Arnold Kling  

I Deny (the significance of) MV = PY

PRINT
Kling vs. the International Mo... Arnold's Bizarre Monetary Econ...

Jeffrey Rogers Hummel writes,


Orthodox monetarists attributed such shocks to declines in the rate of monetary growth, whereas traditional Keynesians blamed declining autonomous expenditures. Both of these sources are captured in the well known equation of exchange: MV = Py, in which MV (money times its velocity) is equivalent to aggregate demand, and Py represents nominal GDP, the product of the price level and real output. In other words, a fall in velocity (V) is equivalent to a Keynesian fall in autonomous expenditures, which can happen only if people in the aggregate are holding (or hoarding) more money. Although this basic truth is sometimes overlooked in the recent debates over fiscal policy, in which economists replay (often with far less theoretical sophistication, despite greater mathematical pizzazz) the forgotten Keynes versus the Classics controversies, a negative shock to aggregate demand must involve either (a) a decline in the money stock's growth rate or (b) an increase in the demand for money.

I am going to take on the paragraph quoted above. Keep in mind that this has very little to do with Hummel. He speaks for the vast majority of professional economists. I am sure that one blog post is not going to be sufficient to change anyone's mind. But I have to start somewhere.

Basically, I am going to attack MV = PY. Even though it is obviously true. It is an identity. It is true even if M stands for Marbles, Manatees, M&M's, or Mackerel. If the supply of mackerel goes up, and nominal GDP stays the same, then a monetarist (mackeralist?) says that the velocity of mackerel went up down and a Keynesian says that people are hoarding mackerel.

What mackerel and money have in common is that they are both stores of value. What is different about them is that money is a standard unit of account, but mackerel are not.

The unit of account function means that a decline in the value of money is by definition an increase in the price level. The price level is an average price of goods and services, using money as the unit of account.

But people want money as a convenient store of value (and medium of exchange). So, if the government is behaving itself, the prices of most goods and services in monetary terms do not change very much from day to day. If the government is not behaving itself, and it is trying to fund a massive deficit by printing money, then we get the phenomenon of too much money chasing too few goods. Money starts to lose its appeal as a store of value. This is hyperinflation.

However, if we go back to the case where the government is behaving, then the quantity of money is more like the quantity of mackerel. It is just one more quantity out there. People who go to their farms, factories, and offices every day do not care about how much money is in circulation, any more than they care about how many mackerel have been caught this week. Ordinary businesses set prices and make decisions based on what they saw in their markets yesterday and what they expect to see there tomorrow.

The Federal Reserve can manipulate some market interest rates by printing money and using it to buy securities. If it buys massive quantities of mortgage securities by printing dollar bills, it can lower the interest rate on those securities. In theory, the Fed could do this by exchanging mackerel for mortgage securities just as well as by exchanging money for mortgage securities.

However, I am not sure how long they can keep it up. At some point, market participants will think, "The Fed is not going to keep buying these things forever. When they stop buying, the price will fall." And such thinking will dampen prices (raise interest rates) today. The longer the Fed fights the market, and the more government keeps spending on other stuff, the greater the likelihood that too much will be financed by printing money, and we will have hyperinflation. As long as we know that the Fed is afraid of hyperinflation, then it has a limited supply of money, just as it would have a limited supply of mackerel.

Suppose that the Fed had decided to print a lot more money at some point last year. The Scott Sumner thesis is that this would have raised nominal GDP. In my view, the fall in nominal GDP was due to the fact that real GDP had to fall. Real GDP had to fall, because the economy was beginning a Great Recalculation. The Great Recalculation was mostly due to the end of the housing bubble and the shrinking of the financial sector. It was probably exacerbated by the panic generated by Paulson and Bernanke. I do not see how the Recalculation was helped by the bailouts, which were huge transfers from future U.S. taxpayers to current large creditor institutions, including many overseas. I do not see how the Recalculation would have been helped by the Fed suddenly printing a whole lot more money. In terms of MV = PY, I see PY as largely outside of the Fed's control--the P part was determined by the combination of habit and gradual adjustment in the Great Recalculation, and the Y part was determined by the frictions involved in the Great Recalculation. So if we could rewind the tape to some time in 2008, hold everything else equal, and have more M, I think we would see essentially a 100 % offset in V. Just as we would if M stood for mackerel.


Comments and Sharing





COMMENTS (25 to date)
winterspeak writes:

ARNOLD: I am similarly pessimistic about changing your mind on this, but let's try.

Say an economy decides to save more money. That is, they will put more of their income in the bank, and spend less of it on stuff.

Given that banks do not lend out deposits, won't that economy's income fall? If their income falls, won't that interfere with their plans to save? Wouldn't this be characterized as a decrease in velocity?

