Bryan Caplan  

Additive Shocks

The Singularity: What's in it... Robin Hanson on Ethics...
A common complaint about Sumner is that he ignores important real shocks, such as the "Recalculation Problems" that Arnold keeps pushing.  I'm still not convinced that recalculation is a big deal, though this graph that Arnold's promoting is the kind of evidence that might eventually change my mind. 

Whether or not I convert to the recalculation story, however, Sumner's main points still stand.  Why?  Because real and nominal shocks are basically additive

To see why, consider two variants on the following thought experiment: Over the last five years we wasted 10% of GDP on mud pies.  Now we've suddenly realized that no one wants mud pies. 

Variant #1: Scott Sumner runs the Fed.  He keeps nominal GDP growing at 5% per year.  What happens?  Unemployment still rises - and measured output falls - because mud pie workers have to look around for a better use of their skills.  But output of non-mud-pies gradually rises as the productive 90% of the economy absorbs the cast-offs of the unproductive 10%.

Variant #2: A misguided utilitarian deflationist (he likes Rothbard's macro, but not his ethics) runs the Fed.  He adjusts the money supply to make nominal GDP shrink by 5% per year.  What happens?  The mud pie industry still collapses.  But unless psychological resistance to nominal way cuts changes far quicker than it ever has in the past, workers throughout the economy quickly become grossly overpaid relative to their productivity.  This leads to massive falls in employment and output in the industries that would otherwise absorb the displaced mud pie workers.

Leading Question: If you subtract the damage in Variant #1 from the damage in Variant #2, what do you get?  You get exactly the damage you would expect if the mud pie debacle never happened BUT our misguided utilitarian deflationist made nominal GDP shrink by 5% per year!  Namely: Massive unemployment caused by excessive nominal wages, and a corresponding fall in output.

P.S. If this seems horrifically Keynesian to you, I have to point out that Rothbard, Dr. Deflation himself, quietly grants my two key points in a crucial section of America's Great Depression:

1. Deflation reduces employment and output when wages are downwardly rigid.
It is, moreover, a common myth among laymen and economists alike, that falling prices have a depressing effect on business. This is not necessarily true. What matters for business is not the general behavior of prices, but the price differentials between selling prices and costs (the "natural rate of interest"). If wage rates, for example, fall more rapidly than product prices, this stimulates business activity and employment. [emphasis mine]
The flip side, obviously, is that if wage rates barely fall at all, and product prices do, the "common myth" is true.

2. Wage rigidity is partly a product of human psychology, not just unions and regulation.
Generally, wage rates can only be kept above full-employment rates through coercion by governments, unions, or both. Occasionally, however, the wage rates are maintained by voluntary choice (although the choice is usually ignorant of the consequences) or by coercion supplemented by voluntary choice. It may happen, for example, that either business firms or the workers themselves may become persuaded that maintaining wage rates artificially high is their bounden duty. Such persuasion has actually been at the root of much of the unemployment of our time, and this was particularly true in the 1929 depression.

Comments and Sharing

COMMENTS (19 to date)
Lee Kelly writes:

So far as macro is concerned, I think wage rigidity is tantamount to expecting steady or rising nominal GDP.

The real and nominal shocks feed on each other -- Hayek's "secondary recession." He also said that grass is green, by the way.

Lee Kelly writes:

Question: why did nominal GDP growth take a sudden nose dive?

It seems to me there are two possible answers:

1. Money demand was steady, but money supply growth decreased or reversed.

2. Money supply growth was steady, but money demand suddenly rose.

What was it? My belief is that it was the second. Money supply growth was steady, but it was inadequate given the new condition of higher money demand. In other words, the Fed was too contractionary relative to the increase in money demand.

What caused the spike in money demand in the first place? I suspect it was an increase in uncertainty following a real shock (i.e. the housing bust and recalculation).

The initial failure of the Fed led to deflation and further miscalculation, which in turn created more panic and increased money demand even more. Meanwhile, the Federal Government did its level best to spread more uncertainty and stand in the way of recalculation.

It seems to me that the Fed probably needs to be more expansionary at the moment, and the Federal Government needs to stop its fiscal "stimulus." They might also reintroduce the rule of law at some point.

Some people imagine that when the Fed "does nothing" we get a glimpse of the free market at work. Many Austrians take this attitude toward the 20-21 recession. But even though the Fed "did nothing" in response to the 20-21 recession, that doesn't mean it didn't intervene in the economy, because its very existence is an intervention, whatever it does or doesn't do.

