Robert Samuelson: Inflation Terminology and the Upside of Recessions

I'm afraid I am going to have to take issue with Robert Samuelson on both of his main themes. He says:

The Upside of Recession?, by Robert J. Samuelson, Commentary, Washington Post: It's increasingly clear that much of our standard economic vocabulary needs revising, supplementing or at least explaining. The customary words we use don't fully convey what's happening in the real world. Let me illustrate with two basic economic terms: inflation and recession. There are also larger lessons.

Start with inflation. ... We all know about oil. Prices are about $60 a barrel. They seem unlikely to return to $28, the 2000 level. The real surprise involves food prices. In the past three months, they've risen at a 7 percent annual rate. We may be seeing the first adverse effects of the ethanol boom. Corn is a main feed grain for poultry, cattle and hogs. Corn is also the main raw material for ethanol... Competition for grain has pushed up corn prices..., almost double a typical level. High feed prices have discouraged meat producers from expanding. The resulting tight meat supplies raise retail prices.

"Poultry is the best example," says economist Tom Jackson of Global Insight. "In the past 40 years, we almost never have year-to-year decreases in production. In the past few months, we've seen production go down." ...

So the government's subsidies for corn-based ethanol are worsening inflation, perhaps permanently. ...

Now switch to recession. ... We've been conditioned to think of recessions as automatically undesirable. The labeling is simplistic.

Hardly anyone likes what happens in a recession. Unemployment rises, production falls, profits weaken, stocks retreat. But the obvious drawbacks blind us to collateral benefits. Downturns check inflation -- it's harder to increase wages and prices -- and low inflation has proved crucial to long-term prosperity. Downturns also punish and deter wasteful speculation. When people begin to believe that an economic boom won't ever end, they start to take foolish risks. Partly, that explains the high-tech and stock bubbles of the late 1990s and, possibly, the recent housing bubble.

Some sort of a recession might also reduce the gargantuan U.S. trade deficit... Almost everyone believes that the U.S. and world economies would be healthier if Americans consumed less, imported less, saved more and exported more. The corollary is that Europe, Japan, China and the rest of Asia would rely more on domestic spending -- their own citizens buying more -- and less on exports.

Ideally, this massive switch would occur silently and smoothly. Realistically, the transition might not be so placid. ... Almost no one wishes for a recession, but the consequences might not be all bad. The larger lessons here involve perceptions. Our regular vocabulary often fails to describe the complexities of a changing economy. We must be alert to new possibilities. Things are not always what they seem.

Let's start with what he calls inflation. His point was about the language of core versus overall inflation and he tries to question that usage, but as he notes:

Since the early 1980s, the two indexes (the CPI and the core CPI) have recorded -- despite many monthly differences -- virtually identical increases.

So he rebuts his own point effectively and I left that part out (though he should be talking about the overall and trimmed mean PCE instead of the CPI and core CPI). The rest of his point is just speculation about core inflation and not worth addressing since I want to move on to his point about economic terminology, in this case his use of the term inflation which isn't technically correct.

I'll use some hypothetical numbers for illustration. Suppose that when the price of oil is $28, the price level is 100, and also suppose that when the price of oil increases to $60, the price level increases to 120 all else equal (again, these are not intended to be realistic numbers).

Now, let the spike in oil prices happen quickly, but due to sluggish wage and price adjustment suppose the resulting 20% increase in the price level takes more time, say 2 years. During this two year period, as the price level rises from 100 to 120, inflation will be reported in the news.

But this is not what economists mean by inflation when they say, for example, that "inflation is always and everywhere a monetary phenomena." To see this, suppose that the change in oil prices and the price level are both instantaneous rather than having the change in the price level drawn out over two years as before. That is, the price level jumps from 100 to 120 instantaneously and stays at the higher level from then on.

In this case, there is no inflation. Prices were stable before the instantaneous jump in the price level, and prices are stable afterward. From that day forward prices remain at 120 and do not increase any further. There is no inflation.

When the change in the price level is drawn out over two years it doesn't really change anything except that the change is no longer instantaneous and hence looks like an ongoing inflation. That is, it looks just like (and is easily confused with) the continual rise in prices that occurs with a continual increase in the money supply beyond what is needed to accommodate economic growth. But at some point, after two years in the example here, the price level will stabilize at 120 and after that there will be no further change in prices and no inflation.

Whether the change is instantaneous or drawn out, it is technically a change in the price level, not inflation, though economists (myself included) are not very careful when they talk about inflation to distinguish changes in prices that reflect drawn out adjustment to factors such as increases in oil prices (which eventually end) and those that arise from excess growth in liquidity (which do not necessarily end).