Now, if your income falls, but you have nominal debt, isn't your real debt burden increasing?

"If the government is not behaving itself, and it is trying to fund a massive deficit by printing money, then we get the phenomenon of too much money chasing too few goods." But Arnold, where does the private sector get its money from? If the private sector is a net saver (which it is), then that means some other sector needs to be in net deficit, and that sector is the Govt. I'm sure you would agree that the Govt can print money. We all know we are no longer on a gold standard (except maybe Hummel, who believes that the US can default, and therefore must believe that it is in some way limited in its ability to create $s otherwise, logically, this is impossible). So therefore, why does the US need to "fund" a deficit? Wouldn't it be more accurate to say that the deficit was funding private savings?

Again, by your own logic, hyperinflation is caused by too many goods CHASING too few dollars. If the dollars are sitting idle in savings accounts (and therefore reserve accounts), then clearly they have no inflationary consequence.

Devin Finbarr writes:

Arnold-

I do not see how the Recalculation would have been helped by the Fed suddenly printing a whole lot more money.

In the housing sector there was indeed a great recalculation. The demand for housing was artificially high and now carpenters will need to find new jobs.

But the current bust is not just in housing, it's across the board. Perfectly good factories, that are perfectly capable of producing automobiles are sitting idle. This is the same mystery of any recession/depression. Before the recession, cars are produced at a massive level. During the down turn the factories sit idle for a few years. Then after the recovery more cars are produced than ever. Why the idle factories? It's clearly not a result of "real" recalculation.

I think the easiest way to see the problem is through a balance sheet perspective. Every person has a personal preference for a paper wealth to expenditures ratio. A senior needs to have enough paper wealth to cover expenditures until death. A young adult might be saving up for a year of travel, etc.

A recession/depression happens when a massive bubble pops. Usually that bubble has been in the credit markets, as a result of maturity transformation in the banking sector. The collapsing credit bubble causes stock market and real estate bubbles to also collapse.

When a bubble bursts people’s total paper wealth drops sharply. Every person has a target paper wealth to expenditure ratio. If paper wealth falls, a person must cut spending or else they will run out of retirement money.

When people cut spending they do so by cutting purchases on luxuries and durables. Businesses producing luxuries and durables see their income fall, and lay off thousands of workers. If the fall in paper wealth is large (10+%) mass unemployment will result.

Thus the causation is burst bubbles destroy paper wealth (ie cause a purely nominal change to balance sheets). Reduced paper wealth causes a rise in the demand for money ( ie velocity, although I hate the term velocity). A rise in the demand for money causes deflation. Due to frictional effects - the non-neutrality of reduced spending, sticky debt contracts and sticky wages - deflation causes unemployment and

The government can stop the deflation and the fall in real output by meeting an increase in demand for money with an increase in the supply of government paper. The automatic stabilizers do this as tax revenues fall and spending rises to pay unemployment. But this is the worst way to restore the economy as it is slow and requires people to lose their job first.

The government could stop the recession instantly by initiating a policy of balance sheet repair. The easiest way to do this would be to simply take all outstanding currency and bonds ( bank accounts, CD’s, treasuries, investment grade bonds) , multiply the face value by 20%, and print enough currency to back the new higher values. The policy is clean, quick, and would avoid the unemployment altogether. Another slightly slower method of balance sheet repair is to declare a tax holiday and print money to fund government operations.

Winterspeak-

I don't see why fixed nominal debt contracts would cause unemployment. They would cause bankruptcy, but bankruptcy does not necessarily produce unemployment. As long as the business makes an operating profit, bankruptcy just results in wiping out the existing equity holders and converting the debt holders to equity holders. Conversely, if a business is making an operating loss, it will lay off workers regardless of its debt position.

Empirically, during this recession many companies have laid off workers that have no problems with debt contracts/bankruptcy (everyone from architecture firms to Microsoft). A fall in aggregate demand results in unemployment because of the non-neutrality of reduced spending. People cut spending on fixed capital, durables, and luxuries first, and those businesses cut workers because demand has dried up for the products and so they need less workers to meet the new lowered demand.


winterspeak writes:

DEVIN: Fixed nominal debt just means that price declines cannot make up for a fall in velocity. In MV=PY, the question is always why lower V cannot be offset by lower P.

Bankruptcy eats into the capital that banks hold, and bank lending is capital constrained (although it is not reserve constrained, as many believe). To the extent that falling bank lending leads to lower aggregate demand, then bankruptcies would drive unemployment (as falling aggregate demand drives unemployment).

Devin Finbarr writes:

as falling aggregate demand drives unemployment

Also, this ends up being circular. It still leaves the question open - why does falling aggregate demand cause unemployment rather than lower prices? Falling bank capital tells us that banks have a pro-cyclical impact on V, but it still doesn't tell us why falling V causes unemployment.

