...there is a crucial issue which explains much of the enormous difference of opinion between critics and supporters of the Fed’s current policy. Critics such as me and Allan Meltzer ... argue that monetary policy should focus on a clear strategy for the instruments of policy. A goal for inflation or other measures of macro performance is not enough if it is simply part of a whatever-it-takes approach to the instruments. Such an approach results in highly discretionary and unpredictable changes in policy instruments with unintended adverse consequences, as we have been seeing in recent years.I have always had this problem with nominal GDP targets, inflation targets, and so forth. Ok, the Fed adopts your target. Now what? If nominal GDP doesn't do what the Fed wants it to do, what should the Fed do about it? Talk more? (Monetary policy is starting to look more and more like foreign policy here).
Supporters such as Adam Posen... are just fine with the Fed using, even year after year, a whatever-it-takes approach to the instruments of policy as long as there is an overall goal. With such a goal in mind, so their argument goes, the central bank can and should always intervene in any market, by any amount, over any time frame, with any instrument or program (old or new), and with little concern for unintended consequences in the long run or collateral damage in the short run (say on certain groups of people or markets) as long as it furthers that goal.
Critics are very concerned about those unintended consequences and collateral damage; they are also concerned about an independent government agency wielding such a great deal of power as it carries out a year-after-year whatever-it-takes approach. Supporters are much less concerned.
Taylor points out a deeper danger. The Fed's "mandate," the list of its "goals," keeps expanding. Beyond just inflation and unemployment, now the Fed is in charge of "financial stability," managing "systemic risks," the health of specific markets (mortgages, exports), the health of specific institutions (too big to fail banks), the diagnosis and pricking of bubbles (when not the deliberate stoking of such bubbles), management of the details of every part of financial system (how swaps get traded, for example) and surely coming soon a federal anti-crabgrass mandate. The list of things the Fed can do in pursuit of these goals is getting bigger and bigger too, while the power of its conventional instruments (setting short rates, quantiative easing) is diminishing. If the Fed doesn't think banks are lending enough, and to the right people, in pursuit of one of its many goals, what stops them from using their regulatory power to just go tell the banks who they should lend to?
We have told the Fed to attain unattainable goals, and given it great power to do "whatever it takes" in their pursuit. The Fed seems to go along. It's fun to be given so much power, in the short run at least. But in a democracy, the price of great independence must be limited power, and the Fed will soon have to choose. Congress already limited some of the Fed's powers after some "whatever it takes" of the financial crisis.
Taylor, of course, would like the Fed limited to the instrument of short-term rates, and to follow the Taylor "rule" for setting them. But the principle is larger than that instance.
Here is the Taylor rule:
ReplyDeletehttp://en.wikipedia.org/wiki/Taylor_rule
Short Term Interest Rate = Inflation Rate + Equilibrium Real Interest Rate + Multiplier #1 * ( Inflation Rate - Desired Inflation Rate ) + Multiplier #2 * ( Log (Real GDP) - Log (Real Potential GDP) )
Inflation Rate = -2%
Desired Inflation Rate = 2%
What should short term interest rate be assuming Real GDP = Real Potential GDP?
“Taylor points out a deeper danger. The Fed's ‘mandate,’ the list of its ‘goals,’ keeps expanding. Beyond just inflation and unemployment, now the Fed is in charge of ‘financial stability,’ managing ‘systemic risks,’…”
ReplyDeleteMaybe Dr. Taylor can explain how 2008-style financial instability (e.g., the risk of no bank being able to trust a counterparty) and systemic risks (such as a total lock-up of the commercial paper market) that would not prevent the Fed from effectively meeting its twin (actually, three) mandates.
Or perhaps he has secret knowledge, of which the Fed is unaware, that these risks have vanished into thin air, thanks to … well, whatever, but NOT what the Fed actually did.
He should share these deeper insights with us. As it is, this complaint comes across as carping that the Fed should have ignored its basic oversight responsibilities and let the economy sink all the way into a huge depression, because some of the Fed's ad hoc actions didn't directly reduce inflation or have an immediate cause-and-effect role on employment.
I don't think anyone is blaming the FED for trying (in the immediate article chain)...but you seem to admit that some of the FED's actions were pointless.
