September 28, 2009
Guest Contribution: Lessons from the 1970s for Fed Policy Today
By David Papell
The Federal Open Market Committee voted last Wednesday to keep the federal funds target rate at a record low of between zero and 0.25 percent. If it was not constrained by the zero lower bound, should the federal funds rate be negative? If the answer is yes, this suggests that the rate should remain at its record low for a considerable period and provides a justification for continued increases in the Fed's balance sheet. If the answer is no, then the Fed may need to raise its interest rate target sooner rather than later.
There has been a lively debate on this topic in the context of the Taylor rule for monetary policy. The debate started with an article in the Financial Times, which cites a confidential Fed staff study that placed the implied Taylor rule rate at negative 5 percent. John Taylor, speaking at the Atlanta Fed, counters that the Fed got both the sign and the decimal point wrong and calculates the rate at 0.5 percent. Glenn Rudebusch, writing in the San Francisco Fed's Economic Letter, argues for negative 5 percent. Most recently, Taylor writes in a Bloomberg.com commentary that the rate should be zero.
In its original form, the Taylor rule states that the Fed's interest rate target should be one plus 1.5 times the inflation rate plus 0.5 times the output gap. According to the most recent Congressional Budget Office (CBO) projections, "core" CPI inflation, excluding food and energy, was 2.0 percent in 2008 but is expected to fall to 1.6 percent in 2009. The output gap is expected to be negative 7 percent in the second half of 2009 and the first half of 2010. Using 2.0 percent for inflation, the implied rate is 0.5 percent while, with 1.6 percent inflation, the rate is -0.1 percent, both around the Fed's current target.
The implied interest rate target can change if the Taylor rule is modified. It is often argued that, because monetary policy is forward-looking, policy evaluation with Taylor rules should be conducted using forecasts rather than actual data. The CBO forecasts that core inflation will fall further to 1.1 percent in 2010. Using this forecast, the implied rate falls to -0.85 percent. More dramatically, Rudebusch and others argue that the coefficient on the output gap in the Taylor rule should be 1.0 instead of 0.5. With inflation forecasts and the larger output gap coefficient, the implied Taylor rule interest rate falls to -4.35 percent.
In "Taylor Rules and the Great Inflation: Lessons from the 1970s for the Road Ahead for the Fed," written with Alex Nikolsko-Rzhevskyy of the University of Memphis, we use research on the natural rate of unemployment published in the 1970s to construct real-time output gap measures for the periods of peak unemployment in 1971 and 1975, and argue that real-time linear and quadratic detrended output gaps provide the best measure of what policymakers perceived the output gap to be at the time. Using these gaps to estimate Taylor rules, we conclude that:
- The Fed did not follow a stabilizing Taylor rule in the 1970s. In order for policy to be stabilizing, the federal funds rate needs to be raised more than point-for-point when inflation increases, so that the real interest rate rises. Our estimates never produce a coefficient on inflation that is significantly greater than one.
- The Fed did follow a stabilizing Taylor rule if inflation forecasts, rather than inflation rates, are used. These forecasts, however, systematically under-predicted inflation following the 1970s recessions and this does not constitute evidence of stabilizing policy.
- The Fed responded too strongly to negative output gaps, with estimated coefficients between 0.7 and 1.0.
In the 1970s, the Fed "stabilized" overly optimistic inflation forecasts and responded too strongly to output gaps, lowering interest rates too much -- especially during and following the 1970-1971 and 1974-1975 recessions, resulting in frequent recessions and the Great Inflation. What are the lessons from the 1970s for Fed policy today?
- The Fed should respond to inflation, not inflation forecasts, especially in an environment where large negative output gaps are causing forecasted inflation to fall.
- The Fed should not tinker with Taylor's output gap coefficient of 0.5.
Using the rule with Taylor's original coefficients, the experience of the 1970s suggests that, even if it could, the Fed should not lower its interest rate target below zero. If the incipient recovery takes hold and inflation stays the same or rises, it may need to raise rates sooner than many people think.