Devin Finbarr writes:

Fixed nominal debt just means that price declines cannot make up for a fall in velocity. In MV=PY, the question is always why lower V cannot be offset by lower P.

A business is always going to maximize its operating profit, regardless of debt service. Let's say that at M*V = 1000, the profit maximizing price of a car is 10. If M*V falls to 1000, the profit maximizing price should be 9. So the car company will drop its prices, regardless of its debts. If debt was the only the factor, we would expect lower prices and no unemployment. Deflation would only result in a transfer of wealth from equity holders to bond holders.

IMHO, the unemployment comes because the fall in V is not neutral. The exchange rate between flour and money might not fall 0%, the rate between labor and money by 10% and the rate between cars and money might fall 20%. Thus the car company cannot just cut its prices by 10% and experience the same quantity of demand. Thus the company can no longer afford to make the same number of cars, and it has to lay off workers. This result holds whether or not the car company has any debt.

To the extent that falling bank lending leads to lower aggregate demand, then bankruptcies would drive unemployment (as falling aggregate demand drives unemployment).

That is true. But I'm not sure how much of the fall in the aggregate demand stems from reduced lending. If you look at the Fed's monthly consumer expenditure data, consumption falls right after the September/October 2008 stock market crash. Judging by a large number of anecdotes, the fall in the stock market caused people to stop spending. Investment plummeted at the same time. Did investment plummet because the banks ran into their capital requirements and stopped lending? Or because the stock market crashed? Or because consumption fell and so anticipated profits from investment fell? Or because companies simply wanted to preserve cash through the downturn? My instincts say it's the last factor, although I haven't researched it in depth.

winterspeak writes:

DEVIN: The reason why a fall in V cannot be offset by an fall in P is because, at a macro level, you cannot change P without changing incomes.

Remember, "velocity" is just an abstract term for "savings", or "income minus expenditure". If the "velocity" of money falls it means that people are saving more. So the MV=PY equation, by its own logic, has saving increasing on one side, and argues that (through lower P) this can be managed by lower incomes on the other. It's obviously nonsense when you put it plainly, and that's ignoring higher real debt loans.

So what happens is Y falls, which means unemployment. This also has the benefit of according with observed reality!

Your question on banks is a good one. Banks certainly fell below their capital requirements, and some (Citi, BofA) are almost certainly still operating below the regulatory capital requirements. But the Obama administration is turning a blind eye, so they keep staggering on. Bank lending is pro-cyclical, so when the economy slows, lending shrinks.

Devin Finbarr writes:

Remember, "velocity" is just an abstract term for "savings", or "income minus expenditure". If the "velocity" of money falls it means that people are saving more. So the MV=PY equation, by its own logic, has saving increasing on one side, and argues that (through lower P) this can be managed by lower incomes on the other.

No, you and/or economists have defined V in a non-nonsensical way. Forget the term V, I think velocity is a terribly mis-chosen word. Let's instead use "Expenditure Multiple" - EM. EM is the desired ratio of expenditures over cash balances. Aggregate demand = M*EM = P*Y. As desire for cash balances increases, EM falls. Prices fall too, until the desired EM is reached.
( also note that you have to define EM the way I do, because in your definition of the equation the units on each side of the equation do not match. A price is ratio of money per given quantity of goods. So the equation is: money * X = money/good * goods/year ) In order to make the equation balance, X must be a scalar over years. EM - Money per year over money supply - fits the equation. Your definition - (Money/year minus Money/Year) - does not.)

The key is that people do not desire a specific quantity of money, independent of the price level. Would you rather have 1 million Centauri Francs, or 1 trillion Orion Pesos? Well, that depends on the price of a Big Mac in each start system.

Thus an increase in demand for savings can be satisfied by a fall in the price level. EM is reservation price, a desired ratio, not an absolute quantity. P can fall to meet M*EM, without a fall in Y.

What I am describing is the Pigou re-balance effect. The question is, why doesn't the Pigou re-balance effect happen in practice? The answer is the frictional forces, specifically as I noted above, the non-neutrality of deflation.


winterspeak writes:

DEVIN: Don't get mad at me, I'm just using the standard terminology that everyone (including you) uses in MV=PY. I'm glad that you see the standard MV=PY means that P cannot adjust the way economists want it to. Y must adjust when V falls.

You can use EM if you want, but you run into the same problem -- one person's expenditure is another person's income. As you say, when the desire for cash balances increases, the expenditure multiple falls, but by your own logic this means that the income multiple falls too (or whatever you want to call the other side of the EM transaction). Once again, you have people finding that the more they try to save, the more their income shrinks.

