If cows were money, an increased demand for milk would cause a recession. People would stop spending their cows to buy goods and services, because if you spend your cows you don't have as much milk. Was the recession caused by an excess demand for milk, or an excess demand for money?
If gold bars were money, an increased demand for bling would cause a recession. People would stop spending their gold bars to buy goods and services, because if you spend your gold bars you don't have as much bling. Was the recession caused by an excess demand for bling, or an excess demand for money?
If dollars were money, an increased demand for savings would cause a recession. People would stop spending their dollars to buy goods and services, because if you spend your dollars you don't have as much savings. Was the recession caused by an excess demand for savings, or an excess demand for money?
The good which people use as a medium of exchange can also be used for other things as well. Cows can be used for milk; gold bars can be used for bling; dollars can be used for savings. An increased demand for the medium of exchange qua medium of exchange can cause a recession. But an increased demand for the medium of exchange qua provider of milk, bling, or savings, can also cause a recession.
Brad DeLong and I are coming very close to agreement. Here's Brad:
"You [Nick] will respond that what I call a "shortage of bonds" is really a shortage of money, as the spending velocity of money will fall as demand for currently-produced goods and services falls. I would say that the spending velocity of money has fallen because people are short of savings vehicles, and so are holding on to their money as a second-best to holding bonds. I would then say that if we had the same amount of money but more savings vehicles--if the government sold bonds and bought a dam, things would be fine, but if we had the same amount of savings vehicles but more money--if the central bank bought bonds for cash--we would not be fine. It seems distinctly odd to call that situation a shortage of money..." (I especially like the sentence I have bolded.)
It has almost come down to a semantic dispute. If I remember (I may not) my philosophy lessons from way back (learned, ironically, just a short distance from Brad's office at Berkeley), we are arguing about the referential opacity of demand functions. If I demand milk, can I be said to demand the medium of exchange, if, as a contingent fact, the medium of exchange just happens to provide milk, and I want the medium of exchange only for its milk?
But it matters. Because the way you frame the cause of the recession may influence where you look for a cure.
If you say that the recession is caused by an excess demand for milk, you start looking for ways to either increase the supply of milk or reduce the demand. Can the government breed some goats, and increase the supply of milk that way?
But if you see the recession as being caused by an excess demand for cows, you also start to think of other solutions. Perhaps the government could tax ownership of cows? Even, as a last ditch policy, make the cows stop giving milk?
If you think of the recession as being caused by a shortage of milk, then taxing cows, or making the cows stop giving milk, look like totally daft solutions. They wouldn't cure the shortage of milk. But they would cure the recession. With a tax on cows, or if cows stopped giving milk, people would start spending their cows again.
But I also see the point of looking at it from Brad's perspective. An open market operation of cows for goats, where the government imports cows, uses them buy goats, and exports the goats, may not work. The total supply of milk is the same, and if cow's milk and goat's milk are perfect substitutes, people have no additional incentive to get rid of their cows to buy the goods and services that are in excess supply. If, at the margin, the demand for cows is a demand for milk only, the demand for cows will increase one-for-one with the increased supply of cows, if that increased supply of cows is the result of an open market purchase of goats.
But why did the demand for cow's milk increase in the first place? Was it that all the goats went dry? Or might the increased demand for milk be a consequence of the recession, and so a consequence of the excess demand for cows? "I had better hang on to my cows, because I won't be able to sell my labour to buy milk in this recession!"
I don't know whether Brad is right about the shortage of other safe savings vehicles being the cause of the increased demand for money qua second best savings vehicle. But, even if he is wrong, this is something that might happen. If it didn't happen this time, or if it were just a consequence of the recession, not a cause, it might happen in future. But correctly diagnosing the proximate cause of the recession as an excess demand for the good that happens to be the medium of exchange, even if that excess demand is not a demand for that good qua medium of exchange, lets us start thinking about radical solutions.
And we need to start thinking about radical solutions, because as history teaches us, and as this recession reminds us (and judging by the Eurozone we are going to be reminded again) stuff happens. Things go wrong. Safe assets become unsafe. Goats stop giving milk. And we need a monetary system that is robust in the face of human stuff-ups. A shortage of milk is a problem, but it didn't ought to cause a recession too. A shortage of safe assets is a problem, but it didn't ought to cause a recession too. And if it didn't cause an excess demand for the medium of exchange, as a contingent side-effect, it wouldn't cause a recession to compound the original problem. And it's just because those side-effects are contingent, and don't follow of necessity, that I insist on my way of framing the problem. A shortage of safe assets may cause an excess demand for money. An excess demand for money will cause a recession. We can break that first link in the causal chain, because it's only a contingent link. The second link follows of metaphysical necessity, unless we switch to barter.
What's the solution that could prevent an excess demand for the medium of exchange, and so prevent general gluts? Silvio Gessell? Barter on Ebay? Funny money systems where you can't hold a positive balance? God only knows. But we ought to start thinking about it. Quasi-monetarism (which is really Keynesianism too, if you are the sort of Keynesian that recognises that Keynesian economics doesn't make any sense in a barter economy) can be a radical approach.
