Bryan Caplan  

Sumner's FAQ

The Shadow Banking System... Pet Health Care Theories...
Every economist alive should read Scott Sumner's FAQ on macroeconomics.  You give him ten minutes; he'll give you a fresh, incisive perspective on the world.  Highlights:

6.  Isn't the real problem . . . ?

No, the real problem right now is not a "real" problem.  The real problem is a nominal problem.  When the growth rate of nominal GDP falls sharply there is always a severe recession.  We have a severe nominal shock, a problem which has been understood by economists at least as far back as Hume.  At the time, it always looks like the "real problem" was some symptom of the monetary shock, such as financial panic.  Thus in the 1930s people thought the collapsing financial system caused the Great Depression, only later did we discover it was too little money.


9.  How can the solution for this mess be the same thing that got us into this mess in the first place?

The solution is stable NGDP growth at about 5% a year, which is not what got us into this mess.  It would be slightly more accurate to say that it is what kept us out of this mess between 1982 and 2007.  We got into this mess when we stopped providing enough money for modest growth in NGDP.


12.  Isn't monetary stimulus ineffective in a liquidity trap.

No.  Temporary monetary injections are never very effective.  Monetary injections expected to be permanent are always effective-even in a liquidity trap (according to well-known Keynesian Paul Krugman.)  What we need is an explicit NGDP or inflation trajectory, including a promise to make up for any short term undershoots.  This will increase the credibility of monetary policy.

13.  Don't we need both monetary and fiscal stimulus?

No.  Monetary stimulus can make NGDP grow as fast as you like, as we saw in Zimbabwe.  Once the Fed has set monetary policy at the level expected to produce on target growth, then there is no role for fiscal stimulus, it can only make things worse.


16.  How can I defend the EMH, when so many studies show people are irrational and markets are inefficient?

Market anomaly studies are products of data mining.  At some level this is known by economists, but the problem is far worse than even most economists realize.  These tests are not reliable.  People are often irrational, but it's not clear that irrationality has much impact on sophisticated financial and commodity markets.  The anti-EMH position has yet to come up with useful public policy advice, or useful investment advice.


19.  Aren't you just a monetary crank trying to solve all the world's problems by printing money?

Yes, but like a broken clock the monetary cranks are right twice a century; 1933, and today.  The other 98 years I am a Chicago-trained, libertarian, inflation-hawk.  Twice a century I put on my Irving Fisher super-hero suit, and emerge from my deep underground bunker.
Scott still hasn't answered all my questions to my satisfaction.  But who else is even trying?

Comments and Sharing

COMMENTS (12 to date)
winterspeak writes:

Bryan: Scott Sumner's emphasis on "monetary policy" only works if you redefine "monetary policy" to be "fiscal policy".

For example, if Helicopter Ben actually *did* start handing out dollar bills, Scott would classify that as "monetary policy" although everyone else would see it as being "fiscal policy" -- where the Government was paying people to be in the right place, at the right time.

He also has no idea, operationally, how bank reserves actually work. He believes that banks lend out reserves. This is not true, banks write loans if they think they can be paid back, and then borrow what they need to hit their target reserve balance. Similarly, if banks shift reserves to Treasuries (which happens all the time, it's why the US Govt issues treasuries in the first place) they just shift the term structure of money, they are not "placing it into circulation" in any way. This is not surprising if you think about it.

He also does not understand how balance sheets work, how private sector balance sheets must net to zero, and how net private sector savings must be funded by Federal deficits.

Not in my recommended reading list.

tms writes:

Recessions are not caused by mis-allocation of resources? They are caused when the government "stopped providing enough money"?

And how exactly does the government "provide money"? It provides inflation, yes, but "money"? How does providing inflation cause anything except, well, inflation?

I am clearly not sophisticated enough to understand these monetarist theories.

fundamentalist writes:

I don't think Sumner gets it that 1) the Feds can't produce the nominal gdp figure that he wants at will and 2) what good would it do?

Sumner complains that the Feds are not pumping up NGDP, but I don't understand how he can say that when the Fed has pushed interest rates to historical lows. Does he want the Feds to pay people to take out loans? The Feds can reduce interest rates, but if no one wants to borrow, what good is that? Greenspan called it pushing on a string. People don't want to borrow in a depression, in fact, they're doing just the opposite: they're paying down debt for crying out loud!

The Feds can also try to buy US paper, but again, people are fleeing stocks, real estate, and corp debt for US paper. Why would they want to sell?

In a depression, people are rebuilding their savings, paying down debt and increasing their cash cache. But to increase NGDP, the Feds have to figure out how to get them to stop doing what they think is in their best interest and persuade them to run out and buy things they don't want or need.