Individual households can increase their savings, but the private sector as a whole cannot.

"The key is that people do not desire a specific quantity of money, independent of the price level." Hah! Changing prices at an economy wide level impacts incomes, and you cannot have savings be independent of incomes.

"Thus an increase in demand for savings can be satisfied by a fall in the price level". Really? A demand for savings can be satisfied by lower (nominal) incomes? And higher real debt burdens?

Devin, I think your argument is that real savings can increase if prices fall. My point is, at a macro level, lower prices must also mean lower income (certainly lower nominal income). To the extent that aggregate demand falls, something that we are seeing in practice today, the result is higher unemployment, and less real output/income as a consequence.

Nominal denominated debt means that people are targeting a nominal number, and not a ratio. If you can find me someone who says "I want an expenditure ratio of x" vs "I want to pay down my mortgage of $300,000" I will change my position, promise.

Until then, we're left with reality, which is an unemployment rate that has doubled, a CPI rate that is flat to modestly increasing, a dramatic fall in asset values, and a dramatic rise in nominal savings. As you say, my story of why P does not fall in MV=Py fits the real world more closely than yours. At this point, I think you're saying "Pigou would be right -- if it wasn't for friction!". Darn that friction -- always messing stuff up!

Personally, I don't like the "sticky wages" Keynesian argument for why P does not fall, which is why I don't make it here. The direct link between prices and income seems more pertinent to me, and I also find Fisher's debt deflation very persuasive. Fisher was someone who probably agreed with Pigou before the Great Depression, but changed his mind and sided with Keynes afterwards. Losing a fortune probably focused his mind, which may be why he changed. I don't think any academic macroeconomists were wiped our like Fisher was, which may explain their relative intellectual sclerosis ; )

Anyway, you can either see the link between prices and income (if you've ever sold anything for a living, it's pretty straightforward) OR you can keep pondering that mysterious friction that keeps Pigou from happening in real life. The final word is yours!

Ben writes:

Winterspeak- if banks don't lend out deposits, what exactly do they lend out? I'm genuinely confused here.

Todd writes:

Winterspeak, you seem to know a lot more about money and banking than I do, so perhaps you can help me understand the difference between savings and investment. From where I stand, it seems as though if people save more, either by putting money in a savings account, or paying down debt, but not just stuffing money in the mattress, then that should increase the ability of banks to lend and keep the economy moving. I understand Arnold's recalculation story as basically saying that even though the money is out there, banks and investors in general are having trouble identifying profitable lending opportunities.

Vangel writes:

What mackerel and money have in common is that they are both stores of value. What is different about them is that money is a standard unit of account, but mackerel are not.

There is a huge difference that you are glossing over. Mackerel are real and to get a bigger supply you are going to have to find ways to catch more of them. That takes time because you will need new capital and labour dedicated to the task. In the case of fiat money no such efforts are required and the money producers can create more of it at will at no extra cost.

But people want money as a convenient store of value (and medium of exchange).

But they don't seem to. If they did they would reject the entire idea of fiat money, which has never been a 'convenient store' of long term purchasing power. Fiat money is only convenient to the state and the banking system.

So, if the government is behaving itself, the prices of most goods and services in monetary terms do not change very much from day to day.

I do not believe that is true. It seems obvious that government can pursue inflationary policies but day to day prices can appear to remain fairly stable. In a society where productivity is rising prices should decline as technological innovation and capital investments lower costs. But those natural declines can be reduced or eliminated by inflation. That would allow governments and the banking system to be the primary beneficiaries of productivity increases. Individuals miss the point because they do not see what would have happened in the absence of the inflation.

If the government is not behaving itself, and it is trying to fund a massive deficit by printing money, then we get the phenomenon of too much money chasing too few goods. Money starts to lose its appeal as a store of value. This is hyperinflation.

Once again, I have trouble with the words that you use. It is clear that one can have high inflation that does not rise to the level of hyperinflation for some time, particularly if there are other pressures in the system. It is possible for people to anticipate actions to head off higher inflation and for bond investors to look for the next Volker to come along and allow them to become much richer. These type of beliefs, and the actions that they drive, could allow inflation rates to rise slower than they would have if individuals could shed their bias and hopes and see reality as it is.

However, if we go back to the case where the government is behaving, then the quantity of money is more like the quantity of mackerel. It is just one more quantity out there. People who go to their farms, factories, and offices every day do not care about how much money is in circulation, any more than they care about how many mackerel have been caught this week. Ordinary businesses set prices and make decisions based on what they saw in their markets yesterday and what they expect to see there tomorrow.

You are right to argue that perceptions matter but perhaps naive to expect government to behave itself and not siphon off the benefits of productivity increases that are hidden by its inflationary policies.