[And I recommend Kantoos (a very good German blogger, but here in Google English translation) for a good discussion of my long-running and productive argument with Brad on this topic.]
{And I will respond to Arnold Kling in a separate post.
Update: Brad DeLong replies.
I'm still not convinced that "medium of exchange" as such matters. What would happen if we used dollars for transactions but quoted prices in terms of cows? (In my own mind, before I read your post, I was thinking of quoting prices in terms of oil, which is easy, since you can look up price oil -- or by extension the oil price of money -- on Bloomberg in real time, and it's very much like using a credit card in a foreign country, where the transaction shows up on your statement in your domestic currency but the value is determined by the exchange rate.) And suppose cow-denominated prices were sticky. Would an excess demand for the medium of exchange cause a recession? I don't think so: it would just lower the money-price of cows, which would lower the money-price of everything else, since I have assumed that prices are sticky in terms of cows. But if there's shortage of cows, it would, I think, cause a recession, even though cows are not the medium of exchange, because everything except cows would then be overpriced.
Posted by: Andy Harless | December 12, 2010 at 10:57 AM
Are you familiar with Karl Polanyi? Your critique, of keynesian not applicable in barter economy, etc. to me sounds similar. but then again the writing of anyone who calls themselves an economist sounds similar to me.
Posted by: edeast | December 12, 2010 at 11:46 AM
Andy: All goods and services are over-priced relative to rental apartments in New York City. Because of rent controls (just assume, for simplicity, that the controlled rents are indexed for inflation). Can that cause a recession? No. It causes an excess demand for apartments. It causes an excess supply of money in the market for apartments, but that's all. It cannot cause a general excess supply of all goods and services. Only an excess demand for the medium of exchange can do that.
The fact that the medium of exchange is usually the medium of account too, plus the fact that prices are sticky, can cause problems. It can prevent the real supply of money rising when there's an excess demand for money.
But suppose cows are money, and there's an excess demand for milk. A fall in all prices, measured in cows, will increase the real value of the supply of cows, but won't increase the amount of milk they produce. An increase in the price of milk, relative to cows and everything else, eliminates the excess demand for milk. This raises the rate of return on holding cows (like the real yield on holding money), and reduces their velocity of circulation. Now, if the price of everything else falls, we can eliminate the general glut of everything else, via MV=PY.
Posted by: Nick Rowe | December 12, 2010 at 11:47 AM
edeast: I did read some Karl Polanyi decades ago. My memory is not good. The Great Transformation? I didn't know he had written on money. But I am not the first to say that Keynesian economics is not applicable to barter economies.
Posted by: Nick Rowe | December 12, 2010 at 11:54 AM
ya great transformation, he's got some papers on money, "semantics of money uses", I've just been studying him for a history paper. After the great transformation he went about describing/researching premarket societies. Markets required medium of exchange money.
Posted by: edeast | December 12, 2010 at 12:02 PM
I don't think the rent control case is analogous, because prices aren't quoted in terms of New York apartments. New York apartments can be isolated from the rest of the economy because they trade in a separate market against money, but if rents were flexible in terms of money and all other prices were indexed to New York rents, then an excess demand for New York apartments would cause the price of everything to go up, and there would be an excess supply of everything.
Posted by: Andy Harless | December 12, 2010 at 12:13 PM
(everything except New York apartments, that is, and maybe money)
Posted by: Andy Harless | December 12, 2010 at 12:15 PM
“A fall in all prices, measured in cows, will increase the real value of the supply of cows, but won't increase the amount of milk they produce.”
Is that a good analogy? The real value of the store of value of money is not a fixed real yield of the medium of exchange value of money, is it? Aren’t the two correlated both in nominal and real terms?
Which leads me to the question – what is your intuition in positioning the cow as the medium of exchange and the cow yield as the store of value? Why not vice versa? Does it matter to your explanation? Can you do it the other way around? (serious question relative to the ambiguity I see in the implicit yield structure).
Posted by: anon | December 12, 2010 at 01:10 PM
Hmm, this OP seems just around the corner from the idea of a Balance-Sheet Recession. When Nick says "excess demand for savings vehicles" I would read in the idea "business paying off of existing debt instead of investing in business expansion".
It's the same as a mortgage. Any financial planning book will tell you that paying off a mortgage in Canada is an excellent financial move. In terms of avoided interest which are paid for with after-tax dollars and no MER, it's hard to beat. My point is that paying off debt, particularly paying it off ahead of schedule, is just as much an act of saving as stuffing the money in a savings account.
So an excess demand for money can be realized in the form of businesses finding it preferable to pay off existing debt early instead of investing in expansion. Furthermore, I like to think of people as Bayesian actors. We define Safety as a Bayesian function where the Probability of A defaulting is seen as a function of B defaulting, where B is another market player. Since we have had a large number of B's lately, people have re-evaluated A to be higher, that is, less safe. So when people feel less safe about investing, they pay off debt.
