Scott Sumner on Monetary Policy
Nov 9 2009

Scott Sumner of Bentley University and the blog The Money Illusion talks with host Russ Roberts about monetary policy and the state of the economy. Sumner argues that tight money in late 2008 precipitated the recession. He argues that the standard measures of monetary policy--growth in reserves or the Federal Funds rate--are misleading. Sumner suggests focusing instead on nominal GDP. He argues that the failure of the Fed to counter the drop in nominal GDP in late 2008 intensified the recession and points to the growth in unemployment. Along the way he discusses the Taylor Rule and other monetary prescriptions.

John Taylor on Monetary Policy
John Taylor of Stanford University talks about the Taylor Rule, his description of what the Fed ought to do and what it sometimes actually does, to keep inflation in check and the economy on a steady path. He argues that...
Scott Sumner on Interest Rates
Scott Sumner, of Bentley University talks with EconTalk host Russ Roberts about interest rates. Sumner suggests that professional economists sometimes confuse cause and effect with respect to prices and quantities. Low interest rates need not encourage investment for example, if...
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Nov 9 2009 at 9:43am

The problem is with my previous held conceptional dogma. That the central bank could devalue the base unit of debt (Federal Reserve Note) and subsequently the bonds held in that unit of account. Therefore, deficits and debt don’t matter.

Only thing is ,as Steve Keen points out, the Fed has the casuality backwards. Banks lend money first and then find a place to find someone to lend it to them! The Fed accomadates lending they don’t dictate it.

The analogy of the economy and any eco system is spot on. Sure you can intelligently design nature for you purpose.

Take a look at those cattle ranches if you don’t believe intelligent design isn’t possible. Only, you better make sure you have dogs and a cowboy with shotgun around.

Nov 9 2009 at 12:32pm


I’ve been hoping this Econtalk would happen, so thanks a lot for putting it together!

Nov 9 2009 at 4:57pm

This was a great podcast. Refreshing perspective from Sumner clearly presented. I came away feeling the podcas had added to my understanding of monetary policy issues. I will make two substantive comments. The first is on futures markets: the trouble with Sumner’s proposal for the Fed to target GDP futures is that futures market participants quickly lose interest in a futures contract with no volatility. If the policy works, the contract would die for lack of interest.

Second, seems to me Sumner’s optimism regarding inflation, i.e., the bond market is correctly forecasting low inflation, trips over his own logic. He says interest rates are low because the market expects low inflation; he also says the Fed can easily restrain inflation by raising rates when necessary. The 10-year yield is at 3.5%. How do you square such a low Treasury yields with a tightening scenario? Seems to me it’s logically inconsistent with either an inflation scenario or a tightening scenario. Conclusion: the market is wrong in that it’s not correctly predicting either future inflation, future interest rates, or both.

Gavin Andresen
Nov 9 2009 at 8:20pm

I think using future nominal GDP as a 21st-century monetary peg is a brilliant idea. It seems to me it would be an entirely self-correcting system.

For example, assume that the federal reserve or Congress will try to cheat somehow (maybe the Fed decides to, I dunno, do something boneheaded like pay interest on reserves or something).

When speculators find out, they’ll simply trade until the price of the futures contract matches their expectation of what the cheating will do to nominal GDP.

The speculators won’t always be right, but I trust the wisdom of people putting their wealth on the line more than I trust elected or appointed bureaucrats.

Charlie Quidnunc
Nov 9 2009 at 9:59pm

Great podcast. But, I kept wanting Russ to ask about the causes of the housing bubble, and the violations of the Taylor Rule that contributed to it.

As a person who makes a living selling into a deflationary economy, Information Technology for the last 25 years, I understand that it’s not as bad as some say. All my customers know that they can buy what I’m selling now for 35% less next year. That’s reality. But if we don’t look at the causes of bubbles and deal with them, they might happen again.

David Johnson
Nov 10 2009 at 12:44am

Tight money in late 2008? I think this may be one podcast I skip. I like to think I have an open mind, but I’m not sure I could get through this one without smashing my mp3 player.

Nov 10 2009 at 2:24am

As always, a pleasure. I often feel like a voyeur.

I was confused at about 30:30 where Russ gently prompted the term malinvestment, which was quickly linked to the concept of secondary deflation, and a secondary affect of falling GDP.

Why was this not a cause, assimilation, and response?

“Secondary affect” — my ass.

Nov 10 2009 at 6:14am

I agree that using NGDP futures as a peg is a brilliant idea. Hetzel’s proposal to use inflation indexed bonds is also good. Russ I felt that you were a bit quick to bring up political economy and it led you to be more dimissive of this idea than it deserves. While political economy is undoubtedly important, fundamentally, the notion that as governments can and do break the rules, they should not be constrained by them at all, is not logical. In fact, it’s anticonstitutional!

Paul Vreymans
Nov 10 2009 at 7:25am

@ Gavin Andresen

The future nominal GDP as a 21st-century monetary peg may be a very original idea but I fear it is just another illusion that one can to create real “money” from thin air after all the previous illusions proved failures.

I fear that at the very first serious external shock people would just dump all the futures they hold so that such monetary system would “run out of bullets” and become a lame duck just as soon as our present monetary system did. Another problem is how to start up the system. As people would only accumulate such futures during booms the system assumes a long prosperous period of before the next external shock occurs. In the present economical mess such assumption looks rather optimistic.

Doug Utberg
Nov 10 2009 at 11:56am

The thing that concerns me about nominal GDP targeting, and any version of targeting for that matter is the fact that it conditions people to become dependent on artificial stability.

By definition, any system that is based on people reading signals will eventually encounter a situation where the people misread the signals and trigger a disruption.

I remain a fan of the Friedman model where money supply growth is fixed at 3%-4% per year, and all other key financial metrics get set based on a market equilibrium. This would doubtless result in some short-term disruptions, but would create an incentive for the market to correctly price and absorb uncertainty.

Any system that is based on the judgment of people in charge will eventually be undone by the power it bestows on them. All people make mistakes and errors . . . it is a fundamental tenant of the human condition.

When power is concentrated, small mistakes destroy entire markets . . . when power is diffuse, mistakes fade into the millions of transactions that are constantly taking place.