You are also not looking at the political incentives during times of crisis. When bubbles pop and there are dislocations in the economy, ruling parties and their opposition will look around and try to score the most political points. When the number of debtors is much greater than the number of savers, or when the debt is mainly in the hands of foreigners, there is little incentive for politicians to look after the purchasing power of the currency. Instead of being voted out of office as tight money policies force a wave of liquidations that allow foreign creditors to buy cheap assets from American voters political parties will choose money printing, subsidies and tax rebates. While such actions can lead to hyperinflation later, political parties are more interested in the next election cycle. Their own short term interests are much more important than the long term state of the economy or the long term changes to the standard of living for individual taxpayers.

The Federal Reserve can manipulate some market interest rates by printing money and using it to buy securities. If it buys massive quantities of mortgage securities by printing dollar bills, it can lower the interest rate on those securities. In theory, the Fed could do this by exchanging mackerel for mortgage securities just as well as by exchanging money for mortgage securities.

This is not true. The Fed can create as much fiat money as it wishes but is limited as to the amount of mackerel it can get its hands on. You seem to have a very large blind spot because you have not looked at the implications of fiat money closely enough. Fiat money is make believe but you are still thinking of it as you did when it was actually backed by something tangible.

However, I am not sure how long they can keep it up. At some point, market participants will think, "The Fed is not going to keep buying these things forever. When they stop buying, the price will fall." And such thinking will dampen prices (raise interest rates) today. The longer the Fed fights the market, and the more government keeps spending on other stuff, the greater the likelihood that too much will be financed by printing money, and we will have hyperinflation. As long as we know that the Fed is afraid of hyperinflation, then it has a limited supply of money, just as it would have a limited supply of mackerel.

I have gone on long enough so I will end this here. But before I go let me note that you are looking at the subject from a very American perspective. I am not so sure that the perspective is not blinding you to possibilities. So let me give you some family experiences that may open your eyes to what the rest of the world has seen.

My grandfather had a mill in northern Greece. It was acquired by savings earned in the US by his father, who went looking for an opportunity to make money and improve his family's standard of living. The mill was very successful and was a good source of earnings. One day my grandfather woke up to find that the old currency was declared worthless by the government. While the family did not suffer much because he did not trust the government and had gold instead many people did.

A few years after my grandfather's adventure by wife's grandfather was relatively happy as a well off merchant just outside Xi'an, China's ancient capital during the Qin and Tang dynasties. Like my own grandfather, he woke up to find the old currency declared worthless by Mao, who had just come to power. Around the same time, my own grandfather had to leave his land behind because of a civil war that took one of his sons and threatened him and his family. Luckily, he had some gold to allow him to start a new life.

I remember as a small child going to a news stand to buy a comic. I gave the man 25 dinars but he rejected it. It seems that the government had introduced a new one that was worth 10 of the old ones and my savings were not worth much any longer. The lesson was clear. Spend money as soon as you get it and save very little only for emergencies that did not involve currency crises.

A few years later my uncle wakes up to find the Yugoslav government has introduced a new dinar worth 1,000,000 old ones. He takes some Canadian dollars that I sent him and promptly paid off the loan he took to build an extension to his house. In hindsight he acted too quickly because the government replaced the new dinar with a new new dinar a few weeks later at the rate of 1,000,000 to 1, and repeated the process one more time a few weeks later. Finally, it threw in the towel and used the German mark as its currency.

A few years after that incident I was in the Chiang Mai airport exchanging my USDs for baht. I told the girl at the counter that she must have made a mistake because I got almost twice as many baht per dollar as I did about three months previously. Because I was busy in China, I never noticed the true effects of the Asian crisis. Sadly, the currency devaluation did not really lower my own costs by very much because prices for the type of goods that my wife and I consumed in Thailand did not change that much. The local savers were certainly harmed as prices for many of the imports that they were consuming clearly increased substantially and local farmers and producers were force to increase prices in nominal terms.

I think that the relative stability that you have enjoyed in the US makes it harder for you to see the true vulnerability of the monetary system. While I do not discount to see some major moves to the upside for technical reasons, the long term trend for the purchasing power of the USD is down. I would not be surprised to see the currency be devalued substantially either by the government or by the markets to levels far below current levels. As such I will follow the advice of my grandfather and invest in real money that can't be created out of thin air by the Fed or my own central bank. Unlike you, I am quite certain that when the chips are down the central banks and national governments will do what they must to save state and local governments and indebted citizens from their foolish actions.

Devin Finbarr writes:

Winterspeak-

Don't get mad at me

The internet is an autistic medium. I like to argue forcefully and directly. Imagine I'm saying everything with a smile on my face.