Now imagine that both the business and consumers want to save, that is, pay off debt, instead of investing. They can't, as a whole, unless they transfer funds to G, the government. Now imagine the Government wants to save too because the business and political classes are worried about government debt levels. Then all three components of the economy want to pay off debt and save. Collectively they will fail, but in the process investment and business expansion will plummet.
In short Nick really hit on the idea of a balance-sheet recession, just cast in different terms.
Posted by: Determinant | December 12, 2010 at 01:25 PM
“A shortage of safe assets is a problem, but it didn't ought to cause a recession too. And if it didn't cause an excess demand for the medium of exchange, as a contingent side-effect, it wouldn't cause a recession to compound the original problem. And it's just because those side-effects are contingent, and don't follow of necessity, that I insist on my way of framing the problem. A shortage of safe assets may cause an excess demand for money. An excess demand for money will cause a recession. We can break that first link in the causal chain, because it's only a contingent link. The second link follows of metaphysical necessity, unless we switch to barter.”
But the shortage of safe assets DOES cause a recession because of the intensity of the demand for safe assets. The shortage is simply too severe not to spill over to the medium of exchange. The side-effect is NOT contingent, given the severity of the demand for safe assets that is at the heart of the problem. The side-effect is assured. The severity of demand for safe assets means safe assets can’t expand fast enough without enveloping the medium of exchange in that demand almost immediately. The severity means that the spill over, rather than being contingent, is assured. So the shortage of safe assets WILL cause a recession on that basis. And the link is only contingent under those benign conditions where the demand for safe assets is not severe enough to cause a recession. The metaphysical process is due to the original severity of demand for safe assets.
So the overall demand for safe assets includes the demand for money as a safe asset, thereby hijacking the effective supply of money in its use as a medium of exchange. I see no way of shielding the medium of exchange from its inherent property as a safe asset – as in eliminating milk production from cows. The answer must still be to increase the supply of safe assets, either by fiscal policy or by radical monetary policy such as buying stocks. That means cow spending (deficits) or cow easing (cows for lousy goats). But if its cow easing, the goats have to be really lousy to get any policy traction – which means putting the easing machine (the central bank) at considerable fiscal risk. So it’s all fiscal in the final analysis.
Posted by: anon | December 12, 2010 at 03:28 PM
"An excess demand for money will cause a recession"
Is that really true? It would seem to depend on the financial institutions that you are assuming, and how you define money.
If money = deposits,
Households can sell bonds to the banks and use the proceeds to open new deposit accounts. The financial sector will have more deposit liabilities, and fewer bond liabilities -- they won't need to sell so many bonds to households, as they will be able to fund themselves more with deposits. For a given rate of interest banks will sell fewer bonds and households will have more deposits up until household deposit demand is met.
If money = currency,
Households can withdraw their deposits for currency. Bond yields will rise as banks fail to attract enough funding. Then the CB will create more currency to keep the rate at target, i.e. up until households, on the margin, redeposit as much currency as they withdraw.
Again, for a given interest rate, household demand for currency will be met.
The above arguments also work for firms. The apply to the non-financial sector as a whole.
We have a flexible money supply, in which households are able to have, in aggregate, more deposits as well as more currency.
Moreover, we have data about household deposit holdings and currency holdings. The data doesn't support the idea that households wanted more deposits or more currency, because if they did, they could have had more by shifting their asset holdings, without needing to spend less on goods at all.
But the one thing that households cannot do by financial operations, is to increase their net worth.
They can convert all their financial assets to deposits, currency, or anything else that you want to call money. But they cannot make themselves richer -- they cannot increase their net worth by doing any of these operations. The only way to do that is if they all saved.
Therefore household net worth would be the natural place to start looking, if you were searching for the cause of excess savings demands.
Posted by: RSJ | December 12, 2010 at 03:47 PM
Which brings me back to the pay-down-debt function. Paying down debt reduces financial risk and improves net worth but at the cost of consumption.
Frame "Excess demand for money" as "Demand for money for current consumption" or "Demand for money to pay down debt". If households and/or firms choose the latter over the former, consumption will decrease as normally there is enough money to service debt as it comes due and maintain current consumption. You can't accelerate debt payment and maintain current consumption with the same amount of money. You need more money or you need to spend less. That's the crux of the excess money demand.
Instead of looking at household currency or deposit holdings, look at total debt, structure of the debt and any rise in debt paydown.
Posted by: Determinant | December 12, 2010 at 04:19 PM
I think my views on this are a mixture of your views and Andy's. DeLong's theory that recessions are caused by excess demand for bonds seems like a special theory, whereas quasi-monetarism seems like a generally theory. DeLong's theory MIGHT be useful at the zero bound (although I have my doubts), but certainly cannot explain the Volcker recession, when T-bills and cash were not at all close substitutes. Cash was the medium of account, and the recession was caused by a sharp slowdown in the rate at which the medium of account was losing purchasing power (plus sticky wages and prices). Excess bond demand contributes nothing significant to that story.