This post written by David Papell.
Posted by Menzie Chinn at September 28, 2009 06:31 PMdigg this | reddit
And then we have wackos that write stuff like:
What unprecedented anti-Fed days these have been! We had our Audit the Fed Congressional hearing, in which the central bank � for the first time in 96 years � was put on the defensive.
Since 1913, the Fed has had it all its own way: booms and busts, dollar depreciation, redistribution to the government and the big banks from the middle and working classes. But just as Andrew Jackson abolished the predecessor of the Fed, we too can knock over this dangerous institution. End The Fed teaches all the fascinating history, and tells us what we can do for the future. I tried to make it easy for everyone to understand, and convincing for those who already do. It gives the constitutional, economic, moral, and libertarian arguments against what Jackson called "the Monster."
Ever seen the Fed's marble palace in Washington, DC, on Constitution Avenue (of all streets!)? That bunch sure knows how to live. I've long had a dream of being the auctioneer when the Fed is sold off for private offices, or maybe a Museum of Sound Money! Help me dull its scissors and then break them, so the Fed can't cut down our dollar's value. Indeed, I believe that people ought to be ashamed to work at such a place; an institution that has done so much damage to American prosperity and freedom, as well as to the freedom and prosperity of the whole world. For example, I want no more bowing and scraping to the Fed chairman when he goes to Capitol Hill to peddle his nonsense. He is just a bureaucrat, albeit a disastrous one.
Together, you and I can change things. Indeed, we must. For all our futures, nothing is as important as cutting the Fed down to size. Join me: let's End the Fed.
Congressman Ron Paul
Posted by: walker o at September 28, 2009 07:55 PM
I'm pretty sure I don't want to go back on the gold standard now that the national debt is $11T, even if they do approve silver coinage finally.
However, I am also suspicious of new millennia output gap calculations. Just heard 2.5 million of our recent unemployed decided to call it quits and take early social security. Good for them. I think old people should retire.
Then we have heard how GM is restructuring for a total US auto market size of 12 million units rather than 15 million units.
We know residential construction is not ever going back to mid decade levels.
And all the non-economists I know don't believe that real estate agents are part of the permanent workforce.
Then there is the financial sector.
So can anyone tell me with a straight face that they can calculate what the output gap really is? Or should be? Or that low interest rates will somehow transform the economy into something new? Remember that Greenspan used that line on us already, and we are supposed to be smarter than that the second time.
Posted by: Cedric Regula at September 28, 2009 09:26 PM
In the US, the 1970s were also characterized by the presence of direct price and wage controls. It would appear that the US government and the Fed thought then that such measures could help control inflation while not having negative effects on output and employment. This could explain the overly optimistic views and forecasts prevailing at that time on inflation.
Note too that if we use forecasts of inflation in Taylor rules, so too should we use forecasts for potential output. But how clear are our concepts of this potential, and how good are our measures of (current... and future!) potential output?
Posted by: Patrick VB at September 29, 2009 03:24 AM
"The Fed did not follow a stabilizing Taylor rule in the 1970s. In order for policy to be stabilizing, the federal funds rate needs to be raised more than point-for-point when inflation increases, so that the real interest rate rises. Our estimates never produce a coefficient on inflation that is significantly greater than one."
I don't think that your first sentence is consistent with your third one. If you have point estimates greater than one (or plausibly greater than one), what justifies your confidence that the Fed did not follow a stabilizing Taylor Rule?
Posted by: Simon van Norden at September 29, 2009 04:58 AM
Output gaps estimates suffer from imprecision; this problem become more acute when we try to estimate the current gap, or forecast the gap.
Using the range of statistically plausible current output gap estimates, what is range of reasonable Fed funds rates?
Posted by: Simon van Norden at September 29, 2009 05:04 AM
If I understand correctly (and please correct me if I'm wrong) Mr Paul wants a gold system. And again correct me but under a gold system, the market adjusts the interest rate, and thus money supply, by raising/lowering the discount rate. Under the Fed, the Fed adjusts the money supply to adjust the interest rate.