You can use EM if you want, but you run into the same problem -- one person's expenditure is another person's income.


I understand very well that the private sector cannot save as whole in fixed nominal terms. Let's assume for a second that you're right, and that the private sector desires a fixed quantity of savings. All the paradox of thrift proves is that they cannot do it. Their attempt to do so will be futile. A falling V will result in a falling PY. But it does not tell you whether the fall in V will result in a lower P or lower Y. Accounting doesn't get you there. By accounting, either a lower P or lower Y can work. In order to determine whether P or Y falls, you have to look at the real world frictional forces.

For example, in response to its collapsed paper wealth, Harvard decided it needed to cut payroll expenses. Harvard can either cut wages across the board by 2%. Or it can reduce staff by 2%. Which does it do? Reduce P or Y? A priori there is no way to know. The MV=PY equation does not tell you. Accounting does not tell you. All accounting tells you is that if everyone in the economy attempts to cuts its spending like Harvard, the attempt to increase cash balances will be futile. But it doesn't tell you whether the attempt will result in lower P or lower Y. In reality, Harvard decided to reduce head count. Why? Simply because of the structure of power within Harvard - professors and administration had a lot more power to prevent wage cuts, than the cut workers had power to preserve their job. In other words, wages were sticky.

If you can find me someone who says "I want an expenditure ratio of x" vs "I want to pay down my mortgage of $300,000" I will change my position, promise.

I can give a huge number of examples. Think of a new retiree meeting with their retirement planner. They plan out an expenditure ratio based on their life expectancy. If their stocks fall, their demand for cash will increase until their balances cover their estimated future expenses. They want a ratio.

I personally think the same way. I think in terms of how much I need to save to buy a house some day, or go on that next vacation, or take a year sabbatical. My savings desires are a function of future planned expenditures.

Universities think explicitly in terms of ratios - they aim to spend ~5% of their endowment a year.

I work at a venture backed startup company. Our dissavings is also a percentage - we want to make the money last X years.

The person holding a mortgage does need to save in absolute terms. But on the other side of that mortgage is me, holding a bunch of CD's. When prices fall, I can increase my real goods consumption, and dissave, because my anticipated expenditures have fallen. So fixed nominal debt results in borrowers saving and creditors dissaving.

To the extent that our financial system does not actually match borrowers and lenders, well, things get a lot more complicated. But my first point would still stand - even if you assume a desire for a fixed nominal amount of savings, you still don't know whether the response to increased savings is falling P or falling Y.

As you say, my story of why P does not fall in MV=Py fits the real world more closely than yours. At this point, I think you're saying "Pigou would be right -- if it wasn't for friction!"

No, my theory of friction fits the real world better because I am deriving it entirely from look at real world forces. My point about the Pigou affect was simply to show that you couldn't get the results you were trying to prove simply via accounting or logic. Either P or Y can fall in response to an increased demand for money. You have to look at the real world to figure out which actually does fall and why.

Personally, I don't like the "sticky wages" Keynesian argument for why P does not fall, which is why I don't make it here.

But it's empirical reality - look at Harvard.

Fisher's debt deflation nicely explains the fall in AD. But it doesn't explain why falling AD leads to lower goods production rather than lower prices.


winterspeak writes:

BEN/TODD: I don't want to hog the comments section, and if I submit a link it will get eaten in moderation. Read "Guest Post: If Credit is Not Created Out of Excess Reserves, What Does That Mean?" at Naked Capitalism for details on how bank lending works.

In short: banks make loans by expanding both sides of their balance sheets, not by loaning out deposits. A bank makes a loan, creating a receivable as an asset, and drawing down its reserve account (also an asset). That loan is then deposited somewhere else in the banking system, creating a deposit (liability) and crediting that banks reserve account (asset) by exactly the same amount as was originally drawn down.

At a system level, the debit and credit to the reserve accounts cancel each other out, and you're left with a receivable (asset) and a deposit (liability). Making the loan (receivable) created the deposit.

Operationally, the bank that ends the day short reserves borrows what it needs overnight from the bank that has excess reserves. The Fed intervenes heavily in this market, because the interest rate banks charge each other is the Federal Funds Rate, and this overnight inter-bank reserve lending operation is the (convoluted) way that the Fed sets that rate. So, banks are not reserve constrained in any way in their lending.

thruth writes:

winterspeak writes: "I don't want to hog the comments section, and if I submit a link it will get eaten in moderation. Read "Guest Post: If Credit is Not Created Out of Excess Reserves, What Does That Mean?" at Naked Capitalism for details on how bank lending works."