Nov 10 2009 at 12:37pm

Russ, your readers might like to look at this collection of links to blogs discussing NGDP targeting:

Paul Vreymans
Nov 10 2009 at 3:41pm

@Doug Utberg

Before thinking of a new monetary system fundamental questions need indeed to be answered about feasibility and moral right of managing business cycles. Do authorities really have the moral right to boost spending and lead citizens into excessive debt just for the sake of smoothing business cycles and avoiding deflation? Is that not making humans to servants of the economy, and should that not be the inverse?

Do citizens not have the right to slow spending when they realize they spent too much?
Do citizens not have the right to hoard cash and protect their savings from bank failures when the risk exceeds the interest rate?
Is it really a government duty to keep upright expensive and unsustainable overcapacity in the car and housing industry?


Deflation is an unavoidable healing process after a period of excesses. It is necessary to re-allocate labor and capital from the sectors where they were wastefully utilized to the sectors where they provide more added value and are most needed.

Real and sustainable recovery is only possible after demand in the bubble sectors declined so much that excessive factor prices (due to excess demand in the bubble sectors)ease. Only then the spread between sales price and production cost in alternative sectors can widen and create such profit potential, that entrepreneurs can no longer resist the temptation to invest and recruit.

All the present monetary efforts and Keynesian stimulus packages are getting us nowhere despite their grotesque magnitude. The American cash for clunkers program and German scrap bonus scheme were unproductive incentives to precipitate people’ car purchases one last time and let manufacturers postpone the inevitable cleanup of overcapacities left by previous distorting stimuli. The grotesque expansion of the monetary base in both the euro- and dollar area does not help to relax credit to the industry. It only helps insolvent banks to clean up their distorted balance sheets resulting from previous monetary expansion.

All these Keynesian one time shots to boost the economy make the next crisis even more unmanageable. The similarity with a junkies needing ever higher doses is obvious and illustrates the unsustainability of stimulus packages and zero interest policy. Monetary policy and artificially low rates are the cause of our problems and not the solution.

Excessive money supply (Printing money) is a forgery which causes severe distortions: It was way too easy access to cheap credit that provided the incentives for the banks’ under-capitalization and their extreme leverage ratios. It were low interest rates which distorted the investment calculus and provided the incentives for disproportionate automation and expulsion of low skilled labor from production processes. It was easy access to cheap money provided big business with an excessive competitive advantage over labor-intensive SMEs. System threatening concentration and monopolies are the result. Industrial overproduction and chronic shortages of service providers are just a couple of the actual symptoms. Massive inefficiencies are the result.


Technological progress results in productivity gains of 2 to 3% each year. Under a just monetary system prices should therefore fall by 2 to 3% and not rise as we witness since the FED was established. The positive inflation targets adopted by Central Banks worldwide seize all the benefits of progress in favor of the banking sector and big business, systematically confiscating 4 to 6% of Joe Sixpack’s and Sally Housewife’s added value.

This inflation leads to severe distortions and is fundamentally unjust. The collective achievement of progress belongs to the whole of society, and in the first place to those who most contribute to it. Worse still is that inflation devalues worker’s savings so much that the real purchasing power is often only one third when they reach retirement.

A new monetary system can only be equitable, and achieve efficient allocation of resources when monetary growth is zero or at the most limited to the growth of the real economy. Only then money will effectively serve its purpose as means of exchange and saving. A return to the gold standard may be the best guarantee thereto.

Interesting related research paper:

Lauren (Econlib Editor)
Nov 10 2009 at 3:53pm

Clarification about terminology, by way of background for readers:

NGDP means Nominal GDP. That is, it means the value of GDP in dollars.

Some online sources confusingly define the acronym NGDP as something they call national GDP. That is not the ordinary or accepted meaning.

GDP stands for Gross Domestic Product. The term “GDP” already means output produced by the home country–the domestic or local country’s own output. The term NGDP merely emphasizes that it is a nominal value–meaning valued in a currency such as dollars from the perspective of that domestic nation. The “N” does not mean the number is additionally modified in some way to be produced by nationals of the country–that’s already accounted for by using GDP.

For more detail, see National Income Accounts and Gross Domestic Product (GDP) in the CEE.

Eric H
Nov 10 2009 at 7:48pm

Another fascinating podcast…like others, I’m glad Scott got the chance to discuss his ideas on Econtalk.

I’ve listened to the podcast through twice, and I need some help:

What is a “nominal” shock? Is this just shorthand for a shock to NGDP, or what?

Also a couple of comments:

1) His idea that the Fed should be “forward looking” implies that someone knows “where the economy should go.” Who is that someone? Or in masonomics terms who is the “we”? And how is this consensus about economic destination derived? Only politics answers these questions, and the political process has shown itself notoriously inept at driving the economic car.

2) Isn’t penalizing banks for holding reserves just another way of saying that banks should loan to increasingly risky lenders? Am I right to believe that banks holding reserves at low rates of interest isn’t a moneymaking venture for banks, it’s a means of reducing risk? If so, and the Fed penalizes banks that hold their reserves, isn’t the Fed forcing the banks to help misallocate resources into another bubble? Don’t banks have a better idea of who to lend to than the fed? I’m reminded of Tyler Cowen’s visit to East Berlin, during which he had a mandatory, 40 ostmark conversion. He bought a couple books and left the rest of the money on a bench. There was nothing he wanted to buy. Likewise, just because the Fed is convinced banks shouldn’t be holding reserves, it doesn’t mean there are awesome opportunities for growth out there that the banks are just ignoring for no reason. Perhaps there’s just nothing out there they want to buy.

Nov 11 2009 at 4:47am

Great podcast. Every time I read or listen to more from Scott Sumner, I learn something new.

@Eric H — ‘nominal’ shock is a very general term: it just means a shock to some nominal variable where a nominal variable is anything like NGDP or prices which is denominated in some kind of currency. The idea is that there are ‘real’ things which actually have value (furniture, homes, tools, services, etc.). But some things affect the economy, not by directly changing real stuff but by affecting currency. If a shock to the economy just affected currency, then that would be a nominal shock with no real shock. In practice, particularly in the short run, nominal shocks and real shocks tend to go together.