DeleteIf you are ok with that continuing then you are ok with an unelected body of government wasting society's time and money as long as they are afraid of instability that they are not even required to define.
How is the Fed setting short-term rates through open market purchases of bonds different from setting NGDP through open market purchases?
ReplyDeleteAnyway, with a change in instrument, NGDP can fit into Taylor's preferred type of policy. The policy instrument can be M2, and the desired setting for M2 can be desired NGDP/velocity. The central bank then sets M2 based on an estimate of velocity. The Taylor rule can be restructured as M2 is set (or high-powered money) as a function of how much NGDP differs from the target, with the multiplier on NGDP as a function of velocity. This is in fact a simpler rule than setting the short-term rate as a function of how much inflation differs from its target and unemployment from its target.
Yes and that makes the same bad assumption that Friedman made - constant velocity.
Deletehttp://research.stlouisfed.org/fred2/series/M2V
Does that look constant to you?
In fact since the Fed started cutting interest rates in 2008, M2 velocity has fallen for 20 straight quarters. I wonder why?
This comment has been removed by the author.
DeleteI did not assume constant velocity. The Fed can make an estimate of what velocity will be based on by how much they are increasing M2 and other potential changes in the economy.
DeleteI appreciate that we don't like the idea of the Fed simply going on an asset purchasing spree until it happens to hit its NGDP target. However, there is an equation telling us by how much M2 must rise to hit an NGDP target just as there is an approximate Taylor rule for how to set the short-term rate in response to deviations in inflation and output. We just don't know the equation yet. If we are afraid of the Fed having too much power/discretion, then we can just create rules that limit the acceleration of M2 expansion to prevent the Fed from purchasing the entire planet in one quarter.
DeleteJoshua,
Delete"The Fed can make an estimate of what velocity will be based on by how much they are increasing M2 and other potential changes in the economy."
The Fed operating on its own swaps currency for interest bearing bonds and back. In essence it trades liquidity for velocity and back. Any increase in liquidity by the Fed (increased M2 or other monetary aggregate) results in a decrease in velocity assuming all other factors are constant (liquidity preference, credit demand, etc.)
Are you claiming that monetary policy has no effect on NGDP? That goes against a lot of evidence. Even if the Fed can't influence RGDP since monetary shocks translate into price shocks, that doesn't mitigate the effect of monetary shocks on NGDP. In fact, if increases in M2 are perfectly offset by falls in velocity, then incredibly expansionary monetary policy will not be inflationary and the Fed has complete control over the real stock of money (as opposed to the nominal stock).
DeleteJoshua,
DeleteIn a simple credit model for an economy there are four factors:
1. Credit demand
2. Liquidity preference
3. Nominal Interest Rate
4. Productivity
You can write an equation for both nominal and real GDP using these four factors. The FOMC has limited control over one of them (nominal short term interest rate). I am not claiming that monetary policy has no effect, I am claiming that absent changes in other factors, lowering nominal interest rates will lower nominal GDP not raise it.
"In fact, if increases in M2 are perfectly offset by falls in velocity, then incredibly expansionary monetary policy will not be inflationary."
That would depend on productivity. If increases in M2 cause productivity to fall - for instance the federal government borrowing money and paying people not to work or worse to destroy productive enterprise, then inflation will likely rise.
What you do with the borrowed money matters as much as the interest rate the Fed charges on it.
Are you claiming that raising M2 causes velocity to fall by even more than the rise in M2? That is how expansionary monetary policy would cause a drop in NGDP. That certainly does not jive with the evidence.
DeleteAlso, monetary policy and fiscal policy need not be linked. We are talking about monetary policy. If the Fed buys assets to raise NGDP, that does not require the federal government to borrow more money. There is a large enough stock of assets in private hands to withstand any purchases the Fed is planning on making. Indeed, expansionary monetary policy often accompanies contractionary fiscal policy (that's the essence of monetary offset).
"Are you claiming that raising M2 causes velocity to fall by even more than the rise in M2? That is how expansionary monetary policy would cause a drop in NGDP."
DeleteI am saying that with unchanging liquidity preference and unchanging credit demand, a reduction in nominal interest rates will lead to a reduction in nominal GDP.
"If the Fed buys assets to raise NGDP, that does not require the federal government to borrow more money."