MikeM writes:

The Fed DID do something in the 20-21 recession, just not on purpose. It cut its discount rate as part of normal operations and that brought the economy back out of recession. The deflation that had been prevailing for a while stopped and unemployment began to decline.

It had no idea it was pursuing 'loose monetary policy' because the idea didn't exist at the time, but it did and it worked. The fascinating thing about the '21 recession wasn't that the free market got us out of it (it certainly did get a leg up from the Fed), it's that it didn't turn into a long, deep depression like happened ten years in the future. THIS was on account of the extreme flexibility displayed by wages and prices, something lacking in the aftermath of the '29 crash.

Vasile writes:

MikeM: The Fed DID do something in the 20-21 recession

Source? Bob Murphy is claiming just the opposite here Banks should rise prices in a Recession.

His source (from comments on his blog) is Alan Meltzer's "History of Federal reserve" (available on Google Books)

... and looking through it I found the following:

Annual Growth of Monetary Base

  • 1920: 3.6%
  • 1921: -15.5%

Yes, the monetary base growth in 1921 was negative.

Vasile writes:

And here is the link to Market Interest Rates in 1919-20. Not exactly accommodating I'd say.

Quite a fascinating read. See this for example, on page 117:

The NBER dates the trough of the business cycle at July 1921, four months before the activist policy began
baconbacon writes:

"Variant #1: Scott Sumner runs the Fed. He keeps nominal GDP growing at 5% per year. What happens? Unemployment still rises - and measured output falls - because mud pie workers have to look around for a better use of their skills. But output of non-mud-pies gradually rises as the productive 90% of the economy absorbs the cast-offs of the unproductive 10%."

Nice story- but as we well know expansion of the money supply doesn't happen evenly, that money is dispersed through agents (lets call them banks). These banks get the money first and have the first look at arbitrage situations- if the Fed's rates are lower than 5% then the banks can simply stick it in any economy imitating system that is earning the 5% and they gain the difference, for free, with no productivity added by their actions. Because they have added no productivity their profits represent simply a transfer of wealth and in this case its a transfer from the workers who don't get that money until later.

The story of the US economy as the Federal reserve has become more active follows this precise path. More and more people go into banking, more and more bankers go into non productive arbitrage, banking becomes a larger sector of the economy, the gap between rich and poor grows instead of falls. New bubbles are formed, in houses and stocks instead of mud pies, and must burst causing another drop in the interest rate by Mr Summers.

This last piece is an obvious, but rarely explored, piece of history. How did we get to a zero federal funds rate? We had a 20% rate in the 80s and every cycle has brought the peak and the trough lower, the graph of the FF rate is simply a sine wave of decreasing amplitude where the cyclical raising of the funds rate is less than the previous, recession busting, lowering.

fundamentalist writes:

There is no reason to accept Caplan's simple mud pie economy as anything resembling the real economy. It's no different than Krugman's baby sitting co-op. It's easy and trivial to come up with a simple story that gives your argument an overwhelming advantage and that's all Caplan has done.

In the first place, mud pies are a consumer good. Austrian econ says that bad investments take place in the capital goods industries, not consumer goods and that's a major point. In the second place, Caplan assumes that the Feds can raise or lower ngdp at will and at exactly the amount they want. They can't.

Scott Sumner writes:

Lee, I basically agree with your comment. One way to think of the rise in money demand is to recall that a real shock can weaken the economy, which depresses interest rates. Many think the Fed cut rates in 2008, but rates used to fall in recessions even before the Fed existed. There is less demand for credit in a recession. As interest rates fall to low levels, the demand for reserves and cash increase, especially when they start paying interest on reserves. The Fed needs to offset this, or NGDP will fall.

MikeM, Check vasile's data. The Fed actually caused the 1921 recession (which was short, but also one of the worst of the 20th century.) The Fed did nothing to help the economy recovery. Wages were cut, and that restored full employment. The best solution in principle would have been no deflation in 1921. But that would have required devaluation of the dollar, not politically acceptable at the time.

bacon bacon, Compare 1929, 1987, 2007, 2008. Burst bubbles cause modest damage (2007-08) or no damage (1987) to the broader economy when NGDP keeps growing. When NGDP is allowed to fall it causes a lot of damage (2008-09). And when NGDP falls in half (1929-33), a catastrophe.

The Fed could have propped up NGDP with even less money than they actually issued. They could have reduced money demand with an interest penalty on excess reserves.

