Log in

Previous Entry | Next Entry


This began as a reply to a post on bgmaster's journal that outgrew its playpen.


To have any sensible discussion about inflation and deflation, you need to decouple two concepts that a lot of people jumble up: price level and money supply. The money supply is the quantity of money in circulation, taking into account the multiplicative effects of banks. The price level is how much goods cost. The price level is partly dependent on the money supply, but it's also based on the aggregate supply and demand for goods.

Inflation and deflation are terms relating to the supply of money, but they're often measured by looking at the price level. That's a nice, easy measure, but we must remember not to confuse the force with the instrument. Inflation is not rising prices, though rising prices may be caused by inflation. Deflation is not falling prices, though falling prices may be caused by deflation.

Inflation and deflation are both bad because they represent a disruption in the unit of account and store of value. They're bad for the same reason bad accounting is bad: they distort expectations and hide economic truths. But rising prices and falling prices aren't necessarily so. They're bad if they're caused by inflation or deflation, but not if they're caused by shifts in aggregate supply and demand.


Armed with this understanding, we can now take a look at what's happening today.The price-level reduction ("deflation") we're seeing today is the result of two things. One is deleveraging and tightening of credit, which reduces the money supply and thus is genuinely deflation. This is bad in theory, but in practice, there was far too much credit during the bubble, and the money supply reduction is being matched by a fall in inflated asset prices (houses, mortgage-backed securities, bank debt and stocks).

The other, though, is a fall in aggregate demand as people restrict their spending to what they can actually afford. This is painful for companies that produced to meet the bubble's expected demand, but it won't continue unabated: the market will destroy marginal suppliers only until supply meets the new demand. That's why we see brinksmanship-playing Wall Street firms, poorly managed car companies, and teetering newspapers dying out: these were all unhealthy companies even during the boom, so in the bust they're the first to go. (The Wall Street firms had a false veneer of health, but fundamentally they had the financial equivalent of osteoporosis.)

Aggregate demand spirals do end if supply is allowed to fall to meet demand. Even when people expect prices to fall, they can only hold off their demand so long—they need to eat, they need to clothe themselves, they need to buy natural gas and heating oil and water and electricity and gasoline; these purchases won't wait. And even discretionary purchases won't wait. In eighteen months you can buy a computer just as fast as the ones on sale today at half the price. Does that keep you from buying? Of course not—a computer in eighteen months is less valuable to you than a computer today. You buy when it's cheap enough, not when it's as cheap as possible. That terminates the spiral. If that weren't true, the computer industry would be a wasteland of deferred demand, not a shining example of innovation and progress and profitability.

But if supply can't meet demand because government bails out marginal suppliers, something else happens. Propping up marginal firms just shifts them a little further away from the margin. The margin then shifts too as it seeks supply to destroy, and suddenly it finds twice as many firms: the ones you saved and the ones that were slightly healthier than those. Now you have twice as many firms to save, and some of them would never have been threatened if you hadn't tried to rescue the others in the first place. The bailouts become more expensive, and even so they don't really save the bailed firms; they just add even more firms to the margin. That is precisely the sort of "deflationary spiral" that Keynes worried about, but note that it wouldn't have happened without the bailout that propped up the original margin.


Earlier I spoke of an inflation-free economy where the money supply neither expands nor contracts. I said that in such an environment prices would slowly fall as firms found more efficient ways to produce; thus, even if you earned the same wages you could buy more with them in the future. As I hope I've made clear, this is not deflation because the falling prices aren't the result of a change in money supply. They don't represent a distortion in the unit of account; they can be predicted to some degree and they happen slowly and naturally, not arbitrarily and quickly in the way changes in the money supply happen. It's debatable whether this is a change in real wages, because it's not clear if "real" is a term that factors out changes in money supply or changes in price level. It certainly does mean that the same wage buys more, though.

But it buys more for a perfectly healthy reason: things are cheaper to make. If things don't get cheaper, then prices don't fall, and we end up with no changes in the price level. There's no external force driving this, just productivity gains. It isn't changing because we suddenly took half the money out of circulation, soaked it in gasoline, and set it on fire, which would be deflation.

Can an inflation-free money supply happen? Absolutely. All you have to do is create a currency whose supply is constrained by some outside force, keep banks disciplined with high reserve requirements, and most importantly, resist the temptation to debase the currency in hard times. Not easy, but within the reach of mere mortals.


( Read 14 comments — Leave a comment )
Mar. 21st, 2009 01:26 pm (UTC)
*flies over Brent in a helicopter and drops a few bags of $100 bills*
Mar. 21st, 2009 02:14 pm (UTC)
*laughs* Oh, it could be worse - the original post (the one through commentary on BG's post) mentioned Atlas Shrugged, and so I immediately looked forward to a catfight economic commentary.
Mar. 21st, 2009 01:52 pm (UTC)
Ok, now for the serious reply. This is one of those situations where I agree with your conclusions, but not some of your arguments. And, being a former (and future) nuke, I am required not to settle for "Right Answer, Wrong Reason (RAWR)."