If you arguing the absurd, sure, if the Federal Reserve decides they should buy Ford's next 10 years of auto production then Nominal GDP would rise. Please look up what the Fed can buy, and you will see how monetary and fiscal policy are interlinked.
"Are you claiming that raising M2 causes velocity to fall by even more than the rise in M2? That is how expansionary monetary policy would cause a drop in NGDP."
DeleteHere is a little equation I came up with to help illustrate:
V = [ INT * ( 1 - LP ) + f'(t) ] / LP
Where:
V = Velocity of Money
INT = Nominal Interest Rate
LP = Liquidity Preference
D = Debt = exp ( f(t) ), f(t) is credit demand as a function of time
Nominal GDP = D * V = D * [ INT * ( 1 - LP ) + f'(t) ] / LP
Leave debt level, liquidity preference, and credit demand constant and lower the interest rate - what happens? Nominal GDP falls. Swapping currency for existing debt lowers the interest rate which lowers the velocity of money.
First, I believe your argument that connects fiscal policy with monetary policy assumes that fiscal policy responds to interest rates. I'm sure there is some truth to that. But, in the current political climate, I'm not sure that the deficit would be any different if interest rates were slightly higher (of course if they were much much higher, the federal government might rein in deficits quickly).
DeleteCan you justify your velocity equation?
I believe there is a mistake in your NGDP equation. NGDP is some monetary aggregate (M2) multiplied by velocity. Let's suppose your velocity equation is correct. The Fed raises M2 by buying debt. This also lowers interest rates (although not by very much in this economic climate). Why do you assume that the fall in velocity caused by the fall in interest rates is enough to offset the rise in M2? How is the Fed lowering the interest rate while leaving D constant? The Fed does not set an interest rate like a price ceiling. Rather, it engages in asset transactions until the interest rate moves.
Joshua,
DeleteFirst, I accept that money ultimately begins as a debt - whether it's yours, mine, or the federal government's. Hence in the equation of exchange, I replace M (Money) with D (Debt). The federal reserve does not print money out of thin air, they (or another bank) lends it into existence.
Simple accounting should explain it clearly. Bank A makes loan to Joe. Joe gets currency and spends currency, bank holds record of loan. FOMC comes along and buys loan from Bank A. Which "money" should count if figuring GDP - the money loaned to Joe, the money residing at Bank A that the FOMC gave to bank A for the loan, or both? The only money that will have an effect on GDP is the money borrowed by Joe. For the FOMC money to have an affect, Bank A will need to make another loan.
Second, this is just a model. I left a lot of things out of the model, for instance equity financing, dividend payments, taxes, differential interest rates (long term, short term, credit risk adjusted), charity, foreign trade, currency movements, etc.
"Can you justify your velocity equation?"
See the bottom of this post for the complete derivation:
http://johnhcochrane.blogspot.com/2013/04/debt-and-growth-in-10-minutes.html#comment-form
Or if you don't want to go through the math, can we agree that interest rates and money velocity are positively correlated? Can we agree that liquidity preference and money velocity are negatively correlated? Can we agree that the derivative of credit demand and money velocity are positively correlated?
"Why do you assume that the fall in velocity caused by the fall in interest rates is enough to offset the rise in M2?"
I don't assume that is the case, I thought I demonstrated (in this simple model) why that is the case.
Nominal GDP = Debt * [ Interest Rate * ( 1 - Liquidity Preference ) + Derivative of Credit Demand ) ] / Liquidity Preference
Set liquidity preference = 0.5, set derivative of credit demand = 0, set D = $50 trillion, and set the interest rate = 5%. Nominal GDP would be $2.5 trillion. Drop the interest rate to 4% leaving everything else constant and nominal GDP would be $2.0 trillion
"How is the Fed lowering the interest rate while leaving D constant?"
The bonds that they are buying are existing debt. They are reducing the supply of bonds that can be bought by the market for a given demand thus pushing up the price of the existing debt and pushing down the market interest rate on the debt.
I believe you are confused. You are suggesting that open market operations only affect NGDP through the interest rate's effect on velocity. Yet, NGDP = M*V and open market operations also affect M. Here is a nice overview: http://en.wikipedia.org/wiki/Equation_of_exchange
DeleteThat aside, your model suggests that fiscal stimulus will boost NGDP significantly. It increases debt and puts upward pressure on interest rates.