Now to Samuelson in particular. First, he calls the change in the price level resulting from a change in the price of oil from $28 to $60 inflation. But as just explained, even if the oil price change causes a drawn out adjustment in the price level that looks just like inflation, the increase in the price level will end once it reaches the higher level and there will be no inflation after that. These kinds of changes where the price of oil moves from one level to another do not cause permanent changes in the inflation rate, only temporary changes during the adjustment period.

He makes the same mistake with subsidies. It's just not possible that "the government's subsidies for corn-based ethanol are worsening inflation, perhaps permanently" as Samuelson claims. A one time increase in subsidies could change relative prices and fuel a short-term inflation, but it cannot fuel a long-term inflation (but it may not change inflation at all - see point three).

Second, looking at individual sectors and noting whether their prices or quantities have gone up or down is not how we assess inflation, that is done by looking at aggregate demand. When there is an oil price shock, it will ripple through the economy and cause a general realignment of prices, i.e. general changes in relative prices. Some relative prices will move such that resources are drawn into the sector, while other will move such that resources move away, so you don't learn much from looking at just a few sectors.

The distributional consequences and potential hardships for lower income households from changing relative prices are certainly worthy of our attention. But learning that the demand for corn is up and the production of chicken is down doesn't necessarily inform us about what is happening to the average of all prices in the economy (and there are, of course, many other confounding factors that affect sectoral relationships). Again, to assess pressures on prices generally you look at the aggregate level - i.e. how aggregate demand and aggregate supply are affected and how they are likely to move in the future. Now it is true that, all else equal, an increase in the price of oil will increase the price level, but you don't look (or need to look) at the price of chicken to figure that out.

Third, subsidies to ethanol do not necessarily increase aggregate demand (and hence inflation) as Samuelson claims, the subsidies have to be paid for somehow and that reduces demand in other areas. They could be paid through deficit spending and the new spending would affect aggregate demand, but in any case it's the aggregate effect that matters and it's not necessarily the case that subsidies increase demand.

More generally, a long-term, 10 or 20 year inflation cannot be driven by increases in government spending and cuts in taxes (which includes increased subsidies). The increase in government spending that would be required to fuel such an increase in aggregate demand and inflation would eventually eat up all of GDP, and the tax cuts that would be required would quickly run into a lower bound, zero in the limiting case, and after that there could be no further stimulus and prices would stabilize. Long-run inflations are driven by increases in the money supply because the money supply, unlike increasing government spending or decreasing taxes, can be increased pretty much without bound and hence sustained in the long-run if a government decides to permanently turn up the printing presses. [Update: More on inflation here.]

Okay, next let's turn to the silliness about the virtues of recessions. Samuelson says:

We've been conditioned to think of recessions as automatically undesirable. The labeling is simplistic.

It's not the labeling that is simplistic here. He also says there are hidden virtues to recessions:

Hardly anyone likes what happens in a recession. Unemployment rises, production falls, profits weaken, stocks retreat. But the obvious drawbacks blind us to collateral benefits.

And one of the main benefits he cites is that it punishes speculators for taking too much risk, i.e. that:

Downturns ... punish and deter wasteful speculation.

This is what Krugman calls the Hangover Theory. I've posted this before, but it applies again:

Powerful as these seductions may be, they must be resisted--for the hangover theory is disastrously wrongheaded. ... The many variants of the hangover theory all go something like this: In the beginning, an investment boom gets out of hand. Maybe excessive money creation or reckless bank lending drives it, maybe it is simply a matter of irrational exuberance on the part of entrepreneurs. Whatever the reason, all that investment leads to the creation of too much capacity--of factories that cannot find markets, of office buildings that cannot find tenants. Since construction projects take time to complete, however, the boom can proceed for a while before its unsoundness becomes apparent. Eventually, however, reality strikes--investors go bust and investment spending collapses. The result is a slump whose depth is in proportion to the previous excesses. Moreover, that slump is part of the necessary healing process...

Sounds just like Samuelson's description of punishment for the excesses that caused the high-tech and housing booms, and the subsequent healing process:

Downturns check inflation ... and low inflation has proved crucial to long-term prosperity. Downturns also punish and deter wasteful speculation. When people begin to believe that an economic boom won't ever end, they start to take foolish risks. Partly, that explains the high-tech and stock bubbles of the late 1990s and, possibly, the recent housing bubble. Some sort of a recession might ... reduce the gargantuan U.S. trade deficit... Almost everyone believes that the U.S. and world economies would be healthier if Americans consumed less, imported less, saved more and exported more.