I add that the goal is to increase lending, and that means that banks should take on more risk than the ~0 risk they are taking on now by holding onto cash. But if banks really are very risk-averse right now, then that means that if we encourage them to hold on to less money, they won’t respond by giving money out to the riskiest people they can find, but to the safest people they can find. They will want to lend to big businesses and individuals with collateral. And that’s exactly what we want.

Nov 11 2009 at 7:13am

Eric H asked a great question:

What is a “nominal” shock? Is this just shorthand for a shock to NGDP, or what?

A “nominal shock” generally refers to an unexpected change in the money supply that was somehow not anticipated, foreseen, or understood in advance by private citizens. It affects NGDP through the “N” or “nominal” side–and through the slippage that Sumner describes. It is an example of a demand-side shock, and it is one of many suggested causes for a business cycle. The way the effects of an unexpected nominal shock might propagate through an economy were demonstrated by Nobelist Robert Lucas in the 1970s.

In the long run, economists agree that a nominal shock has no effect on anything real that matters to people in the economy. But in the short run–which could be anywhere from a few months to a few years, during that period when it is not understood by the public or is confusing to the public–a positive nominal shock could cause a temporary increase in employment, real wages, or both. It would ultimately be followed by an equal and painfully opposite adjustment by the public. For the government to do that intentionally would be a kind of trickery of the public that is frowned upon many economists–and very vocally by monetarists. Doing it unintentionally would have the same temporary effects.

A negative nominal shock (an unexpected decrease or tightening in the money supply) would similarly cause an initial and sudden decline in employment, real wages, and output. A negative nominal shock seems unlikely to be intentional; but that, or at least confusion on the part of the Fed about the incentives created by its own requirements, is exactly what Sumner suggests happened in mid-2008. A negative nominal shock perpetrated by a novice Federal Reserve is also widely accepted by economists as having greatly exacerbated the initial stages of the Great Depression, based on the pathbreaking work by Friedman and Schwartz.

One of the topics not fully addressed in Sumner’s podcast was how up in the air is the question of what causes business cycles. The question of what causes business cycles was posed as far back as the early 1800s and remains one of the greatest unsolved questions in economics. Causes proposed may begin on the supply or demand side, and are possibly based on real or nominal shocks, or may involve animal spirits, speculative bubbles, expectations that are unfulfilled, inventories that get built up, the long time involved between starting a project and seeing it enacted, the mere existence of banking or money as a medium of exchange, or government policies as predicted or as acted on with regard to the money supply, taxation, international interactions, domestic regulation, or other government policies.

But do you need to know the cause to cure the symptoms? There’s the rub.

I do not agree with Sumner that everyone agrees that business cycles are caused by the demand side; but I do not think resolving the causes is critical to thinking about whether his suggestions improve on the existing suggestions for moving forward.

Scott Sumner
Nov 11 2009 at 9:16am

Thanks for all the comments. Here are a few responses:

The NGDP futures market would be unlike other futures markets (where price usually fluctuates.) In this case on the day the contract was issued the price would be stablized by a Fed promise to increase or decrease the money supply as necessary to keep NGDP expectations on target. It is possible that no trading would occur, but that would be a signal by the market that the money supply was already at the appropriate level. However, on following days the Fed would issue new contracts, and the existing contracts would begin to see their price fluctuate, as with any futures contract. In case anyone is interested, this link explains the proposal:

I agree with those who say that the Fed should not try to prevent needed readjustments of resouces out of housing. But I want to avoid a secondary deflation, which can actually cause resources to become unemployed in sectors that were not overbuilt, like manufacturing and services. Those sectors may look overbuilt as well, but that is misleading, they only have excess capacity because of the secondary deflation. As late as mid-2008 they were doing OK, despite the steep housing recession.

Nominal shocks can occur from changes in the money supply or demand. They are the type of shocks that according to the classical economists should have no real effect if prices are flexible. But classical economists like Hume understood prices weren’t completely flexible, and that nominal shocks had real effects, but only in the short run. In contrast real shocks (like higher energy prices, the housing bubble, marginal tax changes and a higher minimum wage rate), affect the real economy in both the short and long run.

Some of you were surprised by my claim that money has been tight. This post explains what is wrong with tradtional definitions of tight money.

scott sumner
Nov 11 2009 at 9:31am

A few more responses;

Paul, Under my system of NGDP futures targeting if the Fed didn’t do its job or ran out of bullets, it would be easy for anyone to get rich. Thus last October I could have got rich shorting NGDP contracts as I knew NGDP growth would come in below 5% (indeed below 0%) But it’s not easy to get rich in the real world, which makes me think the system would work pretty well. I.e., speculators would force the money supply up enough to where NGDP growth was expected to be roughly 5% (or 3% if that target is chosen.)

Doug, It would only create nominal stability, not real stability. Friedman regarded stable NGDP as a goal of his money proposals. Because real factors would still fluctuate, people could not assume stability. For instance, housing and stock bubbles could still occur if people (or government entities) made foolish decisions. But I believe the bubbles and crashes would be smaller, as they wouldn’t be aggrevated by fast NGDP growth on the upside, and secondary deflation on the downside.

Eric, It is a common misunderstanding that the reserve penalty would encourage risky lending. Banks could still swap reserves for safe government bonds, or AAA bonds. It would actually restore the “normal” situation in banking. Normally all government and private bonds offer yields significantly higher than reserves (which used to be zero.) Since T-bill yields are almost zero, reserves interest rates should be negative.

Nov 11 2009 at 9:56am

Even so NGDP still seems like the nominal variable to target because it will always stabilize the economy at the natural / full info level of output (maximising real GDP), which is the only sustainable long run position to be in. As a bonus, policy makers don’t even need to know what the natural level of output is; they just need to stabilise NGDP.

Makes sense to me, anyway.

Nov 11 2009 at 10:33am

“Even so…”

That was a reply to Lauren, BTW!


I have a question about the effect of the Fed paying interest on reserves. How much does paying interest on reserves constrain lending? Is there any empirical work on this?


Bob Layson
Nov 12 2009 at 4:50am

There are two problems with the Fed. and its monetary policy. One, it exists. Two, it has a monetary policy.

Can we not have money for commerce commercially provided? Yes – and we would have were it not illegal so to do.