Joshua,
DeleteIf the federal reserve buys existing debt with money and that money is not lent out it will have no effect on nominal GDP at all. That is why I used NGDP = Debt * Velocity instead of Money * Velocity. Money sitting as reserves in a bank will have no effect on NGDP at all. It is only when that money is lent out that nominal GDP is affected.
And yes government expenditures can boost NGDP significantly (duh!!), unless those expenditures are used to reduce private debt or stuffed under a mattress. The question becomes how will government expenditures affect real GDP.
The Fed is the new central planner.
ReplyDelete“The plans differ; the planners are all alike…” - Frédéric Bastiat
The Fed’s open-ended mandate(s) are an example of a much bigger problem: CFPB, ACA, Dodd-Frank—pick your favorite from a very long list—and you see the handiwork of politicians who are fond of voting for desireable ends but who refuse to recognize the complexity of the tradeoffs involved. The political calculus is easy to understand but the results are regrettable.
ReplyDeleteI agree with most of that. I was a little puzzled by the last two sentences:
ReplyDelete"Taylor, of course, would like the Fed limited to the instrument of short-term rates, and to follow the Taylor "rule" for setting them. But the principle is larger than that instance."
Do you mean you're not on board with the Taylor rule? If so, I agree with that too. One has to go through some contortions to find a theoretical justification for the Taylor rule. In lieu of something better, a 2% inflation target seems as good as anything, and I certainly agree that the Fed's instrument for hitting the inflation target is the overnight rate - currently that's determined by the interest rate on reserves.
Stephen,
DeleteThe Taylor rule was an econometric attempt to model short term interest rate fluctuations instituted by the Alan Greenspan led FOMC. If you take that model and try to use it for the Volcker or William Martin led FOMC, it will be meaningless.
I think it is not so meaningless in Volcker error data. In a purely econometric viewpoint it is pretty easy to find a structural break in the time series indicating the start of a volcker-greenspan era.
DeleteThe taylor rule also plays a pretty big role in lots of theory so it wouldn't be meaningless to analyze other time periods.
of course estimating the coefficients of the taylor equation is a bag of worms in itself. and prof cochrane has lots to say about how fruitless that can be.
Scott Sumner disagrees that focusing on "instruments" results in less discretion. I think he makes the implicit assumption that any level of aggregate demand CAN be hit by the Fed, though it may result in some inflationary tradeoffs. I would chime in by noting that even interest rates aren't an "instrument" but a target. Except for interest paid on reserves, a horrible declaration of deflationary intent resulting from Bernanke's credit-driven (rather than monetary) view of the depression.
ReplyDeleteSumner does not believe that there are tradeoffs between NGDP (aggregate demand) and inflation. He argues that stable NGDP is ideal, not stable inflation. Indeed, he would want less than average inflation during a boom. If we have a large productivity boom and RGDP is accelerating, why would we push to continue to have 2% inflation each year?
DeleteSumner has a very good and simple answer to Cochrane's questions: Set your goals (X percent inflation and X percent employment growth, for example, or even better, a nominal GDP target) and set your policy, contingent on hitting the goals.
ReplyDeleteSuch as, the Fed does $85 billion a month in QE until targets are hit, racheting up by $20 billion a month until success.
Yes, you do more. There is nothing wrong with the doing more and more until targets are hit.
The risk today is not in inflation. That is 1970s style thinking, or the usual reactionary-gold nuttiness.
Prudence is not always conservatism. It is often a fig leaf for feeblemindedness. It does not take much mental effort to say. "Let's just keep doing what we are doing, or do nothing." Prudence can mask timidity, laziness, or an agenda to absolutely protect the status quo, and a raft of other sins.
We could have done nothing and stayed out of WWII. That would have been the prudent and safe thing to do. In fact it was the right-wing that argued for such conservatism.
BTW, I think fiscal austerity is a good idea. There should be radical cuts in federal outlays.
But the Fed has to pour it on, to get the economy going.
Inflation is dead, btw.
And has been dead in Japan since 1992. The results are not pretty. Anybody who thinks inflation is the great boogyman needs to explain Japan.
That is the risk today. That we become the United States of {Japan.