I have read Prof Hamilton's (excellent) article on the gold standard with a takeaway that the problem with such a standard is that you can always go off it. I have also read his concerns regarding a currency crisis.
What I think everyone is concerned about is the value of our medium of exchange, which, on-topic re: inflation, is central to the above. Calling Mr. Paul a wacko (I have no idea) because he wishes to maintain the value of the dollar, and sees, by BLS' own inflation calculator a really large devaluation since the Fed began strikes me as an ad hominen attack.
Posted by: Rob at September 29, 2009 05:26 AM
I laugh as I read all the economists who have the nerve to tell John Taylor what his rule should be. But I guess this is not really that different from the economists so wedded to the demand theory that its inital failure in the Great Depression and its totally being discredited during the Great Inflation of the 1970s simply motivates the theorists to work harder to come up with innovative ways to maintain the illusion.
So what do the lessons teach us? They teach us that:
1. The FED is as political an institution as any other governmental or quasi-governmental entity and that they cannot and will not follow a rule even if they have one.
2. Even if they follow the rule their statistics and projections are so bad that the result is disastrous. In hind sight it is noted that "these forecasts, however, systematically under-predicted inflation...." Did the FED forecast any better in 2008 during the credit crisis? If so why did we have a credit crisis? Don't tell me the FED really can't control the economy? Oh, horror of horrors!
3. The FED responded too strongly to "negative output gaps." You mean they actually over-stimulated? I thought they were omnicient?
So what of the recommendations for today.
1. Certainly the FED should respond to actual circumstances rather than one of thousands of forecasts floating in econo-world, but what response? It appears the assumption is the respons of business-as-usual will give a different result. How about stabilizing the currency so it actually functions as a unit of account and a store of value, freeing up business assets that have been dedicated to hedging against monetary mistakes to be used for actual production?
Tinkering with the Taylor Rule simply gives those advocates of the Taylor Rule an excuse why it doesn't work. The Taylor Rule uses the same tired system of interest rates pushing on a string. The problem is not the rule but the ignorance of the monetary authorities to know what the indicators are saying or to even know what normal is. Can anyone tell me what normal is?
I give John Taylor credit for at least suggesting the FED exercise some discipline, but why does it have to be so complicated when a simple gold rule would solve all of the stability problems with the unit of account. Oh, but then we wouldn't have jobs at the FED and Treasury for all the thousands of demand model economists, oops.
Posted by: DickF at September 29, 2009 05:44 AM
Wacko....? Ron Paul? More like a sane man in an insane world. Riddle me this? Why is it, that we are paying interest to borrow money from ourselves? Answer, to line the pockets of Wall Street Financiers and pad the tax revenues of the state governments that are supported by them. End the fed now. AMEN.
Posted by: Glen at September 29, 2009 06:16 AM
"Ever seen the Fed's marble palace in Washington, DC, on Constitution Avenue (of all streets!)? That bunch sure knows how to live."
Ever been inside into the staff offices? Dingy doesn't begin to describe it. I've been inside a dozen central banks and regional feds; the FRB is dead last in terms of staff offices. (Okay, they tie with the Bank of England.) A colleague moved from there to a nice job with a European central bank; he remarked to me that the private bathroom off his new office was itself larger than his old office at the Fed.
Posted by: Simon van Norden at September 29, 2009 06:53 AM
DickF: great point about capital tied up in hedging programs. That's something I've been harping on for years. The basic problem with our economy is a lack of savings and investment and yet we have billions of dollars tied up in hedging interest rates and commodities that would be a lot less volatile with a stable dollar. How might that capital be put to use if we just stabilized the dollar?
Posted by: Joe Calhoun at September 29, 2009 08:17 AM
Simon van Norden makes two good points:
Our conclusion that the Fed did not follow a stabilizing Taylor rule in the 1970s is not a question of statistical significance. For the real-time linear and quadratic output gaps that we argue best represent research at the time, the estimates are not greater than one. I should have made this clearer.