The talk was most interesting but too many economists appear simply to want to play advisor to the Money Tsar.

Nov 12 2009 at 6:52am

Hmm. I didn’t phrase that very well. When you say interest on reserve requirements, you mean that the problem is that banks are holding money lent to them by the Fed or by depositors, as reserves with the Fed (to earn risk free interest), instead of lending them out–basically a policy equivalent to increasing the reserve requirements… Am I getting this right? So there should have been a large increase in reserves and a concomitant effect on loans. Can we see this?

I’m trying to understand what the effect of paying interest on reserves was / is. I can see that it might change expectations (i.e. make the Fed look hawkish, which it likes, and so reduce expected inflation), but I want to understand what it’s real effects on lending were / are.

Nov 13 2009 at 4:06pm

Great episode. It is one of the few that I have listened to twice, and the only one that I have listened to twice back to back.

Keep up the good work.

paul vreymans
Nov 14 2009 at 9:42am

Dear Mr. Sumner,

Thank You for Your reply. Our previous remark did indeed not consider the obvious possibility of shorting the NGDP futures contracts. Although The NGDP system seems to have advantages over the present system, we continue to have fundamental question about messing around with the money supply altogether, regardless whether the rate of the Money supply is decided by the FED or the outcome of an NGDP futures market.

As “Austrians” we believe easy money causes many distortions and causes asset bubbles, disproportionate expulsion of low skilled labor from production processes, extreme concentration and excessive business cycles.(see our earlier remark).

We continue to wonder why central bankers and many great economists so heavily rely on monetary policy when there is so much evidence that the money supply has no effect on real growth, as Robert Lucas so convincingly explained in his Nobel Prize speech :

Two decades of near zero interest rates in Japan and Switzerland seem to confirm Lucas’ thesis: even persistent and extremely loose monetary policy could not stimulate their sluggish growth. A few years ago our group investigated in a multiple regression model the growth effects of interest policy in the 15 EU states over the 20 year period from 1985-2004 considering delays between 0-4 years and found that the slightly significant growth effect of easy money stimulus was overshadowed by the massive negative growth effect of inflation. A French author found even a remarkable inverse relationship between real output and M3 growth rate in excess of the growth rate of the real economy:

Do You have contrary evidence that easy money gives stimulus to the real economy ?

Thanks anyway for all the great podcasts.

Chee Heong QUAH
Nov 16 2009 at 3:06am

Yes, the Fed eased the money supply and did not adhere to Taylor rule.

In my opinion, the Fed should only control inflation and nothing else!

Hence, your argument does not make sense at all when it was the counterparty risk problem not the illiquidity problem.

Please read John B Taylor or Ronald I McKinnon first.

Chee Heong QUAH
Nov 20 2009 at 1:45am


This time is different from that preceding the Great Moderation.

I think it’s counterparty risk this time and even if interest rates on reserves are set to zero or negative, that won’t stimulate any significant demand for funds or investments.

Remember that before that the rates were already very low, around 2 percent so further lowering won’t make much good but perhaps detrimental to the dollar value .

If you let the market decide, that rate might be better in ensuring normal flow of funds because the price of funds that is the interest rate reflects the risk information in it so that businesses can manage their finances and portfolios effectively.

When you drive rates to zero or negative artificially, decisions cannot be made effectively because the information embedded in the rates is lost.

In fact that’s why we see the rising spread between the risky LIBOR rate and the risk-free Fed funds rate at that time. This reflects the risk-averseness of the market at that time.

Hence, even if rates were set to negative, it wouldn’t help much.

Dr. Duru
Dec 3 2009 at 10:39am

As usual, I enjoyed listening to the podcast and learned a lot.
However, after listening to this twice, I was still left baffled as to why the Fed would boost reserve levels only to hem them in by paying interest? Are the reserves just for show to give the market comfort that IF the economy takes another turn for the worse, the banks have enough cushion to absorb the blows? That’s the only thing I can think of!
Also, it seems we are splitting hairs here as to what the Fed should have done in Oct, 2008. I was not left convinced that if the Fed had rushed money into the system and let (nominal) rates drop to zero, as suggested by Sumner, that we would be much better off than we are now. The Fed seems to prefer gradual approaches given its understanding that monetary policy acts with lag effects. Do we really want to the Fed to swing policy to extremes to every extreme economic event? Seems like a recipe for whipsaw and many unintended consequences.
Finally, regarding expectations and targets. In the short-run, the Fed has indeed had dour forecasts. But for the longer-term outlook, the Fed consistently praises the strength of the economy, insists that its policies will drive the economy to its long-term potential, etc… How much can we gain by having the Fed stretch the truth by insisting its policies can drive nominal GDP to some aggressive target in the short-run? All it takes is for one failure, and the Fed immediately loses the credibility it needs so desperately.

Dec 9 2009 at 3:43am

Sumner does a great job explaining the coincidence of Keynesianism and monetarism. Too many people see Keynes as saying “spend spend spend” when the whole point is he only said to spend under certain circumstances. Keynes said as much to Hayek himself over dinner once – that his policy only applied to deflationary periods.

Comments are closed.


About this week's guest:

About ideas and people mentioned in this podcast:Books:


Podcasts and Blogs:



