorsing it, but I thought I'd put my 2 cents in.)"
I tend to agree. We might look at it this way. We have a choice between "trashing" the balance sheet of the central bank, so people will expect future money printing, or trashing the balance sheet of the government, so people will expect it to default or elee resort to money printing.
If we have to trash a balance sheet to get expectations to change, I would rather trash the central bank's balance sheet.
Posted by: Nick Rowe | May 27, 2009 at 10:15 AM
Leigh: "What do you think is the magnitude of the five effects that you list, compared to the direct effect on AD of government spending and of consumption from tax rebates?"
I don't know. I think it depends on the current debt/GDP ratio, and how fiscally responsible the government is seen to be. I would hope that the main effect is the fifth, or fourth. The other first to third really are banana republic territory.
Posted by: Nick Rowe | May 27, 2009 at 10:19 AM
Nick, I haven't looked closely at this, but is there anyway to solve the puzzle you mention by simply redefining "money creation" as taxes? Thus if money creation is an inflation tax that extracts wealth from current holders of money, who see its value fall, then the government is not getting something for nothing, and should be able to finance a modest amount of spending forever. Does the fiscal theory ignor the very real transactions convenience of media of exchange? Is that why it can't explain how the government could tax this "good" through persistent money creation?
I agree with your point about trashing balance sheets, but would add that we really don't even need to do that if we can turn inflation expectations around (and stop paying interest on excess reserves.)
Posted by: Scott Sumner | May 27, 2009 at 05:59 PM
Scott:
Yes, if you include the revenue from printing money as a form of tax, make sure the interest rate in the PV calculation is the interest rate on bonds, take M off the left hand side (assuming M pays no interest), then you get a perfectly valid equation: the standard long run government budget constraint.
The way I look at it is this: there are two entities, the narrow government, and the central bank.
If we assume that the real rate of interest is greater than the growth rate of real GDP, the narrow government cannot run a Ponzi scheme with the national debt, so you get a well-defined long run government budget constraint where real value of the bonds equals the present value of the real primary surpluses, plus any profit transferred from the central bank.
But if we assume, reasonably, that the nominal interest paid on base money is less than the growth of nominal GDP (or the real interest is less than the growth rate of real GDP), then the central bank can run a stable Ponzi scheme, so when you try to get a long run budget constraint, the present value doesn't converge, unless you have a rule for spending some of Mr Ponzi's profits.
The fiscal theory of the price level tries to consolidate those two balance sheets, using only one rate of interest in the present value calculation, and ignoring the mix of "liabilities" between money and bonds. The result is a total mess, because the mix of liabilities matter, because they pay different rates of interest.
The root of the problem is treating base money as a liability of the government. Base money is irredeemable.
Now, you can draw the analogy between a consolidated government plus central bank, and a corporation that issues both shares (money) and bonds. But nobody will hold shares unless the expected rate of return is at least as good as the rate of interest on bonds. People will hold base money even when the rate of interest is less than the rate of interest on bonds. Why? Duh! Because money is a medium of exchange, and bonds aren't.
Again, Money is not a liability of anyone if the real rate of interest on money is less than the growth rate of real demand for money. It's a stable Ponzi scheme, that works even if everyone knows there's no assets "backing" the money.
Mr Ponzi had to make people think there were assets backing his liabilities. Only because people would not hold his liabilities except at a rate of interest greater than long run growth in real demand for his liabilities. They didn't get any benefit from holding Mr Ponzi's liabilities, except for the interest. They do with money. Money is more liquid than bonds. If Mr Ponzi could have made his liabilities more liquid than bonds, he could have kept in business forever (provided he avoided a run). He would just call himself a bank.
Posted by: Nick Rowe | May 27, 2009 at 07:42 PM
Nick, Thanks, your comments make a lot of sense. It makes me feel better about not spending a lot of time trying to figure out the fiscal view. By the way, here's a little anomaly. In the US most of the demand for cash may be associated with tax evasion. If they eliminated taxes and the government tried to run on seignorage alone, they might collect far less in seignorage. Alternatively, the existence of an anonymous fiat money may lower income tax revenues by more than they gain in seignorage from money.
Posted by: Scott Sumner | May 27, 2009 at 10:32 PM
Nick: "Base money is irredeemable."
Surely base money - as part of the definition of legal tender - can be "redeemed" by paying off your tax liabilities? Sure, you don't get to walk away with a bar of gold or can of beans, but you walk away with reduced tax liability, and a smaller liability is just as good as a bigger asset.
On this theory, the net present value of money would be the discounted total not of the government's surpluses but its future DEFICITS, because that's presumably the amount of tax it will eventually have to raise.
I'm not quite serious. But I like the idea.