As for current output gaps, quadratic detrending produces values close to the CBO estimates. HP and band pass filtering produce smaller estimates, which would imply higher Taylor rule interest rates.
Posted by: David Papell at September 29, 2009 08:52 AM
The Glenn Rudebusch estimated Taylor Rule represents how we have managed monetary policy for from 1988-2008. Dr. David Papell is suggesting that the results of his paper argue for a change in our current monetary policy, one that is in fact more inflation averse. I am very skeptical.
There are two issues here: 1) inflation forecasts and 2) output gaps. The period that Dr. Papell studies, 1966-1979, may not be the most analogous to our own in either of these respects.
Without reviewing all of the data, my general impression is that the Greenbook forecasts of inflation tend to be more inaccurate the higher the inflation rate is. Furthermore they tend to lag the actual inflation rate. More specifically when inflation is trending upward the inflation forecasts tend to be too low.
The period from 1966-1979 was one of generally accelerating inflation. Thus given the pattern I have claimed, one would naturally expect the inflation forecasts to be too low. Furthermore inflation was generally quite high for most of the period, in fact higher than we have experienced in the last 15 years or so. Casual inspection of the data reveals that the inflation forecast during this period were especially bad when inflation got above 5% and during periods of rapid change in inflation the lagging phenomenon is quite visible.
In our current circumstances inflation is now less than 2% and is probably declining. In other words whereas during the period Dr. Papell studied inflation was generally high and accelerating inflation in our near future is low and is likely deaccelerating. Thus I suspect our current inflation forecasts are likely much more accurate that those during the 1966-1979 period and furthermore I expect they are, if anything, too high.
The period from 1966-1979 is also notably different from other periods in terms of the output data. The economy was operating at above potential over 60% of the time. In fact at the beginning of the period it was 6.5% above potential. The largest output gap was 4.5% during the second quarter of 1975.
Christina Romer has stated that she believes that the output gap reached 7.5% in the second quarter of this year. This is far larger than at any time during the 1966-1979 period.
In short, I do not believe a period of high and accelerating inflation coupled with an economy that was operating above potential a majority of the time is relevant to our current situation. I do not believe the results of this paper justify a change in current monetary policy. More specifically I believe that it will be necessary to keep the federal funds rate at 0% for a significant period of time.
Posted by: Mark A. Sadowski at September 29, 2009 09:46 AM
Arthur Burns was not following the principle of Taylor Rule in 1971. He was following the principle of Nixon Rules. In other words, I don't think the performance of the Fed in 1971 should be very meaningful in evaluating today's policy.
Posted by: ISLM at September 29, 2009 09:55 AM
Mark A. Sadowski
"More specifically I believe that it will be necessary to keep the federal funds rate at 0% for a significant period of time."
I realize this is religion among economists (with the exception of ECB's Trichet), but I wonder if you could spell out a 2010 scenario of what this policy would accomplish. Or 2011, if nothing in 2010 is the scenario.
Also, I'm disallowing the answer "Nothing, ever, because it should be negative 5%". We don't want to cause banks to keep their money in their mattresses at home. Besides, we have enough of that happening as it is already.
Posted by: Cedric Regula at September 29, 2009 10:51 AM
So with a better gadget err normative Taylor Rule, we can solve the monetary policy issue? I think I see precedence in the old Soviet bureaucracy....
What if we ignored the "output gap", and just targeted price stability? What overnight rate would that version of the Taylor Rule predict should, as in ought to be, in place right now?
Posted by: GNP at September 29, 2009 11:00 AM
End the Fed!?
Then who would be able to print money infinitely to obscure the insolvency of multiple big bank & protect them from their mistakes? Who would manipulate interest rates continually so that market participants can never know the actual market interest rate...to wit, never know the true cost of capital. Who would serve as the Keynesian death panel for savers who want to actually earn a return on CDs?
