Arnold Kling  

Average Prices, AS and AD

PRINT
Response on Means Testing... The Plight of the Unskilled Co...

Scott Sumner writes,


The recession was caused by a severe AD shock, i.e. falling NGDP. If it had been caused by an AS shock then the recession would have been accompanied by higher inflation. It would really help if people started to think more in terms of NGDP.

For Scott, nominal gross domestic product (NGDP) is fundamental. To me, it is accidental. To me, what is fundamental is real output. Then we have prices for different inputs and outputs, which are set relative to one another, but which are expressed in monetary units. Finally, we have the average price level, which is an index of average prices expressed in monetary units.

I view the money supply as a very wiggly lever for controlling average prices. If the monetary authority pushes really hard for really long, it can get average prices to rise quickly. Then, if it wants to get prices to stop rising quickly, it has to pull really hard for really long in the other direction.

I do not think in terms of aggregate demand and aggregate supply. In fact, I have a very hard time teaching them in my high school economics class, because they make so little sense.

The story of aggregate supply is as follows: start with sticky nominal wages, and with stupid workers willing to supply unlimited amounts of labor at whatever real wage rate happens to prevail. Then if average prices go up, average real wages go down and labor demand goes up. Because workers are stupid, they supply the additional labor, and output goes up. Hence, the aggregate supply relationship--as prices go up, the willingness to supply output goes up.

The (Keynesian) story of aggregate demand is as follows: start with a fixed supply of money, and with stupid investors who think that the short-term nominal interest rate in the Fed funds market is a proxy for all interest rates, including real long-term rates. Then, when average prices rise, the ratio of the money supply to the price level falls, creating an excess demand for money, leading to a rise in short-term nominal interest rates, resulting in a rise in long-term real interest rates, resulting in declines in spending on long-lived assets, such as housing. Hence, the aggregate demand relationship--as prices go up, the willingness to demand output goes down.

The beauty of this framework is that you can explain anything. Did prices and output both go up? Obviously, aggregate demand must have shifted right, causing upward movement along the aggregate supply curve. Did prices go up while output went down? Obviously, aggregate supply must have shifted left, causing upward movement along the aggregate demand curve.

The virtue of the Recalculation Story is that it gets away from this AS, AD framework. It explains the recession as arising from frictions in reallocating resources away from an unsustainable industry structure. I would agree with Scott that the low rate of inflation during the recession is a sign that the Fed could have tried wiggling its lever harder to keep prices higher. However, my conjecture is that if Fed had pulled its lever really hard what we would have had was mostly higher inflation, with little or no improvement in output.


Comments and Sharing


CATEGORIES: Macroeconomics



COMMENTS (11 to date)
Doc Merlin writes:

"To me, it is accidental. To me, what is fundamental is real output."

By saying this you are missing something really important. Nominal goods are connected to real goods through contracts and debts. This means that changes in NGDP are very important.

MV = NGDP
So whenever NGDP drops unexpectedly, people are on average, less able to pay their debts. So the people who were marginal payers default, which lowers asset prices. This is in addition to the drop in prices from lower nominal demand.

You also say that the "tried wiggling its lever harder to keep prices higher." This misses the point entirely. Its not prices that are important, but debt and contracts.

Because of debt, law, and contract enforcement, NGDP shocks (i.e.: fed policy mistakes) have very real effects. This isn't to say that the recalculation story isn't important, it very much is. However, unexpected adverse NGDP shocks are pretty bad too.

AC writes:

Arnold,

I tend to agree with you, I am wondering -- do you think there is any traction at all from fiscal policy during a deep recession (assuming of course it's not perfectly targeted to specific idle resources)

By the way, maybe it's my reader but in my RSS feed I get no indication there is more below the fold

Mark writes:

"By saying this you are missing something really important. Nominal goods are connected to real goods through contracts and debts. This means that changes in NGDP are very important."

Isn't the value of any asset the nominal cash flows discounted by a nominal rate? It would see to me that pushing up nominal activity has an unclear impact on asset prices and therefore debt service. If inflation gets pushed more so than real activity asset prices fall and vice versa.