Posted by: Leigh Caldwell | May 28, 2009 at 02:11 AM
"Surely base money - as part of the definition of legal tender - can be "redeemed" by paying off your tax liabilities? Sure, you don't get to walk away with a bar of gold or can of beans, but you walk away with reduced tax liability, and a smaller liability is just as good as a bigger asset."
Finally someone gets it. Exactly right Leigh.
However, it is future surpluses, not deficits. The surpluses are the real amount of taxes actually collected, the FUTURE government surpuses are the private sector's future tax liabilities. They are related to current and past deficits.
Posted by: Adam P | May 28, 2009 at 06:51 AM
I use Canadian dollars, the medium of exchange, to pay my federal taxes, my provincial taxes, my municipal taxes, my mortgage, my credit card liability, my daughter's rent bill, my Quebec Hydro bill, my phone bill, my gas bill, my bread and milk bill, etc. I can use money to reduce my liabilities to all of these agencies.
Why should we single out just the tax liabilities I pay off to redeem money as determining the value of my money? And why just the Federal tax liabilities at that? Because the Federal government is the only one of those entities that can print money?
Suppose the Federal government of Canada disappeared as a fiscal authority, stopped collecting taxes, stopped spending money, so had zero surpluses, but just left the Bank of Canada remaining as the only Federal entity. All other functions of government devolved to the provincial governments. Would the Canadian dollar disappear too?
No, and I have empirical evidence to say it wouldn't.
Canada would become like the Eurozone, which also lacks a central fiscal authority that can raise taxes.
Posted by: Nick Rowe | May 28, 2009 at 08:14 AM
"Now, you can draw the analogy between a consolidated government plus central bank, and a corporation that issues both shares (money) and bonds."
Wrong analogy.
Money is at the top of the capital structure; not the bottom.
Posted by: anon | May 28, 2009 at 08:31 AM
"Why should we single out just the tax liabilities I pay off to redeem money as determining the value of my money? And why just the Federal tax liabilities at that? Because the Federal government is the only one of those entities that can print money?"
Because the government (central bank) is the only one that extinguishes the money liability when it is used to pay taxes.
Posted by: anon | May 28, 2009 at 08:34 AM
Nick
EUROZONE GOVERNMENTS ACCEPT EUROS AS PAYMENT OF TAXES.
Posted by: Adam P | May 28, 2009 at 08:53 AM
Money is not comparable to equity (at the bottom of the capital structure) because the nominal "market value" of money doesn’t change.
Posted by: anon | May 28, 2009 at 09:07 AM
"Surely base money - as part of the definition of legal tender - can be "redeemed" by paying off your tax liabilities"
'Fiat' currency is 'backed' by the governments ability to tax, and thus in a sense the GDP of the country. So if you wanted to redeem it (supposing you could), you'd be trading a dollar for a dollar worth of GDP, which the gov't would get by taxing, and would pay to you in dollars. So you'd redeem your dollar for a dollar, which would be silly so nobody wants to do it.
I don't know if that's right, but that's the story I tell my libertarian friends when they get on the gold standard paper money isn't backed by anything hobby horse :)
Posted by: Patrick | May 28, 2009 at 09:29 AM
“So you'd redeem your dollar for a dollar, which would be silly so nobody wants to do it.”
It’s neutral in terms of paying taxes and related silliness. Current taxes owed are a cost to the taxpayer no matter what the mode of payment. Nobody wants to do it, not because it's sillier, but because it’s inconvenient as a mode of payment. But that doesn’t mean it can’t be done.
It is a cost to the government. Taxes paid with currency are used to redeem the currency instead of paying for expenditures or servicing interest paying debt. It’s the reverse effect of the benefit to the government of funding via currency in the first place.
Posted by: anon | May 28, 2009 at 09:53 AM
anon @ 8.31:
"Wrong analogy.
Money is at the top of the capital structure; not the bottom."
The analogy between money and shares is one used by proponents of the fiscal theory, not me. So I agree it's wrong, but for different reasons. I don't understand what you mean by saying money is at the top of the capital structure.
anon @8.34:
"Because the government (central bank) is the only one that extinguishes the money liability when it is used to pay taxes."
Now, which is it? The government or the central bank? If you want to build a theory of the price level based on the behaviour of the central bank, its current and expected future seigniorage profits, I am with you. But that's just the modern quantity theory of money, not the fiscal theory. It's the central bank that decides whether or not to extinguish money when it is used to pay taxes. (And currently, I think, the Bank of Canada decides not to extinguish that money, but put put it straight back into circulation, except insofar as the demand for money fluctuates around tax time). Why tack on the government budget constraint onto what would otherwise be a sensible theory of the price level (provided it recognised that the real rate of interest on money is usually negative, and is not exogenous wrt central bank behaviour)?