But suppose the Feds could pull of a miracle and make people spend like the world was ending, what good would it do? At best, prices would rise while production fell. And according to micro econ 101, higher prices for consumer goods would translate into lower wages for labor in consumer goods industries and therefore more use of labor to produce consumer goods. But because there is a depression in place, employers are more likely to work existing laborers overtime instead of hiring more workers.

Besides, the serious unemployment is in the capital goods industries, not consumer goods. Prices would not rise in the capital goods industries, so wages would still be high relative to sales and employers would not hire any new people!

I can't possibly see how fixating on NGDP can help anything!

El Presidente writes:

[hauls tattered drum from closet and strikes repeatedly]

Distribution matters.

[returns drum to closet]

winterspeak writes:

TMS: Government provides money whenever it spends. It withdraws money whenever it taxes. Interest rates play a very marginal role here, although lower interest rates.

fundamentalist: You are absolutely right. The private sector as a whole wants to save (delever). Only way for this to happen is for the Fed to lever.

Niccolo writes:


I liked reading this, but I just take a little issue with point number 6.

True, the shock of a contracting money supply did it, but the stringent labour regulations and harmful fiscal policies contributed nearly as much.

The US economy simply could not decrease unemployment and increase production with the type of economy it had at the time - I don't care if the money supply climbed to Mt. Everest, quasi-socialism just isn't good for growth in business.

Bill Woolsey writes:


There is no need for an increase in borrowing to increase Nominal GDP.

People can pay down their debts all they want and still nominal GDP can rise. And an adequate monetary policy can lead to this result.

Nearly all spending is funded out of income, not borrowing. It is possible for all spending to be funded out of income. There is no need for there to be _any_ borrowing for nominal income to be at any appropriate level.

As for debt repayments, those receiving the repayments can spend the funds on current output. This allows debt to fall and spending to rise.

Debt and debt repayment only impact nominal income to the degree they impact either the quantity of money or the demand to hold money.

They do, of course, shift funds between households and firms and impact the allocation of expenditres, the composition of demand, and, hopefully, the allocation of resources. Problems with credit markets could easily result in an inability to shift resources from low productivity to higher productivity uses.

The only "problem" with incrreasing the quantity of money to get nominal income back on target is knowing how much to create, and then creating too little or too much. Too little, and nominal income stays too low. Too much, and nominal income is too high.

Nominal income targettting does nothing to fix changes in structural unemployment or other productivity depressing shocks. Real income will fall and the unemployment rate will rise. If the nominal income target is maintainted, this will be inflationary. On the whole, firms will be unable to keep up production, there will be shortages of goods and services (of some sorts anyway,) and a higher price level, with inflation moving to that higher price level.

Scott's argument (and I think it is correct) is that falling nominal income has caused the current deep recession and falling sales and surpluses of goods and resources have led to falling production and employment and higher unemployment. This is nearly the entire problem at this time, with any higher structural unemployment and reduced productive capacity being minor.

Further, the financial problems, including people paying down their debts is being caused by the lower nominal income. They are doing this because they are worried about unemployment, business losses, and the like, that are caused by the lower real income.


You should learn some macroeconomics. Aside from being overly focused about trivia regarding banking, (monetary policy isn't about getting banks to lend,) I think your major problem is treating net private saving as given. It's not, and monetary policy can impact it.

Yes, budget deficits impact national savings, and, in equilibrium, they make up the difference beween investment other other sorts of saving. But the other sorts of saving and investment adust to the budget deficit.

Scott (and I) favor increasing by open market operations whatever amount is needed to get nominal income back to its pre-crisis growth path. It is obvious that paying interest on reserves balances at the Fed raises the demand for base money and is counterproductive. Stopping that and charging a penalty rate would help. However, if that isn't enough (which I think it may not be) then the Fed will just have to expand base money more, buying up more types of assets until nominal income reaches target.

I believe (and Scott disagrees or doesn't care much) that given current expecations about future levels of output, the needed monetary policy will have to lower real interest rates on long term and higher risk assets. While describing this as "levering" is confused, I think, I don't disagree. Scott points out that current expecations of future output will change when the Fed is committed to raising base money enough to get nominal income on target, and so the needed levels of real interest rate will rise.

Scott's point, which is correct, is that worrying about interest rates of any sort of an error. What will happen to them is driven by expecatiosn that depend on Fed policy. Low nominal interest rates could be caused by deflation. Or depressed future output. The Fed promising to keep interest rates low tells us nothing about what we should want. And that is nominal income being on target.