So what's wrong with this? Back to Krugman:

Except for that last bit about the virtues of recessions, this is not a bad story about investment cycles. Anyone who has watched the ups and downs of, say, Boston's real estate market over the past 20 years can tell you that episodes in which over-optimism and overbuilding are followed by a bleary-eyed morning after are very much a part of real life.

But ... nobody has managed to explain why bad investments in the past require the unemployment of good workers in the present. Yet the theory has powerful emotional appeal. Usually that appeal is strongest for conservatives... But moderates and liberals are not immune to the theory's seductive charms--especially when it gives them a chance to lecture others on their failings. ...

The point is that Samuelson implies unemployment is the necessary cost of attaining this virtue of reigning in speculators, curing the trade deficit, lowering inflation, and so on. But why do workers have to pay this cost through higher unemployment? Why shouldn't the government step in and stimulate employment and the use of idle capacity rather than having the idle workers watch the idle capital creatively rot and destruct? We can build new capital while the old capital is still operating, there is no requirement that old capital be destroyed before new capital can be "creatively constructed." There is simply no need to let workers pay the costs of punishing and deterring wasteful speculation, let the speculators pay those costs themselves.

Samuelson says in his first sentence that:

It's increasingly clear that much of our standard economic vocabulary needs revising, supplementing or at least explaining.

And he's right. He's made a pretty good case that there's confusion about these basic economic terms.

Comments

[reason]

Mark,
while I largely agree with you, some points perhaps need to be more clearly expressed:
1. What we are seeing today are large changes in relative prices. These make it very hard to interpret general price indices. In particular, the effective rate of general inflation can be very different for different groups of people.
2. moral hazard during the expansion is an important issue.
3. The depth of a recession IS related to excesses to excesses that preceed it. How we combat a deep recession is another issue. International effects can offset domestic fiscal and monetary and make recovery more difficult.
4. The differential effects of a recession on different people (not related to culpability) is a major cause for concern (see point 1).

[Alex R]

As a non-economist, I treat the word "inflation" as being more-or-less synonymous with "rate of change of price level". If economists use the word inflation to mean something different -- maybe something like the "rate of change of the number of dollars of demand per unit of goods" I would be interested in seeing that explicated for non-economists, with an emphasis on how it is actually measured and how it differs from the rate of change of prices...

[knzn]

More generally, a long-term, 10 or 20 year inflation cannot be driven by increases in government spending and cuts in taxes (which includes increased subsidies). The increase in government spending that would be required to fuel such an increase in aggregate demand and inflation would eventually eat up all of GDP, and the tax cuts that would be required would quickly run into a lower bound, zero in the limiting case, and after that there could be no further stimulus and prices would stabilize.

I disagree with this. The implicit assumption you’re making is that there is always a new equilibrium, but in general, there won’t always be one. Consider, as a stylized example, a closed economy where government spending is fixed in real terms, consumption depends on disposable income, investment is insensitive to the interest rate, the economy is at full employment (and initially at equilibrium), and taxes are proportional. Now suppose the government cuts the tax rate. If you solve this as a static model, there is no equilibrium after the tax cut, so prices will rise without bound. If the government wishes to avoid hyperinflation, it will eventually be forced to raise taxes or reduce spending, but there is no reason to presume that it would do this (at least initially) in a way that results in a new equilibrium. Rather, it may simply keep chasing the dynamic adjustment path fast enough to keep the inflation rate from rising above a certain point. And surely the process could take 10 or 20 years before the government finally gets religion.

[Barry]

The larger issue is that people like Samuelson aren't really impacted by recessions. If columnists, pundits and econ professors lost their jobs, they'd see things a bit more clearly. As it is, it's one step below talking about how the peasant population needs to be 'trimmed' by a food shortage.

[Callahan]

To Barry, very well said. My sentiments exactly. Tired of rich talking heads smiling as they "report" on such issues.

[Winslow R.]

Long-run inflations are driven by increases in the money supply because the money supply, unlike increasing government spending or decreasing taxes, can be increased pretty much without bound and hence sustained in the long-run if a government decides to permanently turn up the printing presses.

I agree long-run inflations in the price level are driven by increases in the money supply above the growth in savings and GDP.

What limits 'increasing government spending'? I see no bound except for a political bound.

The 'printing presses' are currently limited by the amount of treasury securities available.

[William Smith]

I've learned to ignore Samuelson. He reads less like an economist and more like a journalist who accidentally attended an economics seminar in college.

[Bruce Wilder]

Like Adam R, I question whether you have made the crisp, and valid, distinction you think you have made.

If the price level changed from 100 to 120, that's inflation.