I suppose if the Fed is able to keep real rates at the front of the curve artifically low and someone is willing to borrow short and invest long, assets could be overpriced for some period of time (bubble?) but that would eventually reverse. Feels like our housing experience.

fundamentalist writes:

"The recession was caused by a severe AD shock,"

Remove the jargon, and all Mr. Sumner has said is "&*% happens!"

Does Sumner really believe that AD shocks are totally random events? Isn't he even a little bit curious about why the shock happened? Or does he assume that the shock wasn't random but instead was a case of the Feds falling asleep at the wheel?

Noah Yetter writes:

For Scott, nominal gross domestic product (NGDP) is fundamental. To me, it is accidental. To me, what is fundamental is real output. Then we have prices for different inputs and outputs, which are set relative to one another, but which are expressed in monetary units. Finally, we have the average price level, which is an index of average prices expressed in monetary units.

YES YES YES! This is what kills me when I try to read Mr. Sumner's posts. He stubbornly treats NGDP as if it were a real thing (real in the metaphysical rather than economic sense) that exists, that can be observed and measured, even altered. It is not. It is a deeply flawed statistical aggregation of billions of billions of actual facts. It changes only when and because the real phenomena it describes change. Changes in it cannot cause changes in anything else.

Scott Sumner writes:

Noah, It is exactly the opposite. NGDP is easily measured, RGDP is a figment of the imagination of statisticians working in Washington. Take a Dell computer that sells for $800. Two year later a model comes out that sells for $900. NGDP (for that specific machine) rises 12.5%. No guesswork at all. Now consider RGDP. A statistician looks at the new machine and says; Hmmm, I think that new Dell computer is 43% niftier than the old one. On what basis does he make this judgment? Who knows, there is no objective measure of the utility derived from computers. So he arbitrarily says RGDP went up 43% for that machine. Of course overall NGDP and RGDP for the entire economy involve doing the same thing for all goods. But with NGDP you are at least imputing something definite, what consumers were willing to pay for each good. We all agree on the NGDP for a Dell computer. For RGDP you are just imputing a bunch of guesses from Washington bureaucrats who haven't a clue how much more utility one machine provides than the next. RGDP is complete guesswork.

fundamentalist, Of course I'm very curious why AD shocks happen, I've got dozens of posts discussing the flaws with current Fed procedures.

happyjuggler0 writes:

Arnold Kling,

"I view the money supply as a very wiggly lever for controlling average prices. If the monetary authority pushes really hard for really long, it can get average prices to rise quickly. Then, if it wants to get prices to stop rising quickly, it has to pull really hard for really long in the other direction."

What happens if the Fed keeps M steady as she goes, but velocity of money *suddenly* falls, perhaps because of uncertainty or friction during a sectoral great recalculation resulting in a sudden greater demand on the part of banks, businesses and individuals to hold actual cash on the sideline?

What happens then is that you get an "unexpected" fall in nominal demand. If your business investment decisions, or your personal bill paying assumptions, are based in inflation adjusted calculations, where does that leave your "real" bottom line? Down is where it leads. Nominal demand shocks matter in real terms!

Now imagine that the Fed didn't keep M steady when this massive desire to hoard cash happened. Imagine instead that the Fed injected new money into the economy, in an amount that precisely offset the reduction in V, keeping MV stable, and thus keeping AD (or NGDP) stable. Now what happens to all those "real" collective decisions made with nominal expectations (and calculations) in mind? They no longer go kaput, and there is no longer a "secondary depression" accompanying the single sector great recalculation that caused the demand to hold cash to soar.

In my understanding, NGDP doesn't matter per se, is is sudden changes in NGDP expectations that matter dearly in a sticky world.

From the perspective of an individual business, is there a difference between a drop in NGDP because of a drop in RGDP, or a drop in NGDP because of a drop in inflation? Either way, you take a *real* hit in your bottom line.

Sumneromics is in effect about floating the economy some cash when the demand to hold cash suddenly soars, and withdrawing that float when when demand to hold cash suddenly plummets. Nominal smoothing thus results in RGDP smoothing, avoiding pointless real economy dislocations that occur from central banks not understanding the importance of NGDP.

"The story of aggregate supply is as follows: [...]Because workers are stupid, they supply the additional labor, and output goes up.[...]"

Some people would say workers pay too much attention to nominal wages because they are crazy, not because they are stupid. I believe Tyler Cowen says pretty much the same thing in his latest book.