Podcast Episode Highlights
0:36Intro. [Recording date: October 30, 2009.] Critical of monetary policy for the mess we are in. Different perspective. What went wrong and when? Divide up the crisis into two parts. Initial part when subprime mortgages became a big issue in late 2007 played out as most people think. Economy slowed down a little bit up until about August 2008; no damage to the broader economy. Where people went wrong is in underestimating the impact of monetary policy made after August 2008: highly contractionary in the only definition that makes any sense. A couple of ways of looking at monetary policy. A lot of people are surprised by hearing that monetary policy was contractionary: didn't the Fed cut interest rates to low levels? Very misleading indicator of monetary policy. In the 1930s, the Fed also cut interest rates to low levels, but today most economists think monetary policy was highly contractionary in the Great Depression. Some people point to real interest rates; but those actually rose sharply in the last half of 2008. Others point to the monetary base--the money actually printed by the Fed--and that did almost double late in the year; but really two reasons why that's not a good indicator. One: when interest rates get to really low levels, people tend to hoard cash. During the Great Depression, the monetary base rose sharply in the early 1930s, yet monetarists like Friedman and Schwartz still regard monetary policy as being highly contractionary. In addition the Fed started paying interest on bank reserves in October 2008, and that explains much of the increase in excess reserves. The traditional indicators are not very reliable. When we look at those, even though on the surface they look expansionary--the low interest rate and the huge increase in bank reserves that the Fed injected into the system: is the argument that velocity--the rate at which people are spending the money--is what dropped, out of either anxiety, uncertainty about the future, caution, the low interest rates. Is that the fundamental mechanism by which these seemingly expansionary moves were relatively ineffective? Yes, velocity did drop; reasons complex. As economy slowed, the Wicksellian real or natural interest rate fell to a very low level. That would be the interest rate that would provide equilibrium in the economy. As interest rates fall, people tend to hoard money more--the demand for money rises. Also, there is financial uncertainty. Money is a safe liquid asset. There was a huge increase in uncertainty about where assets were headed. What people overlook is that monetary policy is really about adjusting the stance of policy to reflect changes in the economy. The reason why Friedman and Schwartz argue that monetary policy was contractionary in the early 1930s despite the big increase in the monetary base and the decrease in interest rates is that the Fed adopted a contractionary stance relative to what was needed. So their actions didn't prevent the broader money supply from falling, or prices or nominal GDP from falling sharply. Monetary policy should not be focusing so much on interest rates or money supply but rather setting a policy stance that's expected to produce on-target growth in aggregate demand. What happened is the Fed adopted a contractionary policy relative to what was needed to keep aggregate demand growing at its normal rate, which is about 5% a year over recent decades. Defining that as the total dollar value of spending in the economy, or nominal GDP. Instead of growing at a normal rate of 5%, it actually fell at a rate of about 2% a year after about 2008. Monetary policy was too contractionary relative to what was needed to offset the falling velocity.
6:07Talk about terminology. Most people come to macroeconomics, realizing it or not, as Keynesians. In our cultural blood. Aggregate demand--Keynesian idea, need to keep spending up. Contrast that Keynesian story with the story of Irving Fisher, early 1900s. Fisher, and later Milton Friedman, use a simple framework for thinking about money--an accounting identity, that MV=PT. That is, the amount of money, times the rate at which it's turning over equals aggregate activity--PT, sometimes PY, where T or Y equals transactions and P is the price level applied to those goods and services. MV=PT, MV=PY. Have to be equal. If we see PY falling, either because there is deflation, with P going down, or Y going down, that is real economic activity, we want to offset that with M going up, to make sure economic activity stays the same or continues to grow. How does that relate to the Keynesian idea? Or is it the same idea? In a way it is the same idea, but the two schools approach a lot of the surrounding ideas differently. If there is a drop in MV or a drop in nominal spending, PY, that's essentially a drop in aggregate demand in the Keynesian language. Where they differ isn't so much in their views about the proximate cause of a recession. Both the Keynesians and monetarists believe that not enough nominal spending is the proximate cause of a recession. Where they differ is what's causing that to occur. The monetarists put more weight on failures of monetary policy; the Keynesians put more weight on consumer pessimism, sitting on their wallets, animal spirits and have doubts about whether monetary policy can actually fix that problem. Third perspective, Austrian perspective. Problem: in both stories, a similar element of animal spirits. In Keynesian [Russ accidentally says "monetarist" here] story, it's that people get nervous, sit on their money, don't spend it; government has to step in and spend more. In the monetarist story, it's people get nervous, velocity goes down, they don't spend as much, and government has to step in and boost M. So in the monetarist story, boost M; in the Keynesian story, boost Y or some measure of real spending. A little misleading. Keynesian story is actually about trying to get nominal spending up. The question about whether that increase in nominal spending results in more real output or more inflation is, according to the Keynesian model, a function of where we are on what's called the aggregate supply curve. If at full employment, get inflation, if there is excess capacity, unemployment, mostly get real output growth. Both talking about driving nominal spending higher. Keynes would admit that if you are at full employment and you do that you are probably going to get inflation. The monetarists are more pessimistic about likelihood of increasing real output; think inflation is relatively more likely outcome. In the 1960s and 1970s, monetarists gained ground because it seemed like the Keynesian policies just resulted in inflation. But those are differences in how they interpret the parameters. Fisher, Friedman, and the Keynesians all have demand-side models of recessions; different from real business cycle or classical models.
11:42Current situation or historical interludes like the Great Depression. What caused the post-August 2008 contraction? Complex: depends on how you look at it from a policy point of view. Can either look at it as velocity going down being the cause, or as the Fed failing to react to that as they should have with a more aggressive monetary policy. For the previous 25 years, the Fed had been offsetting changes in velocity with changes in the money supply in order to keep nominal GDP growing at 5% a year on average. That's the Great Moderation, which suddenly ended in late 2008; Fed allowed nominal GDP to fall sharply below the trend line; we're about 8% below where we should be if they had continued that policy. Because of that decline, real output fell as well--hard for the economy to adjust in the short run to a nominal shock. Play Keynesian for a minute: A Keynesian would respond that's all well and good. You're saying the Fed was insufficiently aggressive in responding to this contraction even though they doubled the nominal base and even though they pushed interest rates down to a quarter of a percent--they had nothing left. Keynesian would say: we have to turn to government spending to get the economy going again. Response: Most economists don't pay much attention to what the Fed is doing. Yes, they increased the monetary base, but they instituted a policy