And you would think that those who love the free-market and are wary of the Fed proclaiming it knows what's best would like this idea. Economists at the Fed can set clear goals as to where the economy should be. Most agree that unemployment could and should be lower, and those who think NAIRU has gone up should be fine with lowering unemployment until inflation starts to accelerate. On the other hand, economists will disagree on what is necessary to get us there. So, the Fed expands its purchases each month as unemployment continues to remain too high and inflation too low. It is absurd to think that Bernanke has some magic formula he can apply to know exactly where interest rates should be to hit unemployment/inflation goals.
DeleteBenjamin why not just set a real GDP target? If the Fed can hit a nominal GDP target, why can't it hit a real GDP target?
DeleteFrank--
DeleteI think targeting the nominal allows the market some squishiness to adjust. BTW, a nominal target, in practice, will mean inflation can be no higher than the nominal target.
The last PCE inflation rate came in at 1.1 percent, y-o-y.
We are talking Japan-style inflation, historically low rates of inflation, sinking rates of inflation.
Not only is the Fed way below its target, it is way, way below what might be a reasonable rate on inflation for an economy that is still not generating real job growth.
You know, when Volcker was Volcker, he got inflation down to five percent, and the WSJ told him enough is enough, and he was regarded as a great inflation fighter, and Ronald Reagan was too.
Now, we are at 1.1 percent and sinking, and everyone is fretting about inflation.
Sometimes, I wonder. When does a concern about inflation become an fixation, then an obsession? Now. That is when.
The 1970s are over, and the structural impediments of the 1970s that helped inflation survive are gone (also Burns erroneously believed you could not fight inflation with tight money).
There are no more unions in the private sector, and international trade is robust, unlike the 1970s. Banks, rails, trucks, airlines, phones deregulated.
If we keep fighting inflation, we will only end up like Japan: Permanent and growing federal deficits, and extremely weak growth and frequent recessions. You cannot recession your way to prosperity.
The safe thing now is a very aggressive, growth-oriented monetary policy.
Joshua--
Maybe so, but whatever rate the Fed sets is artificial, whether it is 5 percent of 1 percent. Then you get into arguments about whether the Fed can set rates at all. Long term rates do what they do.
Right now, the risk is not being aggressive enough, not really blowing the doors open and print lots of money.
Japanification is the risk.
Ben,
Delete"Japanification is the risk."
Unemployment Rate in Japan - 3.7%
Oh yes 3.7% unemployment rate in the United State would be such a horrible thing. Do yourself a favor - ignore everything Paul Krugman has to say about anything.
Benjamin,
ReplyDelete"I think targeting the nominal allows the market some squishiness to adjust."
Just because the Federal Reserve declares a target (either nominal or real) does not mean the "market" has to acquiesce to that target. The "market" can squish around all it wants.
If instead you are talking about an enforced target (Federal Reserve tells "market" to get in line or get lost), then who gives a flying rat's butt what the "market" thinks or wants.
Frank-
ReplyDeleteI am a Market Monetarist---we do not like federal deficits, we like monetary stimulus.
But how is unemployment measured in Japan? And is their labor force growing or shrinking?
BTW, since 1992 and the deflation that defined Japan, stocks and property are off 80 percent, and real wages have fallen 15 percent. (Stocks recently rallied on the Abenomics, they may not be down so much).
For Japan, tight money has been a debacle, and the momentum has shifted to S Korea and China.
China, btw has benefitted from the People's Bank of China, that has a ceiling on inflation at 4 percent.
Another plus of moderate inflation: Pays down the national debt. If Japan had run 3 percent inflation for the art 20 years, their debt to GDP ratio would be one-half what it is today.
What makes 0 percent or 2 percent a divine rate of inflation? What if 3 percent or 4 percent is better for real long-term growth?
Benjamin,
Delete"I am a Market Monetarist---we do not like federal deficits, we like monetary stimulus."
http://en.wikipedia.org/wiki/Market_monetarism
Whose liabilities are Nominal GDP futures - the federal government's or the central bank's? If they are the central bank's they you run into several violations of the Federal Reserve Act.
"Another plus of moderate inflation: Pays down the national debt. If Japan had run 3 percent inflation for the art 20 years, their debt to GDP ratio would be one-half what it is today."