So in that sense I agree with you. But as you know I like the medium of account approach better than the medium of exchange approach, and hence agree with Andy's point about rent controlled apartments--they are not the medium of account. In a sticky wage/price framework, deflationary monetary policy (medium of account) is all you need to get a recession, you don't need to assume anything about an excess demand for the medium of exchange. However, because the MOE is almost always the MOA, the two explanations usually coincide.
Posted by: Scott Sumner | December 12, 2010 at 05:01 PM
Scott: Brad DeLong doesn't say that *all* recessions are caused by an excess demand for bonds. Just this one. He agrees (IIRC) that the Volker recession was quite different. He says that all recessions are caused by an excess demand for some non-produced good, might be money, might not. I'm trying to persuade him to see an excess demand for the medium of exchange as always the *proximate* cause. He's actually a lot closer to a "quasi monetarist" perspective than many might expect.
Yep, you, Andy and I disagree on the relative importance of MOE/MOA. I've gotta think up a clearer thought-experiment, to smoke out the inner MOE-theorist in both of you!
But right now, I'm working on my response to Arnold Kling! Things are getting very lively, after a quiet time.
And I still haven't got around to dealing with Andy! From a MOE perspective, not your own MOA response.
Must look for Bill's take on all this. He's the reliable MOE guy.
Posted by: Nick Rowe | December 12, 2010 at 06:05 PM
Nick:
Great post.
Those of us with the Mises/Austrian background define money as the media of exchange, but don't have a tendency to assume that the demand for money is mostly a transactions demand. People choose to hold money because they find it a good store of wealth. OK.
I also agree that even if Scott is right, and the reason for the increase in the demand for money was almost entirely because of expected and then prove incompetence by the Fed, it certainly is possible that a flight to safety could cause a spillover to an excess demand for money. As I have said, this makes targeting short and safe interest rates or limits on the Fed's asset purchases to short and safe assets a _big_ mistake.
As for the MOE and MOA, the only way the MOA can be effective in a decentralized economy is if shortages or surpluses of the MOA causes monetary disequilibrium, shortages or surpluses of the MOE. Otherwise, the market for the MOA just won't clear, or, more likely, the UOA price of the MOA will change--the MOA won't be effective.
Posted by: Bill Woolsey | December 12, 2010 at 08:53 PM
The shock after Lehman was purely monetary, as T-bills rallied while the equally safe T-notes fell in September 2008.
If markets expect that the monetary policy will be too tight for the next three years, excess demand for savings vehicles will be created:
A hypothetical. The Fed declares in 2007 that for the next three years the fed funds rate will be 15%, and there will be a return to normal policy in the fourth year. Markets will price the 10 year note according to the EH, and the price of the 10 year note will crash. The attractive low price of the 10 year note will create the excess demand for safe savings vehicles.
The reality. Sometime in October-December 2008 markets saw that the monetary policy will be too tight for the some time. In early October 2008 markets thought that the monetary policy will be too tight for a few months. In December 2008 markets thought that the monetary policy will be too tight for the next three years. The attractive low price of the 10 year note (yield approx. 2.2%) created the excess demand for safe savings vehicles.
Theoretically it is possible that a shortage of T-notes will create an excess demand for money, this in turn will cause the recession. But there is no good historical example.
Posted by: The Money Demand Blog (123) | December 12, 2010 at 09:12 PM
"But suppose cows are money, and there's an excess demand for milk. A fall in all prices, measured in cows, will increase the real value of the supply of cows, but won't increase the amount of milk they produce."
That can't be right, can it?
If I have a $ 1000 in cash and a $ 1000 government bond, the real value of both goes up when prices decline.
Posted by: anon | December 13, 2010 at 07:45 AM
Nick, Yes, Kling is much more anti-monetarist than DeLong. I put a comment on Kling's recent post where he asked for my view of the role of money in the Great Depression.
Posted by: Scott Sumner | December 13, 2010 at 02:20 PM
"If dollars were money, an increased demand for savings would cause a recession. People would stop spending their dollars to buy goods and services, because if you spend your dollars you don't have as much savings. Was the recession caused by an excess demand for savings, or an excess demand for money?"
Except that as soon as households get dollars, they immediately rush to purchase bank deposits, bonds, or other financial assets with those dollars.
If they were truly hoarding dollars, then money market rates would go up.
But the marginal cost of a dollar is *zero* now.
People are falling over each other trying to get rid of dollars, so dollars cannot be what is demanded.
The problem is that they are falling over their dollars to buy more financial assets, or non-produced assets, as Brad would point out.
Posted by: RSJ | December 13, 2010 at 06:09 PM
RSJ:
You and Nick share the same assumption that a householder with extra dollars will purchase a financial asset like a deposit or bond. I contend that's not necessarily true. A householder may have debt and seek to pay off that debt. Insofar as this creates a positive return through avoided interest payments, it is the same or better than buying a deposit.
I contend that the extra desire to pay down debt is what drives rates so low. Lenders are competing for borrowers, instead of the other way around.