of paying interest on reserves explicitly to prevent inflation from getting out of control. James Hamilton, Bob Hall have talked about the policy. They bribed banks to hold onto the money instead of moving it out into the economy. In one sense, you can say that's understandable--if you double the money supply, you will get hyperinflation. But how can you say monetary policy is ineffective if they were instituting a policy to prevent it from being too effective. Why did they do that? Their explanation is slightly different but comes to the same thing. Said they wanted to keep control on interest rates. Recall that when they did this, interest rate targets were still at 2%. When they flooded the banking system with reserves, this would have normally driven the Federal funds rate down to zero immediately in the free market. But they wanted to keep their target rate at 2%. So they had to pay banks to hold onto the reserves, not lend them out, to prevent the excess reserves sloshing around from driving interest rates immediately to zero. What should they have done? Should have let interest rates fall to zero immediately. It was obvious at that time that we were going into a very severe downturn; Fed was very slow to react to that fact. They were backward looking in their policy. We had just gone through a high inflation blip--remember the high oil prices--pushing the headline rate of inflation up to 5%. But by September and October the economy was deteriorating and oil prices were plunging; all the forward looking indicators suggested that inflation and real growth were going to come in way below the Fed's targets. Like driving a car down a road steering by looking through a rear view mirror. Instead should be looking forward.
17:19Is it possible that they were too focused on interest rates and worried if they pushed them down to zero, not only is there an inflation worry but we'll have lost our last piece of ammunition in our fight? Strange fear, but possible. Strange: we're not going to do what we need to do in an emergency because we won't have the ammunition to do it some time later. To play the Keynesian--qualitative easing--of course we have to turn to government spending because monetary policy is impotent. Tried increasing reserves; when you finally pushed interest down further, you had nothing left. One more example for inflation hawks. Joan Robinson, a real strong Keynesian, argued back in 1938 that the German hyperinflation couldn't have been caused by easy money because interest rates weren't low. Today: that's an absurd argument. But we are making the same argument. Look at other examples; set an explicit target; constrain Fed by rules. Hayek also thought it should be nominal GDP. Quantitative easing: Talk about what tools the Fed has available--fancy term for creating more money even when interest rates are low. Focus on interest rates is a red herring. Three tools and three steps. First would be to stop paying interest on reserves, and if necessary have an interest penalty on excess reserves so that banks don't hoard the reserves. If that's not enough to increase aggregate demand, can do quantitative easing, putting more cash out into the economy by buying government bonds--bonds that are on the balance sheets of banks or even held by the public. Fallacy of composition: what's true for the individual is not true for the group. If you put a lot of cash out into people's pockets, typically they don't want to hold all their assets in the form of non-interest-bearing cash, so they'll try to get rid of it. As individuals they can do that, but as a society we can't; so the attempt of everyone to get rid of all these excess cash balances will drive up aggregate demand. That's the fundamental principle of that underlies basically all of monetary economics. When we say aggregate demand--and also dollar value of GDP--when that grows 5% it could be a 5% increase in prices with an unchanged real standard of living; or it could be a 5% increase in our real standard of living and prices are constant. Confusing because there is something else going on along the way, for example, in the middle of a recession or doing fine or over-expanded and will get more inflation. Would be a misinterpretation to think that by the Fed injecting money into the economy it will get the economy to grow. It's not going to be real growth necessarily; could just be inflation. Trick to understanding macro is to juggle two balls in the air at one time. One is the long-run classical model, that our real growth is determined by real factors like population growth, technology, productivity, free markets. The other is that in the short run, the fluctuation in nominal GDP that wouldn't make any difference in the long run will affect real GDP in the short run. Nominal shocks like changes in the money supply can have a short-run impact on real output until wages and prices have fully adjusted. That shock could come from a Fed mistake or an unexplained change in behavior by consumers/investors. Can't control real growth in the long run, so let's have growth in nominal GDP that would keep inflation low. Using 5% in blog, which would be about 2% inflation and 3% real growth. Perhaps it should be lower; but need some kind of stable long run policy. What about the business cycle--the short run? To minimize that, should have nominal GDP grow at a relatively stable rate. If there is a sharp change in nominal GDP growth, in the long run it will only affect inflation. But in the short run it can create a business cycle--either too much output, an overheated economy, or a recession if too little nominal spending. Hard to say whether money affects real growth or not. In the long run it doesn't. Nominal growth that's expected--already priced into wages and prices--doesn't have any real effect. It's the unexpected shocks that matter.
25:21Unexpected. If everybody knew that prices were going to grow at 2% a year, everybody would factor that into interest rates. Value of the money changing hands a year from now goes up. Inflation would be irrelevant; deflation would also be irrelevant. People have this bizarre fear of deflation; unusual in our lifetime. If everybody understood that prices fell 2 or 3% a year, they would factor that into wage expectations, their borrowing and interest rates. Where you get real effects is when outcomes don't mirror very closely your expectations. If I lend you $1000 and expect to get $1100 back, and prices are stable; if suddenly prices went up in a way that wasn't anticipated at the time of the loan, those swings in reality versus expectations discourage economic activity. If they are anticipated, those nominal changes are relatively unimportant. In a normal post-war recession quite often inflation slows to a level less than expected. In 1982, bad recession even though inflation 4%, but it was much lower than the expected 10% that we had been having. Caused a lot of unemployment. This time around, went from mild inflation to mild deflation; hurts employment if wages are slow to adjust. But this recession we had a problem we didn't have in the 1982 recession--went into it with a fragile banking system devastated by the subprime crisis. With an unexpected deflation, also, it is hard for people to repay loans. Also had the debt crisis get much worse because people had borrowed money anticipating the inflation we had, but instead prices started falling in 2008. One additional point: better to look at this process from the perspective of nominal GDP rather than the perspective of inflation. Inflation figures are very inaccurate; also, nominal GDP can be thought of as the nominal dollar income that the public has to repay their debts. It's the total command over resources. So, by the middle of this year people's nominal income, including corporations and the public was about 8% below what it would have been to follow the usual trend. People had about 8% less dollars to repay their loans with than they had anticipated. What was originally half a trillion or trillion dollar mortgage crisis spread to many other kinds of debts, including industrial loans, better quality loans. Most not due to original bad lending practices but secondary effect of falling nominal GDP. This is what the Austrians call a secondary deflation. The first crisis described the start in late 2007; Austrian term is malinvestment--too many houses, too big houses. Secondary deflation made the problem much worse; same as in Great Depression. If you are just looking at news it kind of looks like it was one big problem. Can see this in terms of the U.S. housing market--just a few states. No decline in most other states at all. When nominal GDP starts falling, it affects the whole economy. Housing prices started falling in states like Texas which had avoided the whole initial problem.