This is real simple. To reduce federal debt a government must either collect more in tax revenue than it spends or it must sell equity instead of debt. You are making a bad assumption that income levels rise with inflation.
Remember this statement you made: "There are no more unions in the private sector"
That also means there are no more COLA agreements which means income may not rise in lockstep with inflation which means that tax revenue may not rise in lockstep with inflation.
"What makes 0 percent or 2 percent a divine rate of inflation? What if 3 percent or 4 percent is better for real long-term growth?"
There is nothing that makes 0 or 2 percent a divine rate of inflation. It could be -1%, -3%, -5%, any number of values. Why do you think a positive rate of inflation is better for real long term growth?
Frank-
ReplyDeleteWhat market Monetarists believe (and Scott Sumner is a better proponent than I, and I encourage you to read his blog) is in money illusion (and I add, social norms).
It is considered punitive to cut a worker's pay. Bad juju. But to leave pay steady, in a time of mild inflation is okay.
This means wages can adjust to weak demand--in times of moderate inflation.
Some say workers will learn to get the COLAs. But as you point out, often workers do not have a choice. They have to grin and bear it.
So moderate inflation is good. It allows the labor market to adjust to changes in demand.
Inflation need not raise revenues to cut the federal debt. The debt relative to GDP shrinks if we have general price inflation. So, let's say we do not add to the debt, but the nominal GDP increases by 6 percent. We are thus reducing our debt to GDP ratio.
Cochrane is very worried about mounting federal debt. But, with inflation and some monetization of the debt, that problem is easily solved.
The Fed need not ever sell bonds it bought through QE. It can just hold to maturity and send proceeds to Treasury.
I refer the question on a NGDP futures market to Scott Sumner. For me, it is not necessary. An aggressive, pro-grwoth central bank is all that is needed.
BTW the People's Bank of China has a ceiling of four percent inflation. Works out great.
Ben,
ReplyDelete"Inflation need not raise revenues to cut the federal debt. The debt relative to GDP shrinks if we have general price inflation. So, let's say we do not add to the debt, but the nominal GDP increases by 6 percent. We are thus reducing our debt to GDP ratio."
I thought we were talking about reducing the debt itself not the debt to GDP ratio. And raising the price level (inflation) may not even raise nominal GDP if the inflation imported. I realize you are talking about general price inflation (all goods - imports and exports), but the Fed doesn't have control over what goods are resource constrained within the United States.
"Cochrane is very worried about mounting federal debt. But, with inflation and some monetization of the debt, that problem is easily solved."
ReplyDeleteCochrane doesn't understand why federal debt is "bad juju". See here for the thinking that dominates the Republican Party:
http://www.bloomberg.com/news/2013-05-08/boehner-accidentally-explains-why-his-deficit-position-is-phony.html
There are good reasons to reduce the federal debt and threat of bankruptcy or insolvency is not one of them.
Finally, here is the U. S. effort to "grow" its way out of debt since 1950:
http://research.stlouisfed.org/fred2/graph/fredgraph.pdf?&chart_type=line&graph_id=&category_id=&recession_bars=On&width=630&height=378&bgcolor=%23b3cde7&graph_bgcolor=%23ffffff&txtcolor=%23000000&ts=8&preserve_ratio=true&fo=ve&id=GDP_TCMDO&transformation=lin_lin&scale=Left&range=Custom&cosd=1950-01-01&coed=2013-01-01&line_color=%230000ff&link_values=&mark_type=NONE&mw=4&line_style=Solid&lw=1&vintage_date=2013-05-09_2013-05-09&revision_date=2013-05-09_2013-05-09&mma=0&nd=_&ost=&oet=&fml=a%2Fb&fq=Quarterly&fam=avg&fgst=lin
This includes all debt (public and private). And it been pretty much an abject failure (including the great inflation of the 1970's). The easiest way to reduce debt is to replace debt with equity.
I am fine with monetizing debt (the federal debt) or converting private debt into equity (following the bankruptcy of a corporation for example).
ReplyDeleteOther tactics might include elimination of the homeowner mortgage interest tax deduction and the elimination of the onetime forgivance on home-sale capita gains for people over age 55.
No taxes on dividends might help too.
Ben,
DeleteBut you are uncomfortable having the federal government sell equity because?