What is Ron Paul thinking?
Posted by: bryce at September 29, 2009 11:14 AM
Since 1980 the Taylor rule widely used gives controlling inflation about double the weight of controlling unemployment or the output gap.
Prior to 1980 the Fed appeared to follow a Taylor rule that gave controlling inflation and unemployment about equal weight.
However, much of this divergence developed during the Volcker era. Under Greenspan the actual Fed funds rate was much closer to a rule that gave unemployment and inflation about the same weight.
If you run a regression with a dummy variable for the Volcker years the pre-Volcker and post-Volcker
era seem to be following roughly the same rules that is based on giving unemployment and inflation rougly equal weight.
Posted by: spencer at September 29, 2009 11:41 AM
We may well conclude that at the end, the Fed should use inflation expectations instead of inflation forecasts or actual inflation.
Even better, What about using alleatory inflation numbers?
Would it make any difference?
Posted by: IM at September 29, 2009 01:08 PM
It's hard to keep up with the conflicting statements on whether inflation or deflation is the biggest worry these days. Richard Fisher has said today that deflation is the biggest concern so I guess interest rates will remain as they are for a while?
Posted by: Senan at September 29, 2009 04:13 PM
The question is not what would one expect to accomplish by keeping ZIRP, the question is what would one expect to accomplish by not keeping ZIRP. Raising short term interest rates when the output gap is this high and inflation this low would be highly counterproductive. We need monetary policy to be more stimulative (if that were possible), not less.
Also, lending is not down because of ZIRP. ZIRP is necessary because lending is down. If you recall, the credit freeze preceded ZIRP by a few months, not the other way around.
In any case the Rudebusch estimated Federal Reserve Taylor Rule is :
FFR = 2.07 + 1.28 x Inflation - 1.95 x (Unem. - CBO natur.)
Based on this and FOMC forecasts I think it's safe to say that the Fed will keep ZIRP in effect at least through the end of 2012.
P.S. The interest rate the ECB pays on overnight deposits, which has taken on a more prominent role as the de facto floor for very short-term lending in the past year, has been 0.25% since April. Thus although the ECB headline rate may be 1% they are already essentially following a stealth ZIRP.
Posted by: Mark A. Sadowski at September 29, 2009 05:17 PM
Intuitively it seems that adding output gap to forecast inflation is double counting, since supposedly forecast inflation is based output gap and current inflation.
Posted by: HZ at September 29, 2009 05:18 PM
Mark A. Sadowski and HZ
Great point HZ. Wish I woulda thought of that, but let me also say current inflation is based on current output gap. We definitely have double accounting.
The other thing I believe is that low interest rates doesn't help employment, other than maybe keep too many bankers employed, since the only thing I see this doing is keeping alive an over sized financial sector thru corporate welfare. We have zombie banks feeding off a steep yield curve, and that is all that is happening.
But raising this point with economists or the administration and suggesting maybe rates should be 1% or 2% so they are REAL ZIRP rates, rather than nominal is kind of like going to an evangelist's prayer meeting and telling them that there is a plan that the Democrats are going to steal Christmas. You get about the same response.
The other thing we have learned about Fed policy is they like to lower rates fast and raise them slow. So when the zombies do come back to life the Fed has the problem of raising rates by a large amount, but wanting to do it slow so as to give financial markets time to adjust to the new normal.
So now I'll quote Mark:"The question is not what would one expect to accomplish by keeping ZIRP, the question is what would one expect to accomplish by not keeping ZIRP. Raising short term interest rates when the output gap is this high and inflation this low would be highly counterproductive. We need monetary policy to be more stimulative (if that were possible), not less."
Like I say, spoken like a true economist, but I gave my reason for higher rates.
Posted by: Cedric Regula at September 29, 2009 08:18 PM
Regardless of whether we should or should not "end the Fed", here's why Congress will not end the Fed:
The Fed has enormous powers to print money, control short-term interest rates, and pump liquidity into the economy. These powers are flexible, can be implemented quickly as economic conditions change, and without much congressional oversight.