33:08Clarifying question: Made the observation that it's difficult to measure prices as a whole--consumer price index (CPI), etc. Attempt to create a basket of goods is very challenging. Example from blog: consumer price index in the middle of this year, core inflation still up about 1%. Looked at components: housing almost 40% of the core inflation. The government claimed housing prices were up 2% between mid-2008 and mid-2009. That's because of the peculiar way they measure. Implicit rent: government tries to figure out what your house would rend for. They don't look at the market or housing prices, which have been falling. But people sign long-term rent contracts; they will be locked into a long-term schedule. Why does nominal GDP avoid that problem? It also has P times Y. It puts more weight on new construction. Nominal amount of new houses constructed fell sharply; which caused construction workers to lose their jobs. Distinction between GDP deflator and the CPI? Yes, but the GDP deflator is also flawed; but it's better. Nominal GDP is measuring the size of the nominal shock that will eventually cause the price deflator to be 10% lower. In the short run what will happen is that prices will fall less than 10%, but output will also fall. Nominal GDP will fall 10% because prices and output will each fall 5%. Would still cause that a negative nominal shock. In the long run, you'll just end up with 10% lower prices. Consumer price index, which includes a lot of sticky prices such as rents or prices in catalogs, it's not going to pick up what's going on real time in the economy when there is a sudden nominal shock. What picks this up best is many asset prices--stock prices, commodity prices fall much more quickly than the CPI. Not in favor of targeting asset prices, but it's understandable why people are pointing to them. Their prices appear to be more flexible.
37:33Suggesting that the Central Bank should be keeping an eye on nominal GDP as the best current measure of whether the economy is shrinking or growing and therefore whether monetary policy needs to respond. How does that differ from the Taylor rule? He wants to maintain a stance taking into account both the growth of the economy and inflation. Correct? Inflation and how the economy is doing relative to its full employment level. Taylor Rule is similar; but the big difference is John Taylor prefers a backward-looking approach. Take historical data, looking at past inflation and past real GDP growth data. Sumner suggesting that the policy should be forward-looking: always set policy such that nominal GDP is expected to grow at the target rate, say, over the next 12 months. Taylor also wants to use the Federal Funds rate as the primary level. Sumner skeptical. Could instead create a nominal GDP futures market, in a sense make money redeemable into these futures contracts. In essence, the Fed would stabilize the price of a nominal GDP futures contract, much like they used to stabilize the price of gold under a gold standard. Nominal price was fixed; Central Bank would either buy or sell gold at that price to maintain it. Problem with the gold standard was that sometimes when the price of gold was stable, other prices would change, so it didn't always produce an optimal result. But it did avoid huge amounts of inflation. Let's create an asset better than gold and have money convertible into them--these nominal GDP futures. How would that literally work? Contract that pays off depending on what nominal GDP turns out to be a year from now. So if speculators thought we'd have 7% nominal GDP growth, but the Fed's target is 5%, they would take a long position, buying these contracts from the Fed. That would be a signal to the Fed to do an equal decrease in the money supply in order to lower expectations of nominal GDP growth. Why would I buy that, though, knowing that the Fed is working against me? Sounds like: let's have a race, and I'll bet that you are going to win. You are going to bet that you are going to lose. But it's easy for you to run slower. Hard for you to run faster; but you are in control. Why would I bet you that you are going to run slowly? Assuming the Fed is serious in trying to hit its target, so the Fed takes a passive position. Not quite right to say the money supply is under control of the Fed; it's really under control of the speculators. Goal is to set up a system that works automatically. Fed automatically does the opposite. Can make that policy work. With any policy, even a gold standard, there is always the risk that the government will try to profit from it. In the Great Depression, the Fed almost doubled the price of gold and made a lot of money on their gold stocks. Have to assume that the Central Bank will adhere to the rule.
43:29Different political economy question. Milton Friedman, in 2006, made a comment on this program. In the 1970s in grad school, taught that the Fed was worried about the growth of some monetary aggregate--M1, M2. So in those years, Friedman was often advocating an automatic growth rule in M2, say 3% a year, roughly equal to the rate of productivity, so we would have a stable price level. Somewhere in the 1980s and 1990s, Central Bankers stopped talking about aggregates like M1 and M2, and started talking about interest rates: set interest rates to set prices and stabilize inflation. Same goal, different target. Standard argument as to why they switched was that we couldn't measure M1 and M2 as accurately. Hard to measure inflation, so instead of having inflation target we'll have a Federal Funds Rate target because that's easier to measure. Asked Milton Friedman why the change; he chuckled and twinkle in his eye, and said that's what they say, but they are really just trying to keep a stable money supply. Milton sent an Excel spreadsheet. The reason they talk about interest rates is it makes them look more impressive; they are steering the ship. This was at a time when Alan Greenspan, Chairman of the Fed, was considered a genius, maestro. What of claim by Friedman that you still want to look at monetary aggregates? Money supply is better than interest rates as a stance of policy. Back to hyperinflation example: money supply correlates better with inflation than interest rates. But still going beyond M2 to the aggregates themselves, inflation or nominal GDP, and especially expectations of them, are really the best target. Why stop at an intermediate target? It's true that an open market operation affects the monetary base and that affects M2 through the multiplier, and then depending on velocity that affects nominal GDP. But not really trying to control M2--trying to control inflation or nominal GDP. In one of Friedman's last books, Money Mischief, he actually sort of endorsed a proposal by Robert Hetzel to have the Fed do exactly that--target inflation expectations. Target the gap between an indexed bond and a conventional bond. Gap between those two interest rates is roughly what the market thinks inflation will be over the period of the bond. Friedman said good things about that proposal. Not unsympathetic to the ideas. When he wrote that book we didn't even have an indexed bond market in the United States. Different question: John Taylor's been critical of the Fed during the 2002-2004 period for deviating from the Taylor Rule, lowering the Federal Funds rate much below what the Taylor Rule would have suggested. Monetary policy too expansionary in that period. Agree? Perspective on that period? Don't think interest rates are a good indicator of monetary policy, so low interest rates are not a sign of easy money. Low interest rates in Japan or in the United States in the 1930s not a sign of easy money. Friedman said low interest rates are usually an indication that money has been tight and that you have had deflation in an economy, or very low inflation. Understand Taylor's argument; but look at nominal GDP. Mixed view: In early part of that period, growth was pretty slow and low interest rates appropriate. Recession in 2001. Later part, expansionary policy, nominal GDP growing a bit too fast. However, the mistake we made doing too much expansion, inflation, was very small compared to the mistakes we made in the 1960s and 1970s; didn't have huge housing bubbles in those decades. Don't think the Fed's easy money alone can explain the recent situation and the size of the subprime fiasco. Other mistakes were made, either in private sector or by regulators that go beyond monetary policy.