"The (Keynesian) story of aggregate demand is as follows: start with a fixed supply of money, and with stupid investors who think [...]"

Isn't this too complicated? What about simply adding the demand functions of individual goods?

Gregor Bush writes:

Arnold,
I've enjoyed your debates with Sumner over the last year but I'm much more in agreement with his ‘collapse in AD’ view than your ‘Recalculation Story. The former just seems to fit the data much better. Consider an obscure industry: Japanese ceramics. If you look at the production data for Japan's ceramics industry, it's a dead flat line from 2002 until September of 2008. Output never rises or falls by more than one or two present in a month and any small declines in production are quickly reversed. Then from October of 2008 to February 2009 production plummeted by 30%. Since March of 2009 output has increased by 20%, albeit from a lower base.
Was output of Japanese ceramics really at unsustainable highs? And was the drop off in production caused by a sudden realization by investors that capital and labour had to be reallocated to other sectors. If so, why has production rebounded so quickly? Production of Korean electronic components id an even more striking example: it fell 41% from September to December of 2008 and then rose 925 from January to July of 2009.
There are many industries in countries all over the world that have production charts which correlate almost perfectly with the inverse of US credit spreads. This seems more consistent a collapse demand caused initially by a scramble to conserve cash (and resulting spike in the cost of capital) and compounded by a dramatic weakening in the outlook for sales over the next two years (the realization that the policy response would not be powerful enough to prevent such a decline). The rebound in these industries has been consistent with a decline in the cost of capital and an improved outlook for sales. There was never any ‘need’ for Korean electronics production to collapse.
I think the only substantial ‘recalculation’ was a shift of resources out of residential construction in the US. But that had already been underway for 30 months prior to September of 2008.

Noah Yetter writes:

Scott, you are totally wrong, and I'll show you why. Take the $800 Dell computer and add a $1 Avocado to it. Now what do you have? Eight hundred and one dollars worth of making stuff up. You cannot add dissimilar things together and get wisdom. Yes the dollars can be added together but dollars do not matter. Money is a claim on real resources and those are what matter.

Your argument against "real" GDP statistics is certainly correct, but the flaw is not in the "realness" it is in the GDP. GDP is not a thing. It is emergent. You cannot measure it. There is no such thing as a shock to it. Policy cannot control it.

Noah Yetter writes:

Now imagine that the Fed didn't keep M steady when this massive desire to hoard cash happened. Imagine instead that the Fed injected new money into the economy, in an amount that precisely offset the reduction in V, keeping MV stable, and thus keeping AD (or NGDP) stable.

Dear god man, listen to yourself. This is the kind of sloppy thinking that results from over-reliance on aggregates and a failure to think about what they represent.

First let's make the ultra-heroic assumptions that 1) we know what V is (and was), and 2) we know what M is, so that we know how much additional M we need to keep MV stable. Given that, HOW would the Fed increase M? What policy lever do they pull to do that? Which actual people do which actual things to make that happen? Not statistics, actual observable events that occur in the actual observable universe. We all know that what the Fed actually does is conduct open market operations. But that doesn't just "increase M" what it does is increase the M that's held by banks. Do you really and honestly think that if the Fed figured past and current V, figured current M, solved, and sent a huge pile of fresh Benjamins to Goldman Sachs that it would stop or reverse a recession?

We joke about money-dropping helicopters because we all know the concept doesn't work. There is no way in the real world to just increase M. The new money has to have an entry point. It has to be lent or spent by actual people on actual goods in a process that's not too politically objectionable. And that takes time, and causes distortions.

And that's all disregarding the utter silliness that is "everyone suddenly wants to hold more cash" story. Someone who loses their job is not suffering a bout of animal spirits restraining their spending, they are suffering a shock to income. The tech bubble and the housing bubble didn't pop because people suddenly and randomly stopped spending, they popped because investments in things no one wanted to buy finally became unprofitable, resulting in income shocks as malinvestments were realized and liquidated.

When comparing "AD shock" and "Malinvestment/Recalculation" explanations of cycles, the latter have the advantage of being TRUE. All you have to do is put down the statistics and look out a window.

Comments for this entry have been closed
Return to top