Administration/fiscal powers to spend are less flexible - more long-term, subject to congressional approval/oversight, and delays of implementation. Through "Fed capture", the "Administration of the day" uses the Fed's powers to supplement its own policies. The Fed is theoretically independent, but Fed cooperation is easily obtained -the Fed demonstrates its usefulness, enhances its political standing with the "powers that be" , and the Fed can always find some plausible justification for their actions.
Both Democrats and Republicans have a vested interest in the status quo of the Fed, but for different reasons. Democrats like the spending power abilities of the Fed; Republicans protect the interests of financiers through - influencing Fed regulation of financial institutions, lobbing for low interest rates to bolster bank capital, and the availability of bailouts on demand. Both Democrats and Republicans view as intriguing, the Fed's powers to devalue the dollar by creating inflation .
Furthermore, I don't see any scenario on the immediate horizon that will change the dynamics of the current situation.
Posted by: fwm at September 30, 2009 06:22 AM
David Wessel, in his Wall Street Journal column today, argues that households are only beginning to reduce their debt. If this is true, then it is a bad idea for the Fed to raise interest rates anytime soon. Raising interest rates at the same time deleveraging is curtailing private spending is a recipe for stagnation.
Moreover, this type of monetary policy, concurrent with household deleveraging, would exacerbate the federal government's debt load .
Posted by: SMG at October 1, 2009 05:20 PM
Whether Fed Funds is at zero, or probably anywhere up to 2% has zero impact on existing credit card debt.
It also has no affect on any other existing fixed rate debt. And it is possible you get a flattening of the yield curve when inflation expectations are contained. I'm not saying they aren't still, but that can change quickly once the economy looks like it is on the path to recovery.
If it impacts Libor or other benchmarks that are used to re-calc adjustable debt, it may have an impact there. But here I would rather see the spread over Libor charged by the banks negotiated down. Or principal reduced. Many of these are the questionable mortgages that were written with predatory terms. Why not go after the predators instead of putting the entire economy at risk from monetary policy with a difficult exit strategy. If borrowers are faced with a reset that will result in foreclosure, the lender may get a little more reasonable.
In commercial real estate the Fed is working on "lender of last resort" plans again.
The Fed is also saying they may set up interest paying Fed CDs so banks can invest at the Fed. Banks also have been buying lots of Treasuries. See anything wrong with this picture? You can be in the biz of collecting near zero deposits from customers, and turn around and buy Fed CDs and Treasuries and make a little interest. Shouldn't the USG just offer that deal to the public and cut out the middle man?
The current wisdom is we have a liquidity trap, so we need negative interest rates because no one is lending or borrowing and the banks are sitting on excess reserves.
So I say this is a "glass half full" view, and why not start withdrawing the excess reserves because no one is borrowing the money anyway. I think the Fed over shot the liquidity need, and now is the time to make corrections.
Posted by: Cedric Regula at October 2, 2009 04:23 AM
First, David Papell concludes his post by stating:
If the incipient recovery takes hold and inflation stays the same or rises, it may need to raise rates sooner than many people think.
My point is that spending is likely to be weak if deleveraging by the household sector continues. I doubt that the Fed would want to raise the federal funds rate anytime soon.
As for your point:
why not start withdrawing the excess reserves because no one is borrowing the money anyway.
Can't the Fed can do that without raising the federal funds rate by reducing the interest rate it pays to hold excess bank reserves?
The only reason why the Fed would want to raise the federal funds rate is because it is worried about the economy overheating.
The idea it's possible to divorce the quantity of reserves being held from the conduct of monetary policy is discussed in a recent article published in the New YorK Fed's Economic Policy Review, Divorcing Money from Monetary Policy by Todd Keister, Antoine Martin, and James McAndrews.
Posted by: Anonymous at October 2, 2009 07:57 PM