You cannot assume that the liabilities of a household are zero. Typically they are not, usually mortgage or consumer debt or both. Why focus on a complicated definition of the role of money when a simple micro choice will produce the same result?
Posted by: Determinant | December 13, 2010 at 09:52 PM
"What's the solution that could prevent an excess demand for the medium of exchange, and so prevent general gluts?"
Stop assuming every instance of price deflation and/or real GDP contraction is an aggregate demand shock.
Start thinking about increasing the amount of medium of exchange with currency and not the demand deposits created from currency denominated debt because demand deposits can be paid off (I think) and demand deposits can be defaulted on. Most of the time there is no worry about the amount of currency going down.
Posted by: Too Much Fed | December 13, 2010 at 10:27 PM
"What's the solution that could prevent an excess demand for the medium of exchange, and so prevent general gluts?"
Start thinking in terms of how currency and the demand deposits created from currency denominated debt differ from each other, in terms of budgets (balance sheets) of the major entities of an economy including the time differences between spending and earning currency denominated debt can create, in terms of retirement date changes and retirement inequality, and in terms of an economy going from supply constrained to demand constrained.
Posted by: Too Much Fed | December 13, 2010 at 10:35 PM
Scott Sumner said: "Nick, Yes, Kling is much more anti-monetarist than DeLong. I put a comment on Kling's recent post where he asked for my view of the role of money in the Great Depression."
How about the role of the medium of exchange, specifically there were too many demand deposits created from currency denominated debt and not enough currency?
That means an economy can be "price inflated" with the wrong medium of exchange.
Posted by: Too Much Fed | December 13, 2010 at 10:41 PM
"If cows were money, an increased demand for milk would cause a recession. People would stop spending their cows to buy goods and services, because if you spend your cows you don't have as much milk. Was the recession caused by an excess demand for milk, or an excess demand for money?"
If cows were money an increased demand for money would not cause a recession because other monies (harnesses, yokes, farm implements) can step into the gap and satisfy demand. An increased demand for milk shouldn't cause a recession because milk comes from goats, sheep, horses, yaks etc., and higher prices for this milk will draw in supply, keeping animals out of the slaughterhouse and in the milking barns.
The trick to avoiding gluts comes down to removing all monopolies and coercive measures that limit people's ability to switch from cows-as-money to cow yokes-as-money. This means free banking, not a central banking monopoly unresponsive to the demand for liquidity. The trick is also to ensure that people have the ability to turn to alternative milks, and that suppliers have the ability to provide this milk. This means free capital markets, not markets in which large monopolies own some of the safest assets yet choose not to issue new claims upon those assets when people so desperately demand them.
Posted by: JP Koning | December 13, 2010 at 10:43 PM
About the safe assets/savings comments, are people looking to buy the highest yielding asset that won't go down in value or be defaulted on (think FDIC insured CD's or even paying off a credit card account with a greater than 10% interest rate)?
Also, are people believing that the demand deposits from currency denominated debt have brought forward from the future price gains in assets and when the demand deposits are defaulted on or even paid off, the prices on the assets will "spiral backwards" in time to a lower level? Once at the lower level, people will buy the assets again.
Posted by: Too Much Fed | December 13, 2010 at 10:52 PM
JP Koning said: "The trick to avoiding gluts comes down to removing all monopolies and coercive measures that limit people's ability to switch from cows-as-money to cow yokes-as-money. This means free banking, not a central banking monopoly unresponsive to the demand for liquidity."
How does that apply to the monopoly of currency?
Posted by: Too Much Fed | December 13, 2010 at 10:55 PM
anon said: "If I have a $ 1000 in cash and a $ 1000 government bond, the real value of both goes up when prices decline."
What if the bond goes down in value or is defaulted on?
Posted by: Too Much Fed | December 13, 2010 at 11:00 PM
JP: There's just a chance that a competitive supply of monies might solve the problem. Trouble is though, money seems to have a network externality, and be a natural monopoly. There *are* competitive monies in existence, those LETS things, for example. But they never seem to get very far, even in a recession.
Posted by: Nick Rowe | December 14, 2010 at 12:09 AM
Determinant,
"You cannot assume that the liabilities of a household are zero. Typically they are not, usually mortgage or consumer debt or both."
I am not assuming this. Repaying debt is like closing a short position. Whether you close a short position or open a new long position, in both cases your savings add to your net-worth (which is the definition of savings applicable to households).
But I agree that the "savings demand" can be driven by many sources: firms require a certain amount of earnings in order to meet creditor demands, and households may have mortgages or other obligations that require nominal savings each period.
So the savings demand is not just psychological, in the sense that when your net worth drops, you decide to save more in order to fund retirement. There are also inflexible savings demands that must be met each period.
"Why focus on a complicated definition of the role of money when a simple micro choice will produce the same result?"
I'm not sure what you mean here. The point is the micro has freedom that macro does not, so you cannot look at a partial equilibrium result and deduce the cause of recessions.