Relative price changes can happen without changing the overall price level, and if such relative price changes occur without changing the overall price level, then there is no inflation.

Inflation is a "monetary phenomenon" because inflation is a decrease in the value of money relative to all goods in market exchange (measured by the prices in exchange of a suitably weighted sample basket of goods).

As a matter of policy, it is arguably wiser to accomodate the rising cost (increasing relative price) of a basic commodity within a general inflation, but that's a whole other topic.

[Sonia]

There are weighting issues here. A relative increase in food and fuel prices will cause some decline in consumption of fuel and shifting to lower cost calories and to lower price other stuff or less of it (in a few cases to suboptimal calorie consumption but that is rare in this country). The CPI will not pick these shifts immediately and the CPI will overstate inflation. The upward adjustment in the CPI will increase Social Security payments and federal pensions (and in the past when many wages were indexed these as well) shifting some real income to beneficiaries of these payments and away from heavy users of fuel and food. Because prices may be inflexible downward there may be some drop in employment in the sectors that lose consumption. The effect of changes in price levels on employment may be somewhat different in the current world with large international flows than previously. Consider construction, a substantial portion of the new employment in construction was immigrant (often unducumented) labor. A drop in construction may cause less of drop in measured employement than previously.

[kharris]

Samuelson labors under the twin burdens of having to write frequently on economic issues and being "not Paul". As a result, he tends to be quite shallow.

The complaint he levels just repeats, in reasonably convincing prose, arguments heard from barstools around the country. "Inflation is what I say it is" and "recessions are needed to restore economic health".

The first view is at the heart of much of what passes for educated discussion on the internet and elsewhere. Disagreements over definition are tedious and unproductive, but often intentional as an effort to dictate terms of the debate. If inflation is what Samuelson says it is, then Samuelson's point is secure. So he redefines inflation.

"Recession is needed to retore economic health" confuses the period of rapid growth needed to return to trend with some greater miracle, or it is a plea to get my gardeners and house painters to show up on time and stop asking for more money. The point about the trade balance is correct, but if what we are most interested in is welfare, then restoring balance by degrading welfare seems an odd goal.

[Summary]

Okay, so the fed is only trying to prevent rapid general price increases due to excess money growth. It doesn't care about relative price increases due to other causes. The price of food could double due to gov regulations, and this would have no effect on fed policy. The fed has no mandate to keep the cost of essentials affordable.

[Detlef]

Now, let the spike in oil prices happen quickly, but due to sluggish wage and price adjustment suppose the resulting 20% increase in the price level takes more time, say 2 years. During this two year period, as the price level rises from 100 to 120, inflation will be reported in the news.

But this is not what economists mean by inflation when they say, for example, that "inflation is always and everywhere a monetary phenomena."

Well, it seems that this is exactly the way the European Central Bank sees it?

(Sept. 2005 statement to the European parliament)
He warned that further oil price hikes could lead to higher wage and price demands, which could push inflation up.

Mr Trichet told the European Parliament's economic and monetary committee that oil prices will drive euro-zone inflation up more than previously expected. The ECB sees annual inflation of between 2.1% to 2.3% in 2005 and 1.4% to 2.4% in 2006.

He said that the main danger of higher inflation comes from "potential second-round effects in wage and price-setting behaviour triggered by ongoing oil prices."

[Winslow R.]

One other thing...

and the tax cuts that would be required would quickly run into a lower bound, zero in the limiting case, and after that there could be no further stimulus and prices would stabilize.

not true as you forget negative taxes, including economist's favorite the EITC. Something I also tend to forget :)

[calmo]

Thanks for updating us about what goes on in those bars Kharris and I agree with a tag like Samuelson, Robert has his work cut out for him trying to live up to expectations...sorta like Elmo Einstein who has an opinion too in this great democracy...where WaPo can give even Elmo a regular spot.
The recession he urges on...I got this far before ordering another stiff one:

The labeling is simplistic.

Hardly anyone likes what happens in a recession.

thinking this (labeling) was his intro for building his case for Robert Samuelson, the Recesssion Hero: the only man who likes what happens in a recession...because he knows what the collateral benefits are.
Well forgive my labeling but this does not rise to the level of Plonking Type I and people here are not up to entertaining Elmo and where he strayed from those "sensible remarks".

[Dave Iverson]

Krugman on "the silliness about the virtues of recessions" : "...nobody has managed to explain why bad investments in the past require the unemployment of good workers in the present."

Although I don't like to talk of "virtues of recessions", and certainly not of depressions, I'm not convinced the Krugman's thesis is correct.