Adam:
"EUROZONE GOVERNMENTS ACCEPT EUROS AS PAYMENT OF TAXES."
Yep. So do Canadian provinces and municipalities accept Canadian dollars as payment in taxes. Should we add the expected surpluses and debts of provinces and municipalities into the Fiscal theory of the price level? My credit card company also accepts Canadian dollars; should we add BMOs balance sheet in there too? Where do we stop? I would argue we should stop at the balance sheet of the central bank. And the balance sheet of the central bank is a very weird balance sheet, since most of its liabilities pay no nominal interest, and negative real interest.
And whose liabilities are the Yap stones, and cowrie shells? And Yap stones don't even have any real value as commodities.
anon @ 9.07: I would side with the fiscal theorists on that one. The market value of money, in terms of goods, does change. And that's what we need to explain.
Patrick: I would just say that fiat (I prefer "fiduciary") money, has no backing (the assets of the central bank are mostly irrelevant), except the trust that central banks won't print too much, and that other people in future will continue to use it as a medium of exchange, based on the inertia of network externalities. Indeed, what is amazing is how strong those network effects are. Even in Zimbabwe people continued to use Zim dollars, right until the end, even when the real rate of return was incredibly negative, and dominated by almost every other asset. Even fresh milk, which would not be an obvious candidate for a medium of exchange, dominated Zim dollars in rate of return, but they used Zim dollars as a medium of exchange.
anon @ 9.53: sorry, but you lost me there, though some of what you say sounds right, if I interpret you correctly.
Posted by: Nick Rowe | May 28, 2009 at 12:11 PM
Nick, I've already explained this and linked to a paper where Sims explains it. Provinces, municipalities, credit card companies all have nominal budget constraints. Entities that can create money have only real budget constraints, not nominal ones, thus their real budget constraint determines the price level when denominated in the currency they create.
The ECB has a real budget constraint, it is capitalized by the taxing power of the member nations:
http://www.ecb.int/ecb/orga/capital/html/index.en.html
The only thing you've said, here or on Scott's that may be a true counter example is the Yap stones. But do we know anything about their system of government?
Posted by: Adam P | May 28, 2009 at 12:38 PM
Adam: Sorry. I missed the link to Sims' paper. Was it here?
The ECB has a de facto monopoly on producing money in the Eurozone. It chooses how much profit to make (subject to the inflation target upper limit). If it makes a profit, it must be distributed to the shareholder governments in proportion. But would it really make any difference to the price level if it decided to give that profit to someone else? Why can't our analysis just stop at the profit it makes, and ignore the balance sheets of the entities to whom it gives that profit?
Now it is logically possible for a central bank to make a loss and require a bailout, but only if 3 things happen together: if its assets go bad; and if the demand for its money falls; and if it doesn't want to allow inflation. In that case yes, the balance sheet of the entity doing the bailout would matter, to determine if it could afford the bailout. But I can't think of any recent examples (though there must be some). The Bank of Canada generally just hands over a steady $2 billion a year to the government, and it has a net assets of about $50 billion.
My bottom line is this: the fiscal theory of the price level emerges as a limiting case of the quantity theory, when the interest rate on base money approaches the interest rate on bonds (so the central bank makes zero profits), and when the government primary surplus is exogenous with respect to the national debt (interpreted as bonds plus money).
I don't know about Yap government, except for one anecdote, which might actually strengthen your case! When the German colonial authorities had problems getting the Yap islanders to pay taxes, they painted a little black cross on the stones, signifying their "confiscation" by the German authorities. The islanders paid their taxes, and the Germans cleaned off the black crosses!
Posted by: Nick Rowe | May 28, 2009 at 01:56 PM
http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/04/an-alternative-universe-with-gold-price-control.html?cid=6a00d83451688169e201156f58c61b970c#comment-6a00d83451688169e201156f58c61b970c
we already had a lot of the same discussion in the comments surrounding that link.
Posted by: Adam P | May 28, 2009 at 02:12 PM
http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/04/an-alternative-universe-with-gold-price-control.html?cid=6a00d83451688169e201156f58c61b970c#comment-6a00d83451688169e201156f58c61b970c
we had a lot of this discussion already in the comments surrounding that one.
Posted by: Adam P | May 28, 2009 at 02:20 PM
Thanks Adam. Only a month ago, and my memory had totally failed! That's worrying.
I'm going to try to get my brain around that paper. Not sure I'm up to taking on Sims in an argument.
My initial reaction, like to Cochrane's paper, is that he has important insights. To say that the fiscal theory of the price level is a metaphor that limps on 2 legs is not so say it can't move at all.