The point that Nick would argue (I think), is that when households decide to save more in the form of bonds, then this requires others to invest by the same amount. Therefore saving in the form of bonds cannot result in insufficient demand for goods, since the corresponding borrowers will demand goods with the proceeds of the funds that were borrowed.
But that assumes that the credit markets behave in a very peculiar way.
The assumption is that all borrowers are borrowing in order to purchase final output. But in the credit markets, some participants are borrowing in order to lend again. Others are lending because they have borrowed. Others may be borrowing to buy land or other non-produced assets.
It's not the case that everyone selling a bond is going to use the proceeds to buy goods, neither is it the case that everyone purchasing a bond is deferring the consumption of goods.
Now you can argue, that after netting out any transactions that do not consist of deferral of consumption or the purchase of a capital good, you get that aggregate savings = aggregate investment. That's of course true. But the interest rate that clears the bond markets is going to be the rate that clears the all the transactions, not the net transactions.
It could well be the case that the interest rate does not clear the demand for final borrowing and final lending, even though it does clear all transactions. That's an aggregation error.
Here is an example. Suppose the economy consist of 3 people: A is neither saving nor investing. A spends all of his income but still likes to buy and sell in order to rebalance his portfolio.
B wants to defer consumption and C wants to borrow in order to purchase a capital good.
B's supply curve is r(Q) = Q/20. C's demand curve is r(Q) = .1 - Q/20.
If there was a commissar, interrogating market participants and preventing them from selling or buying assets unless they spent the proceeds on final output, or obtained the funds for purchasing assets only from the deferral of final output, then the equilibrating quantity would be 1, and the rate would be 5%.
B would lend $1 to C at 5%, and both B and C would happy, with no excess demands.
But suppose that the market already contains 10,000,000,000,000 worth of bonds yielding 6%, and these are held by A.
Now as soon as B attempts to put in a bid for bonds at 5%, or even slightly below 6%, A will sell him all the bonds that B is willing to buy. And simultaneous to that, C will only be able to afford to borrow Q = .8 at 6%.
So both C and B are price takers, and there will be an excess demand of 1.2 - .8 = 0.4 spilling out into the other markets.
The final equilibrium reached is incomes decline up until the amount that B succeeds in saving is the (lesser) amount that C disaves at the excessively high rate.
In general, the amount of net investment funded by the credit markets is several orders of magnitude smaller than the amount of credit market debt or the amount of transactions. And there is no a priori reason to believe that just because there is an interest rate that clears borrowing = lending in the sense of gross transactions, that this same rate clears the *demands* of the insignificant minority of market participants that are deferring consumption or borrowing to purchase final output.
The influence of these two groups on the rate of interest is basically zero. You have daily turnover of trillions of dollars, of which, a few billion is spent on actual investment.
Posted by: RSJ | December 14, 2010 at 01:08 AM
RSJ:
I really dislike models that assume households save by purchasing bonds. It really produces a different result than reality. It also misses several pertinent facts about modern consumer behaviour that affect how households.
First, a few observations:
1) Mortgage interest rates, lines of credit and other household borrowings always carry a higher interest rate than deposit rates (Canadian bank products are a good representative sample here). This is a simple result of banks engaging in their normal business model, borrowing at X% and lending at Y%, where Y >X.
2) In Canada mortgage and consumer debt must be paid for with after-tax dollars. Therefore when comparing the return on a debt repayment, you have to scale up the payment by household's marginal tax rate to compare it with other investment rates. So the nominal return on a debt repayment gets a 35% boost just because of the tax implications. If you combine this result with observation [1], debt repayment is nearly always a superior financial move for households.
3) The combined result of [1] and [2] is that a proper model of households with disposable income that they want to "save" should recognize that households will typically have debt (mortgage debt at least) will want to pay off their debts first before purchasing a financial instrument. This is why I don't care about bonds. Bonds are secondary. They are second in the pecking order of household savings. Not all savings are directed into bonds. The first destination of savings is debt retirement.
Ok, now that I have stated this, what does this mean on a macro level? Easy. Lets say under normal circumstances a household pays 20% of its income for debt serving. However there was just a economic shock, and the household wants to save more. So it wants to increase its debt payment to 25%. This is the first iteration of Nick "demand for money". A household has two choices in this case: increase its income or decrease its consumption. Decreasing consumption is the far easier choice (its immediate too), so by "saving", i.e. paying off debt, its consumption falls. Work this out on a macro level and you get a recession through falling demand or demand/supply mismatch (if some households try to earn more income and others cut their consumption).
Bonds don't enter into it. It's households deleveraging that causes demand to fall.
Posted by: Determinant | December 14, 2010 at 05:09 PM
"A household has two choices in this case: increase its income or decrease its consumption."
What if a household sells an asset (maybe even a retirement asset like some stock) to be able to keep spending in the present?