The problem seems to be one of trust. How are government authorities (or anyone else) going to get the system moving again immediately in the wake of collapse? And if they could, how are they going to overcome the political psychological problems of taking action? Usually a time lag is necessary both to quell the animal spirits of speculative frenzy and to allay the fears of people burned by any recent collapse.

Who is to lend what to whom? Why Would they dare? What many writers are unwilling to entertain is the lag time necessary to induce either bankers to loan money or credible businesses to borrow money after bubbles burst.

A time lag is needed in order to allow the psychological cloud of pessimism to move far enough into the backdrop to once again begin the cycle of guarded optimism and trust that a better day is possible.

From my, Can Minsky Respond to Krugman's criticism of 'The Hangover Theory'?

[2slugbaits]

Mark,

Wouldn't your definition of inflation suggest that we should look to the broadest measures of inflation (e.g., GDP deflator) rather than things like typical shopping baskets for urban consumers?

Also, suppose most of the commodity price increase is due to high import prices and two countries trade at a fixed exchange rate, so the money supply does not directly affect the price of imports. Under your definition of inflation, should those imported goods be counted in the inflation index?

[James Killus]

When economists say, "inflation is always and everywhere a monetary phenomena," they are engaing in tautology, having defined "inflation" as a "monetary phenomenon" and "non-monetary changes in in the general price level" as something else.

In the hypothetical example of prices moving from 100 to 120 if it makes no difference how long it takes, what difference does it make whether it was too much money or a shortage of goods? Must we now go over the past century of so of the CPI and extract which parts were due to changes in supply and which were changes in demand?

Now I would have thought that it would be perfectly reasonable to talk about "supply side inflation" just as easily as "demand side inflation." For that matter I've long informally referred to the ongoing price reduction in computing power and similar technological gadgets as "technological deflation" because it's a convenient and descriptive term. Is there some agreement among economists to create technical terms that renders simple concepts into obfuscatory jargon?

[bakho]

Just curious. How do you reconcile the "oil shocks are not inflation" with Volcker's late 70s monetary policy to stop what looked like inflation but was really an oil shock? Were double digit interest rates in the early 80s justified? What am I missing?

[jdrietz]

RE: Killus -

If the price level goes up from 100 to 120 due to monetary expansion, the same amount of goods are still being produced. Incomes eventually adjust upwards as the "new" money works its way through the economy. Consumers thus regain their initial purchasing power though the adjustment period can take some time. Not a good thing, but temporary.

If the money supply doesn't change, but there is general decrease in production of goods, prices do go up. However, consumers can't buy as much as before because the goods aren't there to buy. Assuming production stays at this lower level (just as the money supply in the first example stayed at the higher level), consumers are worse off in the long run.

RE: Recessions' silver linings

I'll take a slightly different approach than Samuelson. Rather than over-investment, I would say "wrong" investment or mal-investment, to borrow from the Austrians.

Consider the housing boom. The spree of home buying, the result of low interest loans, caused construction companies to kick into high gear and greatly increase construction employment.

Now that the housing market is tanking (which may well lead to a recession), construction companies have to lay off employees, as will related industries. It is a natural adjustment to an understandable drop in demand.

There is no doubt that it is unfair to the workers wo were laid off (more on this later). However, if wages levels are flexible, unemployment will be temporary as the unemployed workers will find positions in other industries. This will result in lower wages; however, lower wages are offset by lower prices due to recessionary deflation and lower labor costs.

How to solve the problem? Avoid booms caused by "easy money" and the distorted investment that results.

Just in case you didn't see it coming :-) my vote is to let the market set the interest rate, rather than put the power in the hands of a few chosen economists at the Fed to decide (guess?) what is best for the economy.

[reason]

Dave Iverson,
surely you and Krugman are talking about two different things. Krugman is talking about the waste of resources (especially workers) during a recession, and how to get them productively employed and you are talking about the credit cycle. You can both be right and not contradict one other.

[Dave Iverson]

Reason,

My point is simply that Credits Cycles and Employment Cycles are umbilically linked. You can't deal with one without thinking of consequences in the other. In order to avoid the excesses of "booms" (particularly latter stages of such) you are also going to affect workers.

In order to jump start an economy post "bust" you have to be thinking in terms of both money and credit AND employment. You can't just get workers back on the job immediately -- or maybe you can, by deficit spending (assuming you have space to manoeuver, which many believe is absent in the US right now). Government attempts to get people back to work have implications for the credit market.

It all smacks of Garrett Hardin's Second Law of Ecology: You can never do just one thing! Too often economic arguements are cast up "as if" all else can be held constant. Seldom is such a reality in governemntal policy-making.