I think that piece is a little misleading. The basic argument is: loans create deposits. This is mechanically true, but the reserve requirement still bites in the end. Each time a bank creates a new loan, they have to hold a little in reserve against the corresponding deposit. This is textbook stuff, see the wikipedia entry on "Money creation"

I think the more relevant statement in the context of this discussion is that an exogenous increase in deposits doesn't necessarily result in new loans. For example, in addition to reserve constraints, banks might be capital constrained (for a variety of reasons). If the capital constraints are binding, the easing of the reserve constraint that results from an infusion of deposits may have little impact on banks desire to lend.

Of course, Winterspeak hasn't really specified what he has in mind that's preventing the deposits from stimulating lending, so I'm just speculating...

winterspeak writes:

thruth: No, reserves never bite. Banks make the loans they want, and then borrow the reserves they need. Wikipedia, and most textbooks, get this wrong and repeat the "money multiplier" myth. Canada has reserve requirements at zero. And if comparing current m0 to m1, m2 etc. doesn't convince you that the money multiplier is bogus, nothing will.

Capital requirements, which are totally different from reserve requirements, absolutely constrain bank lending.

thruth writes:

winterspeak wrote: "No, reserves never bite. Banks make the loans they want, and then borrow the reserves they need."

I wasn't claiming that reserves matter empirically, rather that they do at least in principal put an upper bound on credit creation. The article you suggested certainly isn't clear about this (which is obvious from some of the comments).

"Capital requirements, which are totally different from reserve requirements, absolutely constrain bank lending."

Which is the point you should have been making all along. What still needs to be fleshed out is why the banks can't raise capital to fund new lending in the face of the precautionary demand shock. Saying "new lending is unprofitable" is not a satisfying answer as there is always a price at which a loan becomes profitable. One plausible explanation is the debt overhang story, in which hoarding reserves makes more sense than raising new capital for a firm on the brink of insolvency.

Joe Calhoun writes:

Arnold,

I've also been following Scott Sumner's blog and I think that the problem with his theory is that he doesn't seem to care much about the quality of GDP, nominal or real. I have no idea if the Fed could have created enough money to keep nominal GDP rising, but it seems to me that even if they could it would have just created another situation like the housing bubble where "investment" didn't yield anything productive.

As for the rest of your argument, I am of the opinion that the recalculation could have proceeded without the total collapse of last fall if Bernanke, Paulson and Bush had reacted differently. I see what happened last fall as essentially a massive and sudden case of regime uncertainty. After a year of bailouts by the Fed and Treasury the sudden change in course on Lehman created so much uncertainty that the entire country kind of stopped to see what would come next. I also believe that as it became obvious that Obama would be elected a further set of uncertainties were introduced. Absent the uncertainties, which were created primarily by the past interventions in my opinion, the recalculation could have proceeded relatively smoothly. Yes, we might have had a recession or we might have had a few years of slow growth with rising unemployment, but I see no reason to believe we needed the full stop we got instead.


winterspeak writes:

thruth: Reserves do not matter empirically, operationally, or in principal. They do not put any upper bound on credit creation, as credit creation creates reserves. I've gone through the accounting for this in a previous post.

This is a very expensive and foolish myth that needs to be put to rest.

Banks have cut back on lending because 1) their capital has eroded and they cannot lend as much, 2) new capital is expensive in this environment, 3) weak economy means people are less likely to pay loans back, and 4) private sector just isn't as interested in taking on more debt right now as it ties to save. Banking is pro-cyclical by its very structure.

If you don't understand the role reserves play (and don't play) in banking, you cannot understand how banks work.

thruth writes:

winterspeak wrote "I've gone through the accounting for this in a previous post."

Please point me to the relevant post. I'll come back to your other points later.

winterspeak writes:

thruth: Look at the post on this thread dated "Posted September 22, 2009 10:53 AM". Draw out t-tables and go through the transactions.

thruth writes:

winterspeak wrote: "Look at the post on this thread dated..."

I think the communication problem we have here is the money multiplier is an equilibrium concept and you aren't treating it that way.

To illustrate, imagine economy with 2 banks that have identical balance sheets:

Assets
$90 loans
$10 reserves
$5 other assets

Liabilities
$100 deposits
$5 equity

If RR is 10% of deposits, neither bank can do any additional lending on their existing deposits, because they won't have reserves to cover the new loans (even by borrowing reserves, since neither has excess reserves to lend the other - one way around it is to raise capital, but that's a different story)

From what I can tell, you are presuming banks can simply set their deposits at whatever level they desire, but clearly at given equilibrium prices there is a fixed supply of deposits (and likewise demand for loans). It is in this equilibrium context that the RR is binding. A truly exogenous positive shock to deposits would alleviate the constraint and allow marginal lending as the economy shifts to a new equilibrium.

winterspeak writes:

thruth: By law, banks need to have the reserves at the required level overnight, not every second of every day.