Posted by: Too Much Fed | December 14, 2010 at 05:31 PM
If the sale was successful, then that means there was a counterparty that bought the asset had enough money to complete the transaction. Since in this case there is enough supply and demand of funds to clear on this transaction, the "excess demand for money" is temporarily satiated. However since assets are limited, this just postpones the inevitable in the absence of a reliable stream of income, e.g. employment or retirement earnings. There is an equilibrium between the two where capital is annuitized.
Speaking of retirement, I am a big fan of life annuities for those without defined-benefit pension plans. Annuities remove the risk of outliving your money. Without annuitization the risk of decreased consumption is still there. Actually the case of the senior who outlived their money is a great real-world example of your idea, Fed.
Posted by: Determinant | December 14, 2010 at 08:24 PM
Determinant,
Any net interest income that the bank earns is passed onto households who are creditors of the bank (or employees of the bank, in the case of bonuses). When you write it all out, the household sector gets back in wage and capital income any interest payments that it makes.
So yes, for the individual buying the house it's good to focus on debt repayment after going into debt. But that doesn't give an explanation for the macro behavior. You want to explain why aggregate debt is decreasing. Why are there now fewer borrowers rather than those repaying debt? Or for that matter, why did aggregate debt increase faster than GDP (or the capital stock) prior to this crisis?
And for that, you need to talk about interest rates and return expectations.
You have two different definitions of the what the rate of interest should be -- should it be such that it removes the possible arbitrage from borrowing to buy an asset, waiting for 1 period, and then selling the asset -- i.e. is it set by expectations of future asset prices -- or is the interest rate going to ensure that the number of households desiring to defer consumption at a given rate is exactly equal to the number of households desiring to invest at the same rate?
During the .com bubble, did households become suddenly patient, or did they expect asset prices to increase? And when the bubble burst, did households become more impatient? In other words, is asset demand determined by the utility of deferring consumption for a period or is it determined by expectations of future price increases?
Which of these two mechanisms will set the rate?
Only by imposing some rational asset valuation assumptions can you assume that the results of these two valuation approaches yield the same rate. B
But there is no one to enforce rational asset valuation assumptions, and in that case, investor profit motives will ensure that the first mechanism sets the rate, and that rate can deviate from the rate that would have obtained if the interest rate solved some intertemporal utility maximization problem of a single household that was not allowed to book capital gains or otherwise re-sell the asset to someone else.
Posted by: RSJ | December 14, 2010 at 10:45 PM
"Any net interest income that the bank earns is passed onto households who are creditors of the bank (or employees of the bank, in the case of bonuses). When you write it all out, the household sector gets back in wage and capital income any interest payments that it makes. "
Sorry RSJ, I don't agree with that. It's false. Banks are chartered companies, they pass on their profits in the form of dividends. Unless by "creditor" you meant "owner", then your post misses the mark. Most bank stock in Canada is owned by institutions, namely mutual funds, pension funds, and any other form of professional investor you care to name. If a household recognizes profit from bank stock, it will be credited to pension or retirement investments which can't be sold easily, if at all. If it's in a pension fund (and no pension funds is seen without bank stock here) then the money is locked away and can't be touched.
I contend that that plumbing in your model isn't complete.
However when you mentioned that paying off debt is like closing a short position, you really did say something I agree with. My theory is that any model should have households closing their short positions before purchasing financial assets. The reason is that short positions or mortgages as debt constructions have legal and financial penalties for default. I MAY save by buying a bond. I MUST pay my debt obligations. That's I contend that any accurate model must have households paying off debt first. I want models to be clear on that behaviour.
When Bayesian actor households react to a large shock and recalculate their financial position, they may choose to pay off more debt. That capital gets returned to banks but the banks can't lend it out to more conservative Bayesian households. The capital piles up in banks.
Posted by: Determinant | December 15, 2010 at 02:12 PM
Discussion in the Woolley household "so Nick is milking this for all that it's worth."
Posted by: Frances Woolley | December 15, 2010 at 04:25 PM
Determinant,
Yes, I was also thinking of equity holders as creditors. Perhaps a bit sloppy on my part. But you see what I mean -- all the money gets passed back to households. So the issue is not that some households have savings demands, but that the investment demand is less, at whatever the current interest rate happens to be.
That's why it's an interest rate problem, fundamentally. But here, households do take gains, or "total return" into account. An individual household may well be prepared to pay $X in mortgage payments each month in order to obtain rental utility. But if they think the house is going to decline in price by 20%, then they will require a much lower X, easily demanding even negative interest rates to compensate them for the expected loss. On the other hand, the seller of the house is not going to take a 20% haircut, in some cases because they cannot -- e.g. they can't afford to bring a check for 20% of the property's value to closing, or the bank can't afford to recognize the capital loss. So you can get into a situation in which everyone would become insolvent if they began to rationally value assets. But also, they would need to be convinced that the price is *permanently* lower by 20%. If they think that, in 3 years time, the market will recover, then they won't sell for 20% lower now, and there will be 3 years of market failure.
Similarly, if the buyer thinks the house prices will go up, they may be willing to pay much more than X in order to obtain rental utility + savings.