BTW, my preferred policy would result in much higher interest rates, not lower as you imply. Low rates are an indication of tight money, not easy money, as Milton Friedman often pointed out.

Ryan Vann writes:

Scott Sumner,

First, I'd like to say I'm a big fan of yours, and love your Money Illusion blog (keep up the good work).

Anyway, I'm having a little problem conceptualizing your view. You said,

" preferred policy would result in much higher interest rates, not lower..."

Is that because you advocate an interest on reserve policy (banks will demand higher rates on loaned money to accomodate the imposed interest on reserves) or is there some other reasoning behind this? Perhaps phrased in a less muddled fashion, "Would an interest on reserve policy tend to increase interest rates?"

Joshua writes:

I confess I have never understood this. It's obviously possible for the government to move GDP around, because of the definition of GDP. But the argument that when demand for money rises, more money should be printed, doesn't make sense to me, unless we are to believe that the "demand for money" rises mysteriously due to animal spirits rather than for an actual reason.

Scott makes the point right here that when GDP falls during a burst bubble, things get worse than when it doesn't. But is the fall in GDP the cause, rather than a symptom? I wonder to what extent economists who believe it is the cause also invest heavily in the stock market.

And do we really think money demand can be reduced by raising taxes (or interest penalty, whatever) on "excess reserves"? Surely it's when money demand is elevated that the holders of those reserves will consider them *less* excessive, and therefore more crucial to preserve.

Joey Donuts writes:

Lee Kelly:

I think you are correct about the increase in demand for money but I think you have the wrong cause.

If you look at the difference in the price of oil between Nov. of 2007 and Mar of 2009 from $65 a barrel (the price for several months preceeding Nov, 2007) You will see that 240 billion dollars went from accounts with higher propensities to spend to accounts with lower propensities to spend. This was just in the US world wide the shift is larger. Basically from oil consumer accounts to oil producer accounts.

The Fed during the summer of 2008 were net sellers in its open market operations. Why because banks had begun to borrow heavily at the discount window (I contend because Oil Producers were drawing down accounts and investing in short term securities) The Fed saw, according to their own statements, an unusual increase in demand for US gov't securities and in order to make their target federal funds rate sold securities. This also put a strain on the banks capital requirements.

If the banks hadn't had so much junk on their balance sheets, this shock to the monetary system wouldn't have been seen as such a big crisis because the banks could have met the increase in demand for US securities by selling off some the banks holdings of US securities.

The oil price shock is, I believe, the single most important cause of this recession. The price shock was equivalent to an increase in taxes of $240 billion with no increase in expenditures by the recipients of the higher oil revenues.

baconbacon writes:

"bacon bacon, Compare 1929, 1987, 2007, 2008. Burst bubbles cause modest damage (2007-08) or no damage (1987) to the broader economy when NGDP keeps growing. When NGDP is allowed to fall it causes a lot of damage (2008-09). And when NGDP falls in half (1929-33), a catastrophe."

There are two main problems with this statement, the bigger of the two seems to be that you are ignoring the current position we are in. We burst a bubble in 2001, drop the federal funds rate from 6.5 to 1, fuel a new bubble- raise interest rates to 5.25 and burst that bubble. Then we drop rates to effectively 0 and increase the federal reserves balance sheet 2.5 times. Now just about every fundamentally driven investor says that equities are overvalued, Robert Shiller recently says that areas of the US housing market are back in "bubble territory", the US is paying historical lows on its debt despite issuing historical highs in debt- basically we have at least three bubbles forming right now.

What can the federal reserve do now? US debt has an averaged maturity of 50 months (and falling) if it raises interest rates the cost of servicing US debt and hence the deficit will spike. Banks balance sheets are still in precarious positions, forcing them to take back the crap they dumped on the fed (literally stocks of now bankrupt companies) would drive many of them bankrupt. The US housing market is buoyed by low interest rates (and $8,000 in free money) without which housing prices will plummet.

Any attempt to raise rates will burst the bubbles that are forming now. Any delay in raising rates will push the bubbles larger and increase the problem until rates are raised, while likely continuing to weaken the dollar.

So just like we put off problems from the 2001 crash with a bubble we are trying to put the 2008 crash off with a series of other bubbles. Only we have to revert to more and more desperate means of re-inflating each time. To claim that the 2008 crash wasn't as bad as the 1929 crash because we sunk ourselves in an additional 2-10 trillion dollar hole before that scenario plays out is woefully misguided.