AS I UNDERSTAND monetary policy (something that I'm not an absolute expert on, though I AM rather interested in it and am always seeking to learn more), it is correct to regard inflation/deflation as rising/falling price levels, because they are symptoms. The cause is quite unrelated to that fact.

As for the cause, Keynsians belive that it's caused by interest rates and govenment deficits. The Say/Friedman/Sowell line of economics believes, instead, that inflation is caused by, as Reagan put it, "Too many dollars chasing too few goods." I think you know me well enough, by now, to know which side I take. This follows from Alfred Say's principle teaching - that is that wealth is not money; it is goods and services. The money is just the score card.

Now, as for your two listed causes for the current deflation, I have two comments.

First, the tightening of credit is, as you indicate, actually the removal of a form of "soft" money from the economy. It's clear that you understand that and base your conclusions on it. I just feel that it's important to state that explicitly. Most people only understand money as it exists at the M1 or M2 level.

Second, deflation as a result of lower spending has a definite floor. Business that can not sell goods for enough to recover cost will not stay in business past the point where they realize that spending will not bounce back in the immediate future. In other words, we're limited to the price of production plus about 5% (which is where many investors decide their money is better used elsewhere).
Mar. 21st, 2009 10:05 pm (UTC)
I'm generally arguing from the Austrian school here (Mises and Hayek, if you're familiar with them).

There is an important difference between changes in the money supply and changes in aggregate (goods) demand: changes in aggregate demand are sending true price signals about the preferences of consumers, while changes in money supply are sending false signals concocted by the central bank. A contraction in aggregate demand is a very real phenomenon that tells suppliers to reduce production, while a contraction in money supply, though it has the same effect (there are fewer units of currency seeking goods), does not represent a change in demand that suppliers should react to. Similarly, a rise in aggregate demand calls for suppliers to increase production, while a rise in money supply sends the same signal (there are more units of currency), but doesn't represent a change in consumer preferences.

When suppliers react to true signals (aggregate demand changes), they're simply adjusting production to match the needs of consumers. But when they act in response to false signals (money supply changes), they're misallocating resources, and that misallocation ultimately has to be corrected. For inflation, when the expansion in the money supply sends a false aggregate demand increase signal, that corrective phase is called a recession, and involves factors of production becoming unemployed so they can be directed into new employment in line with true consumer demand.

But if the central bank responds to the crisis by expanding the money supply again, the correction process ends before it's truly finished. Some of the factors are once again misallocated, usually into some area of the economy that's experiencing some genuine growth (but not nearly enough to warrant the factors added to it), and the process begins anew—except this time it's worse, because there's a "misallocation debt", so to speak, that the new misallocations are being added to. The cycle only ends when the central bank finally stops inflating to end crises—as in the postwar decontrolling of the economy or Volcker's 1982 shock therapy.

We find this pattern again and again in U.S. history. Between 1921 (the end of the last recession that was allowed to run its course naturally) and 1929, there were actually two or three near-recessions that the Fed averted by inflating; the ferocity of 1929 was partly a result of the built-up imbalances. The 50s and 60s saw similar recessions fomented and suppressed by inflation, culminating in the 70s and early 80s. And there were at least three partially-suppressed recessions between the mid-1980s and 2007. The severity of the current recession is the result of what we didn't suffer then.
Mar. 21st, 2009 04:51 pm (UTC)
> All you have to do is create a currency whose supply is constrained by some outside force

This is, of course, the theoretical underpinning of tying money to things like the gold standard.

> prices would slowly fall as firms found more efficient ways to produce

I don't want to detract from your (entirely correct) point in any way, but it may be worth mentioning that this is frequently not what actually happens. Instead, what we often see (particularly with electronics) is that the same dollar buys a better product, rather than simply more of the original. The relevant measure is not so much price as it is consumer surplus -- the amount of value retained by the consumer in a transaction.

So, to rephrase what you said a bit more precisely: The amount of consumer surplus would slowly rise as firms found more efficient ways to produce.
Mar. 21st, 2009 06:32 pm (UTC)
Ah, but often the ability to purchase a better product at the same price is coupled with the alternative to buy the same product at a cheaper price. The old model may be repackaged and rebranded, but the option still exists, unless it becomes too obsolete to bother selling or buying.