Posted by: Nick Rowe | May 28, 2009 at 03:36 PM
BTW Nick, here's a quote from David Andalfatto:
"Krugman's proposed "monetary policy" for escaping a liquidity trap is an inflation target regime. What he does not realize is that central banks are not typically accorded the privilege of printing money and injecting it into the economy (this is a fiscal operation).
Central banks are (for the most part) relegated to controlling the ratio of money/bond ratio. Money is to be understood here as non-interest-bearing government debt and a bond as interest-bearing government debt. "
He seems to think governments print money, not central banks.
Posted by: Adam P | May 29, 2009 at 02:24 AM
Adam: here's my take on that question.
I would write the LR government budget constraint as:
B/P = PV(primary surpluses) + PV(profits from central bank)
and the flow budget constraint for the central bank as:
Profits from central bank = iBb/P
where Bb is the bonds held by the central bank, i is the nominal interest rate on bonds, and Bd=M.
So an OMO does affect the LR government budget constraint, because it increases the PV of the flow of profits of the central bank, and so has LR fiscal consequences.
But the TIMING of those fiscal consequences is the government's choice, as is the decision whether to change PV(G) or PV(T).
Of course, under Ricardian Equivalence, the timing of T does not matter, only PV(T) matters.
If my old memory is still working (it's many decades since I worked this out), you can consolidate the CB and govt LR budget constraints, "book" the profits as the central bank prints money, and re-write the consolidated LR budget constraint as:
(B-Bb)/P = PV(primary surpluses) + PV(dM/P)
where dM is the increase in the stock of money.
Money is base money throughout; I am ignoring the central bank's operating cost, and the interest rate used in the PV calculations is the interest rate on bonds, adjusted for inflation ex post. And assuming money pays no nominal interest.
It's interesting to see how my LR budget constraints compare with the fiscal theory:
(M+B-Bb)/P = PV(surpluses) (I think by "bonds" the fiscal theory means "bonds in public hands", and excludes those held by the central bank, right? Because that would be double-counting, if we have M as a liability too.
Comparing the FTPL to my LR govt budget constraint, since (M+B-Bb)=B, we see that the imp0licit assumption of the FTPL is that the central bank makes no profits. And, if money paid the same interest rate as bonds (forever), it wouldn't. So the FTPL equation emerges as a special case of mine.
Sims, Sims. Sims.
Later. Got to fix an exhaust pipe.
Posted by: Nick Rowe | May 29, 2009 at 08:15 AM
Typo in the above. "Bd=M" should read "Bb=M"
Posted by: Nick Rowe | May 29, 2009 at 08:23 AM
I've been trying to wrap my head around Chistopher Sims' paper (amongst other things). I still don't feel I have my head fully around it. But here are some observations:
Let's take Sims' analogy of government debt to a corporation's shares.
Set aside the observation that government liabilities are the national unit of account (Sims recognises this difference, and it matters in the sort run, but shouldn't in the long run).
Suppose we had a corporation financed only by shares, and the shares paid a fixed dividend, but the dividend was payable in more shares, not cash. (Set aside the semantic question of whether we call them shares or debt). The corporation's profits can be used to retire shares.
In this case I would agree that the real value of the outstanding shares is determined by the PV of the corporations' profits.
Sims says this is exactly like government debt. The real value of the outstanding debt is determined by the PV of (real) primary surpluses. And that determines the price level.
So, what are the differences between the two cases?
The first differences is the direction of causation. I would assume the corporation maximises profits, and those profits are exogenous wrt the number of shares outstanding, and the interest rate promised on those shares. I don't assume governments maximise surpluses, and I don't think the surpluses they do run are exogenous wrt the national debt and interest obligations.
The second difference is that all corporate debt/shares has to offer a rate of return that matches the competition from other assets. There is one government "liability" - money- that is special, in that people will hold it even though they expect its rate of return will be much lower than competing assets. It has a strongly downward-sloping demand curve. So the rate of interest used in the PV calculation is not exogenous wrt the real value of the stock of money.
The third difference is that the (real) rate of interest on money is typically negative, so the PV calculation is weird, and Ponzi-esque.
A fourth difference is that (sometimes) the issuance of money is handled by a separate body (the central bank), that has some degree of independence from the rest of the government.
I would say that the weird rate of return characteristics of money stem from it's being used as a medium of exchange.
Again though, the Sims paper is interesting and insightful. And it certainly makes for an interesting comparison for countries which do and do not issue foreign currency denominated debt. Even if i don't fully accept the analogy, that doesn't mean there isn't some truth to it.
Posted by: Nick Rowe | June 01, 2009 at 07:02 PM