So, in your example, bank A makes a loan ($1) that is then deposited in bank B. At COB, their balance sheets look like this:

Bank A:
Assets
$91 loans
$9 reserves
$5 other assets

Liabilities
$100 deposits
$5 equity

Bank B:
Assets
$90 loans
$11 reserves
$5 other assets

Liabilities
$101 deposits
$5 equity

At 10%, bank A is short reserves (needs $10.1, and it has $9) while bank B is long reserves (needs $10, has $11).

Overnight, bank B will lend bank A it's extra $1 in reserves, so bank B is back at its target, and bank A is now short just $0.1 instead of $1.1 as it was initially.

Since the banking system as a whole is now short of reserves, bank A will get the extra $0.1 it needs at the discount window. This is what the banking system must do if it is short, as a system, of reserves.

More likely, the Fed will intervene in the overnight market and buy/sell/repo treasuries (which make up the "other assts" that banks hold) to adjust reserves that way. When the Fed wants to drain extra reserves from the system it issues treasuries, and banks move money from their reserve accounts to their treasury account (all at the Fed) the same way you move money from the checking account to your savings account. So, the extra $0.1 reserve will either come from "other assets" or from the discount window, depending on how the Fed decides to square the books.

When the banks open again the next morning, bank A will pay back its overnight loan from bank B (and the interest rate it pays IS the Federal Funds Rate, that's the whole point of this). And it will continue to make loans subject to capital constraints, and ignore reserve requirements.

Bank lending is constrained by capital requirements on the supply side, and desire for loans from good credit risks on the demand side. Reserve requirements are just how the Fed decides to set short term interest rates.

The money multiplier is a myth, as reality has been demonstrating ever since the Treasury slugged $800B of reserves into the banking system a year ago, and has watched it slosh around ever since. But as I said, if m0, m1, m2 etc does not convince you, and accounting does not convince you, and the Fed's own operational details (which are well documented) don't convince you, then I'm not going to convince you either!

thruth writes:

"Bank lending is constrained by capital requirements on the supply side, and desire for loans from good credit risks on the demand side. Reserve requirements are just how the Fed decides to set short term interest rates."

which impacts the opportunity cost of lending. hence, RRs matter. QED.

"The money multiplier is a myth, as reality has been demonstrating ever since the Treasury slugged $800B of reserves into the banking system a year ago, and has watched it slosh around ever since. But as I said, if m0, m1, m2 etc does not convince you, and accounting does not convince you, and the Fed's own operational details (which are well documented) don't convince you, then I'm not going to convince you either!"

the relevance of this is:
1. money multipliers are far less meaningful than they used to be due to a move to capital based regulation
2. we're in a liquidity trap -- neither reserve nor TARP capital injections have done much good. (Don't ignore that there's now interest on those reserves)

winterspeak writes:

"which impacts the opportunity cost of lending. hence, RRs matter. QED."

RR is the mechanism through which the Fed sets short term interest rates. They could do it with a reserve requirement of zero, as the Canadians do. Or they could do away with RR altogether like Sweden, New Zealand, and Australia and set interest rates through any number of other mechanisms.

I never said banks don't care about interest rates, but they don't care about the implementation mechanism. I said, and it's a fact, that banks lending is not reserve constrained. A bank's ability to make a loan is in no way impacted by whether it is long, or short, reserves. Ask anyone who manages reserves at a bank.

Banks ask "do I have the capital to make this loan?", and "can I make a profit on this loan?" before making a loan. They never ask "do I have the reserves to make this loan" because it does not matter.

I'm done with this. Thanks for the chat.

kharris writes:

OK, I hate to say this but, isn't the notion that PV is on its own path and that manipulating M forces V to respond ignoring a lot?

Great Recalculation? Sure, call it whatever you want. The thing is, part of whatever you call it was a perturbation in the financial sector. It may be that policy induced changes in M force offsetting responses in V when the economy is running along as it normally does. However, having a large number of financial links come unlinked at the same time is, in this equation, represented by V falling. That is not a response to M. If V falls because we have done a bunch of silly things in the financial sector, then the Fed can aim at keeping MV stable by increasing M.

Now, as I understand these things, M is not so likely to respond automatically to changes in V. Whatever may have happened independently on the PY side of the equation, any drop in V greater than required to accommodate the independent drop in PY would lead to a further drop in PY, unless there is a policy decision to increase M.

In short, what I think I see here is an effort to ignore the largest bout of financial turmoil in decades, just waving PY around and pretending the collapse of financial intermediation was a reflection of that, instead of an independent source of influence in the economy. Policy makers shoved M up because V was down. V did not fall because policy makers shoved M up.

Comments for this entry have been closed
Return to top