Ryan Vann writes:


You asked
"And do we really think money demand can be reduced by raising taxes (or interest penalty, whatever) on "excess reserves"?

I was under the impression that the goal of charging interest on reserves is not to stimulate money demand, but increase money supply. It seems to me, that money demand, associated with production, during economic slow downs would be high (if not just to avoid liquidation).

The problem is that banks are much more hesitant to lend in the face of uncertainty. By charging interest on reserves, it provides a bit of disincentive for banks to hold on to money.

I could be entirely mistaken though.

steve writes:

Bryan, you seem to be proposing that a fed which caused massive deflation in the face of voluntarily chosen wage rigidies would cause disruptions and unemployment not related to any real world disruptions otherwise requiring people to alter their behavior. I don't see any problem with this statement. You are essentially saying that the fed decreed an increase in wages and the value of current savings. Naturally adjustments will follow.

But, this is hardly the same thing as deflation caused by an increase in the demand for money. If you interpret this strawman argument as saying the fed should increase the money supply in the face of increased demand for money, then you are essentially saying that the consumers desire to save purchasing power should be defeated by erroding the purchasing power of the money they save.

Would this increased demand for money cause disruptions and unemployment in the face of voluntarily chosen wage rigidies? Again, I think it would for those who made that choice, in defiance of reality, just like an entreprenour suffers a loss when they misjudge reality.

If the fed increases the money supply in the face of increased demand for money, they are attempting to defeat the savers in their attempts to save and fool those who demand wage rigidity into accepting lower wages.

In other words, the fed proposes to 'fix' the increased demand for savings by sowing confusion and opposing adjustments to the economy demanded by the savers.

steve writes:

What if the increased demand for money is not from consumers attempting to save purchasing power, but rather to pay off mortgages from a popped bubble? What should the Fed do to fix a previous error?

All I can say to that is "Oh what a tangled web we weive when at first we attempt to deceive."

Vasile writes:

Scott Sumner:

Compare 1929, 1987, 2007, 2008...

Scott, are you sure, that your models... aren't cheating on you? You know, like they did with Philips? They showed him a likeable curve, which later turned out to be... quite different from the original one.

A more serious objection would that you kind of dismiss both the 1920-21 depression and Volcker depression, where in both cases the policy was to pay no attention at all to NGDP. With rather satisfactory results.

Here is what (quoting Alan Meltzer) ended the F(orgotten)D(epression)

Falling prices raised real balances [...] The public used its increase in money balances to purchase goods and assets

Supposing that to yours, (Bryan Caplan's, ..) proposals would have been given way during FD, then there would have no price deflation, and no increase in real money balances. With the public less eager to purchase goods and assets.

Vasile writes:


Steve, that's exactly it. Demand for money is demand for delayed consumption. A shift in preferences which favours future consumption at the expense of current consumption.

Of course, lowering the interest rate (the cost of borrowing) is making current consumption cheaper. Not sure, how much it can help. More during normal/boom time, less so in recession time.

And of course, price inflation makes future consumption costlier. Sometimes, a lot more so. In this way, it can shift consumption into present time, or in extreme cases into hoarding of other (non-monetary) goods. Paintings. Old violins. Physical gold.

So yes, one can (to some extent) influence demand for money, the relative prices of future vs. current consumption.

But.. It is not free of collateral damage, and... Why should one advocate a politics of price fixing?

MikeM writes:

Scott: I have the same data. I'm aware of the Fed causing the '21 recession (the Fed jerking the interest rate around for various reasons causes MOST recessions), I'm just pointing out that when they dropped their discount rate in '21, that's what brought us out of it.

You can't think of any post-Fed economy operating under an entirely 'free market' regime, although it closely approximates one at times. Since interest rate changes in this period are primarily driven by the Fed's discount policy, rather than market forces, there really isn't much free market going on. Like I said before, the impressive thing about the '21 recession was the extreme wage and price flexibility. Had the discount rate not been lowered the price deflation would have continued and so would have wage deflation because the money supply would have continued collapsing.

Doc Merlin writes:


The modern world has much less price and wage flexibility. Seems to me mostly due to benefits mandates and high payroll taxes. In the 30's the lack of flexibility this was mostly due to price and wage fixing.

Could our current unemployment crisis be in large part due to anticipated increases in the fixed costs of hiring from medical mandates?

Comments for this entry have been closed
Return to top