51:37Back to political economy issue: Think of a whole range of stances we'd like to see the Fed take. Some would be very discretionary--they have a lot of choice, a lot of freedom to steer the economy, micromanage it. Or more restrained, more hemmed in: have them take a stance to try to keep nominal GDP on some kind of target growth rate. Friedman pushed for a target of a monetary aggregate, non-discretionary; Taylor, non-discretionary restraint on Federal Funds rates. All attempts to encourage less discretion on the Central Bank. Given the political reality, is it feasible that any of these are likely to be successful? Might we be better off with a different kind of Central Bank or different kind of monetary system--private money, a gold standard, something that might be easier to enforce? Futures contract scheme you were talking about--complicated--harder to enforce the constraints on the Fed. Central Bankers are human beings who want to have a reputation for skill. Political reality. Is this the best way to go? Hard to say. Problem with getting the government out of money is that it's not really clear what the monetary system would look like. Gold standard combined with free banking; even some free banking experts skeptical. Gold standard doesn't necessarily promote economic stability. Could have periods of inflation and depression if there is an increase in demand for gold. May not be linked to gold; but then what would you link the dollar to? Don't know what that system would look like. With all its flaws, could argue that the discretionary system we've had in the past 25 years hasn't really done any worse than the past systems in terms of the business cycle. Reality of the modern world: governments like to have a lot of control; move forward incrementally. One incremental improvement: set an explicit target. That's been done in other countries, but not yet in the United States. Another: target the level instead of the growth rate. Spell out where they want something like the price level to be and commit to try to return to that trajectory if they deviated from it. A little more optimistic. Look back at the last 100 years of Fed history; can see some really gross errors that were made; can also see the Fed learning from the mistakes. Never quite repeated the same mistake of the Great Depression, letting M2 fall 30%. After the 1960s and '70s rethought strategy toward inflation; strategy improved. Now this crisis has revealed a third problem: don't know how to handle sudden changes in expectations and velocity; too backward-looking in their policy. Can perhaps learn from that as well. Before this crisis we had the 25 most stable years in American history in terms of the business cycle. Followed by a really bad recession which both Taylor and you have put at the feet of the Fed, though with different explanations. Too much hubris. Six years ago we were confident we had mastered this business cycle thing.
57:50What seems to be true: if you look back at the last century, the Fed does seem to have gotten better at moderating inflation, which was its job. It's also now looked on as the killer of business cycles, which is difficult to do. We could give them a B+ because we haven't had hyperinflation, deflation, and price level has been fairly stable. However: when Alan Meltzer was on this program, he was confident that the reserves the Fed has on their books and the banks have been hoarding are going to go flying out the door; inflation will reignite; pressure to raise interest rates and political pressure for them not to. What are your thoughts? "Good economists don't make forecasts. They infer market forecasts." Right now, the futures markets, CPI, bond market all forecasting very low inflation. Why? Because the Fed has an operating procedure that allows them to break out of inflation. Though they don't watch expectations enough, there are inflation hawks at the Fed. Every cycle we've had since 1982, we've gone to lower and lower trend rates of inflation. Think that this one is the same; we're now coming out of this recession at an even lower trend rate of inflation. But we haven't doubled the monetary base at any of those times. They'll either keep paying interest on that or they'll pull those reserves out of the banking system. One point on the free market view: realistic view; country is not going to let nominal GDP bounce around according to the gold market. Problem you have is when the Fed makes mistakes and creates deflation, it looks like a failure of the free market. Conventional view is the free market is not working. Want to keep nominal GDP growing at a steady rate partly for libertarian reasons--the free market will look much better to the average person, won't be looking for bad policies like bailing out banks that are really just treating the symptoms of bad monetary policy.
1:01:37Given those past mistakes, what could the Fed be doing now? We've seen a mix of both monetary and fiscal policy. The Fed's made a massive increase in reserves; they've lowered interest rates. That's had some impact? At the same time the Federal Government has pledged to spend an enormous amount of money. What would get us out of the mess now? Tempting to look at the low interest rates like the Fed is doing a lot, but the countries that are successful, like Australia, which hasn't had a recession since 1991, actually have much higher interest rates because higher rates are a sign of prosperity. Lower interest rates are a sign of failure. Need to set an explicit target, for, say, nominal GDP and promise to do whatever is necessary to get expected growth up to that target. Or it could be an inflation target. Would have to increase reserves and lower the rates they are paying on those reserves to encourage that money to get out into the economy. Also possible that just setting a target would so increase expectations that the same amount of money out there would help, changing velocity. Doubling of the monetary base and low interest rates are symptoms of deflation. Need a broader strategy. Might be higher interest rates, or lower interest rates. Could charge a negative penalty on excess reserves. Paradox: crystal ball, would much rather interest rates be 4% than 0%--recovering rather than in the Japanese situation. Both the fiscal stimulus and the monetary policy haven't worked very well. Both suffer from people being anxious about the future. Governments and economists are not very good at changing people's expectations. We don't have a good model of giving people confidence. Monetary policy can be very powerful, even too powerful as in Zimbabwe. Fed hasn't tried to shift people's expectations. Instead the Fed has basically predicted failure. Wasn't the doubling of the base their way of saying to the world they were going to be aggressive in trying to increase the rate of growth of prices? Lars Svensson: the Fed should always target its forecast. Starting last October 2008, the Fed started forecasting inflation and real growth at levels far below what everybody believed the Fed wanted. So the Fed started forecasting failure. Could say there was nothing more they could do. But interest rates at that time were 2% and they were paying banks to keep banks from letting interest rates fall below 2%. So they could have done much more. Even when rates hit zero, they can an almost infinite amount of quantitative easing or the printing of money. The Fed wasn't taking any action to move expectations up to its target.

More EconTalk Episodes