But obsolescence is an entirely different argument.
Mar. 21st, 2009 04:53 pm (UTC)
Yes, it worked so well with the gold standard
Mar. 22nd, 2009 11:15 am (UTC)
Maybe you're right. After all, it's not like the Industrial Revolution was a very prosperous time, or the building out of the railroads improved peoples' lives.
Mar. 22nd, 2009 12:17 pm (UTC)
And so was the 1500-1700s in Spain, which had a very stable economy based on gold.
Mar. 22nd, 2009 12:28 pm (UTC)
Noted. I never mentioned gold because it is not ideal for money by definition: it's only good if it's hard to mine more. Spain during that period had massive gold inflows that they took great care to keep within the country. In that situation, gold is a terrible form of money, nearly as bad as paper printed by incompetents (I don't count the Fed in that category).
Mar. 21st, 2009 07:56 pm (UTC)
Constraining by some outside force doesn't really work so well, though. See Megan McArdle's posts here and here on the problems with the gold standard.
Mar. 21st, 2009 09:49 pm (UTC)
On another note, though, you're right about the whole aggregate demand thing, which isn't going to return to previous levels because people are (thankfully) starting to save again. But there's a whole lot of economic pain to go through to reach equilibrium again (and, indeed, what generally happens is that we overshoot in the other direction and then bounce back), and there's also a risk that we don't correct, or at least don't correct for a very long time. This is part of the rationale behind stimulus - ease the pain during the downturn, and prevent it from turning into what Roubini describes as an L-shaped recession instead of the normal U-shaped recession.

One of Keynes' insights was that, well, deflationary spirals don't work themselves out very well - again, the entire period between the mid-1870s and 1890s was one large period of constant deflation and economic pain (and people were on the gold standard then), and of course the Great Depression was a horrific deflationary spiral. I think basically our main disagreement is that I don't think your idea would work, and I think it would leave the economy - as it was before fiat currency - much more prone to very long and very nasty recessions and deflations. Low levels of inflation are much less damaging than low levels of deflation and are easier to control, so it's a less damaging bar to set.
Mar. 22nd, 2009 12:51 am (UTC)
The argument against the stimulus is that it doesn't ease the pain, it just spreads it out over time. The analogy I've heard is that it's like peeling off a band-aid—you can do it quickly or you can do it slowly, but you can't avoid the fact that it's going to hurt.

I need to look up the numbers, but it's only by measures of price levels that the 1873-1890s "Long Depression" looks like horrible pain. First of all, it's worth noting that there was genuine deflation—of the M1 money supply, no less—starting in 1873, because the US had just de-monetized silver, leaving only the gold supply to serve as money. The Treasury was actively withdrawing U.S. Notes (specie-backed paper currency) from the economy to reflect the reduced monetary base. But the only measure that shows the depression lasting past 1879 is price levels, which (as I mentioned) were falling slowly since the U.S. had no inflation at the time but was seeing major productivity gains. Other measures of the economy, like employment, productivity, and GDP growth, were just fine during that period.

If you look back through history, you will never find anything like the Great Depression. That's because never before had governments tried to prop up wages and prices when they were trying to fall. Long, nasty recessions just don't happen in environments with a stable money supply, and they're never quite so long and nasty without a lot of other interference as well.
Mar. 22nd, 2009 03:44 am (UTC)
> people are (thankfully) starting to save again.

Unless they're the government, or anyone who happens to listen to Newsweek. I've never been able to discern the logic by which, if spending caused this economic mess, doubling down on spending is the only way to cure it.

> I think basically our main disagreement is that I don't think your idea
> would work, and I think it would leave the economy - as it was before fiat
> currency - much more prone to very long and very nasty recessions and
> deflations.

As far as the backing of money goes, I am genuinely unconvinced that there is any particular difference between whether the money is backed by ties to an external commodity or to the ability of the money to pay taxes. Dollars are nothing more than a tool for communicating value through an economy. The particulars of what a dollar is valued against is ultimately irrelevant as long as you, I, and everyone else agree about what a dollar is worth.

That said, pinning money to an external commodity does have one dramatically positive effect: It makes the money supply relatively difficult to tamper with. What we have seen on Wall Street dating back into the end of Bush's presidency is a strikingly high degree of volatility which I hear the news so helpfully describe as a "lack of investor confidence." Strikingly, many of these swings have been tied not to the revelation of new corporate news, but to political pronouncements about the number and allocation of dollars being added to the economy. Uncertainty is bound to happen when fickle politicians are free to do with money what they will (see AIG, the Dodd amendment, and the bus the company's execs have now been thrown under). Nobody in their right mind would want to invest in an economy which can be turned upside-down in a day, which is precisely what fiat money allows.

Since it is a lack of investment (and the accompanying lack of value creation) which ultimately drives a recession, I would much sooner place my hope in a system that is resistant to the kinds of short-term and almost arbitrary disruptions which lower investor confidence.

> ease the pain during the downturn

As far as this goes, the essential characteristic of a bubble is that some particular investment is overvalued. The overvaluation is unsustainable and, eventually, will lead to a downward correction by the market (i.e. the bubble will burst). As much as anything, the stimulus is an attempt to apply a ratchet to the market, prevent he downward correction, and basically preserve the bubble. Catch is, bubbles can't be unpopped. Everyone now knows that their $700,000 house is really only a $200,000 house, and even though we can make the house become worth $700,000 by inflating the money supply to reduce the value of each dollar, the value of the house never changes.
( Read 14 comments — Leave a comment )