So now imagine how distorted prices would become after a 30 year period of secularly increasing asset prices, and earnings growth rates far in excess of GDP growth rates, all funded by increasing household indebtedness. Businesses need to be convinced that the last 30 years is a fluke, and that earnings will be permanently lower going forward. They need to fall by about 50% to go back to historical multiples, and this plunge in profitability needs to happen without anyone going bankrupt or getting fired. Asset holders need to recognize that many of their assets are worthless. Banks need to recognize the same. How likely is that to occur, when everyone is levered up and has committed to delivering excessively high returns to each other? Pension funds and capital investors need to be convinced that their reasonable return is only 6%, despite getting 8-9% over the last 30 years. How are they going to do that without acknowledging that they are insolvent? So you can get into a situation in which basically everyone would be insolvent if they allowed the goods markets to clear.
None of this "real" -- in the sense that if you are just looking at how real resources are allocated, you don't see massive misallocation and you don't have an explanation as to why firms would liquidate both labor and capital rather than cut prices and accept lower levels of earnings, or why there are ghost neighborhoods of new houses standing empty for years. This is fundamentally a financial problem that become invisible when you consolidate everyone's balance sheets.
Posted by: RSJ | December 15, 2010 at 09:17 PM
"Yes, I was also thinking of equity holders as creditors. Perhaps a bit sloppy on my part. But you see what I mean -- all the money gets passed back to households. So the issue is not that some households have savings demands, but that the investment demand is less, at whatever the current interest rate happens to be."
I see what you mean, but you didn't see what I meant. Pension funds and other retirement savings are special, very special indeed. That is where most household "benefit" from bank stock lies. Company Pensions and Locked-in Retirement Vehicles are not fungible, by law. A household can't sell, pledge them, borrow against them, or in any way use them for present consumption. From the perspective of the present, the funds are out of the picture. If they are in a personal retirement account they are most often in a mutual fund, so the ownership is still indirect, or even if they are directly held in a retirement account, there are huge tax implications for using them for present consumption.
The law creates a special walled-off pocket for retirement funds. You can't figure them into present consumption or to pay off immediate debt. It's entirely unrealistic to do so. A household can't realize the value of their retirement holdings in that way.
I don't know how a debate on the nature of money and the role of household disposition of income evolved into the nature of mortgage payments.
I don't follow your reasoning on mortgage payments. X is the "rental utility level", the level of rent a household currently pays and at which it will switch to home ownership. So good so far. I don't think your idea about demanding negative interest rates works in Canada, due to the different nature of our mortgages. Mortgage interest rates are locked in for a period of 6 months to 5 years, after which the outstanding principal is refinanced at current rates. A 25-year mortgage consists of five mortgage terms of 5 years each, with 5 refinancings.
First, I can't see banks willing to lend on the expectation of a 20% loss in value of the home. Second, a renter who wants to convert to ownership who has that expectation wouldn't want to purchase because of the very real risk the bank wouldn't refinance the full outstanding balance because the current value of the home won't provide sufficient security. The renter-cum-owner would be on the hook for the unsecured, outstanding balance (according to Canadian law).
In such a scenario you would see prices falling to the clearing rate, a rash of mortgaged owners turning their homes into rental units (to carry the mortgage), low sales volume and long time on market (reflecting sticky prices from mortgaged vendors).
But that's not the current Canadian reality because Canadian banks never relaxed their lending standards. The rules stayed in place.
Posted by: Determinant | December 16, 2010 at 12:06 AM
"I don't think your idea about demanding negative interest rates works in Canada, due to the different nature of our mortgages."
It doesn't work anywhere. That's the point.
Canada is a nice, clean example, because mortgages are recourse. In that environment, the would-be seller may not be able to afford to sell at the market clearing price, because he may not have enough equity to absorb the loss.
Therefore the asset price doesn't decline, or rather it takes a long, grinding period of declining prices, during which time the spot market for houses is not clearing.
So this is an example of balance sheet effects preventing the goods markets from clearing, because the various actors cannot afford to recognize the losses on their balance sheets from selling the good at the market price.
This is no different than a firm that has a certain required cost of capital not being able to lower prices in response to a decline in demand for its products. The firm *must* have a certain margin, and when it can't get that margin, then it liquidates capital rather than maintaining the same capital stock but earning a lower rate of return on that stock. And the reason why the firm can't lower its margin is because the pension fund that owns the firm requires that margin, otherwise it will become insolvent.
So you have these interlocking balance sheet claims that prevent the goods markets from clearing, and I am just giving examples of this.
This is a separate issue from whether households decide to repay debt before buying new bonds.
I'm not disputing that point, only arguing that you don't need these effects in order to explain gluts.
General balance sheet considerations are sufficient to explain the glut, without making distinctions about the priority of debt-repayment over balance sheet expansion.
Posted by: RSJ | December 16, 2010 at 12:59 AM
Nick; Are you amening Woolsey? on the moa/moe.
Posted by: edeast | December 17, 2010 at 10:30 PM