Liquidity trap issues may be relevant as to whether or not the Fed will have to purchase something other than T-bills to get base money high enough to keep nominal income on target. So? Once the Fed runs out of T-bills, it will have to buy other assets if nominal income is still too low.

fundamentalist writes:

Bill: "Debt and debt repayment only impact nominal income to the degree they impact either the quantity of money or the demand to hold money."

And when does debt not affect the quantity of money? The main way that the money supply expands is via loans to businesses. The state can borrow and spend, too. NGDP cannot increase without an increase in the money supply. It's simply not possible. In a depression, the money supply shrinks because people stop taking out new loans and repay the loans they made, in other words, they deleverage. That deleveraging and the loss of wealth from bad investments causes prices to fall.

I would be very interested in knowing how you increase the money supply to get NGDP higher without getting someone to borrow money.

Bill: "...falling nominal income has caused the current deep recession and falling sales and surpluses of goods and resources have led to falling production and employment and higher unemployment."

I think you have cause and effect exactly backwards. Falling NGDP is a result of falling sales and production and higher unemployment.

Winterspeak writes:

Bill: I understand macro just fine, although it is not a field that curently covers itself in glory

if you want to target nominal income, interest rates are a bad way to do it. In the US Zirp is comtractionary as it reduces nominal incomes, while credit remains expensive. Charging interest on reserves will be further comtractionary as it reduces private sector income further

the best way to increase nominal income is simply to stop taxing. This is cleary fiscal but Scott will claim it is monetary

Bill Woolsey writes:


As a matter of fact, business loans only make up a small fraction of bank assets. Less than $1 trillion out of $13.

Measured checkable deposits also make up a small fraction of bank liabilities. Less than $1 trillion out of the.. well, a bit less than $13 trillion.

Because of sweep accounts, however, actual checkable deposits are more than the measured amount. Maybe twice as high. Still, it is much smaller than bank assets.

Bank assets make up maybe 40% of total debt.

My view is that the money supply could increase by 500% or so without any private debt at all. Basically, it would all be existing monetary debt.

If that is too little to meet the demand, then it is true, then there must be some level of remaininng indebtedness for the money supply to be adequate.

However, that doesn't mean that total debt must expand. It could shrink, and by increaseing the proportion financed by monetary instruments the quantity of money could rise.

Only when all debt is monetized does an increase in the quantity of money require an increaes in debt. And the status quo is no where near that.

Bill Woolsey writes:


So, all that is needed is a price floor on interest rates, and nominal income can be increased without limit! You have created the secret of prosperity.

Lower interest rates may lower the income of lenders, but it reduces the expenses of borrowers, and the money they spend on other things raises the incomes of those selling those products.

Your "theory" that zero interest policy lowers nominal income would be similar to a theory that a good harvest of wheat lowers nominal income. Lower wheat prices surely lowers the nominal incomes of the wheat farmers, right?

So, we can raise nominal income as high as we like by putting on price floors on wheat. Gee, the wheat farmers make more. They buy more tractors... on and on....

Oh.. but we forgot about what the wheat buyers would doing. Paying more for wheat. Maybe spending more on wheat and less on other things, depending on elasticity.

Let's raise nominal income by raising the minimum wage. That raises nominal income, right? And the workers can buy more.. and on and on..

Oh, but the firms paying the workers have less. And if they raise their prices, those buying the products of minimum wage labor have less money to spend on other things (or maybe they spend less on those products.. depends on elasticities.)

The "income effect" of the lower interest rates is a transfer. But there remains a subsittution effect, towards more consumption and investment now.

Oh, and by the way, if someone wants to earn interest, let them purchase 30 year Baa bonds. I think the yield is 7%.

winterspeak writes:

BILL: If the Federal Government wants to increase nominal net (after tax) private sector income, there is a very very simple way it can do so immediately. Hint: it has nothing to do with monetary policy.

And you make my point for me. Lowering interest rates increases the nominal income of some private sector areas, and decreases the nominal income of others. So you need to sum losses and gains to see the net effect. In my opinion, increasing NIMs suggests that the current ZIRP policy is having a net negative effect on private income, and is thus contractionary (overall). As you say, it cuts both ways, and I have seen no evidence pointing out that it is, right now, expansionary (overall).

Your wheat example is an entirely private sector one, so I don't see what relevance it has to how the Federal Government (one sector) can help the private sector (another sector) de-lever. Various private sector actors shuffling assets between them has no impact on the private sector as a whole. That's the point of Macro. Since you recommended I study it in your first post, I assumed that you understood this point, but am now confused.

Comments for this entry have been closed
Return to top