Scott Sumner  

Is deflation bad?

Reply to Scott Sumner on Globa... Where I Dissent from Nathan Sm...

Here's David Andolfatto:

Everyone knows that deflation is bad. Bad, bad, bad. Why is it bad? Well, we learned it in school. We learned it from the pundits on the news. The Great Depression. Japan. What, are you crazy? It's bad. Here, let Ed Castranova explain it to you (Wildcat Currency, pp.160-61):

Deflation means that all prices are falling and the currency is gaining in value. Why is this a disaster? ... If you hold paper money and see that it is actually gaining in value, it may occur to you that you can increase your purchasing power--make a profit--by not spending it...But if many people hold on to their money, this can dramatically reduce real economic activity and growth...

In this post, I want to report some data that may lead people to question this common narrative. Note, I am not saying that there is no element of truth in the interpretation (maybe there is, maybe there isn't). And I do not want to question the likely bad effects that come about owing to a large unexpected deflation (or inflation). What I want to question is whether a period of prolonged moderate (and presumably expected) deflation is necessarily associated with periods of depressed economic activity. Most people certainly seem to think so. But why?

So far, so good. There are lots of reasons to be skeptical of the notion that expected deflation is bad:

1. Wages and prices should adjust to expected changes in the price level.
2. Changes in the price level are not important; it's changes in NGDP that impact the business cycle.

What puzzles me is the evidence Andolfatto uses to evaluate this claim:

Let's take a look at some more recent data from the United States, the United Kingdom, and Japan. The sample period begins in 2009 (the trough of the Great Recession) and ends in late 2013. Here is what the price level dynamic looks like since 2009:

[Graph showing Japanese deflation]

Over this five year period, the price level is up about 7% in the United States and about 11% in the United Kingdom. As for Japan, well, we all know about the Japanese deflation problem. Over the same period of time, the price level in Japan fell by almost 7%.

Andolfatto then claims that Japanese real GDP per capita is up about 15% since the first quarter of 2009. I think the actual figure is about 10%, but even that is higher than the real GDP per capita growth rates for the US and Britain. Unfortunately this data doesn't tell us anything, as the claim (by people like Paul Krugman) is not that expected deflation leads to lower growth, it's that expected deflation leads to lower levels of output (due to downward nominal wage inflexibility.) It is unexpected deflation that leads to lower growth.

Then Andolfatto discusses the rise in Bitcoin prices, which means deflation of goods prices measured in Bitcoin:

One might think that given the prospect of continued long run deflation (price appreciation), that people would generally be induced to hoard and not spend their bitcoins. And yet, available data seems to suggest that this is not the case:
At this point data is presented that Bitcoins are used in about 65,000 transactions each day. I'm not quite sure what to make of that data---how many transactions would occur each day if Bitcoins were being hoarded? I also wonder about the claim that Bitcoin holders faced the "prospect of continued long run deflation." How do we know this?

In my view deflation is neither good nor bad, just irrelevant. On the other hand falling NGDP matters a great deal, and should be avoided.

HT: Michael Byrnes

Comments and Sharing

CATEGORIES: Macroeconomics

COMMENTS (8 to date)
David Andolfatto writes:


Thanks for your comments. But I'm afraid that you are the one that has left me puzzled. For example:

Unfortunately this data doesn't tell us anything, as the claim (by people like Paul Krugman) is not that expected deflation leads to lower growth, it's that expected deflation leads to lower levels of output (due to downward nominal wage inflexibility.)

First, of course I was not peddling the data as conclusive of anything; it was meant to be suggestive.

Second, over the course of 5 years, what is the difference between "growth" and "level?" Clearly, the real pc GDP in Japan is at a higher level than its two counterparts, relative to 2009.

Third, what sort of evidence would, in your view, "tell you something?"

In terms of Bitcoin owners expecting secular deflation, you are right, I do not know for sure that they think this. It is based on anecdotal evidence. But please tell me this: What do you think an owner of bitcoin is expecting in terms of price appreciation? Do you really take seriously the notion that they are expecting inflation? If so, tell us why.

Finally, while you care about NGDP, I care about RGDP. That's because I can eat RGDP.


dannyb2b writes:

"It is unexpected deflation that leads to lower growth."

Only in the short term though correct? Why is everyone so worried about short term effects anyway if money is neutral?

W. Peden writes:

[Comment removed for supplying false email address. Email the to request restoring your comment. We'd be happy to publish your comment. A valid email address is nevertheless required to post comments on EconLog and EconTalk.--Econlib Ed.]

Nick writes:

So I think there are about 13mm bitcoin in circulation. 60k a day gives us like 23mm a year (don't make fun, I no math good). So each coin is turning over twice a year, which is low.
I'm not saying this proves anything about deflation, but I'm puzzled as to why that 60k/day number would be viewed so positively.

Scott Sumner writes:

David. When you move from inflation to deflation, Japanese equilibrium RGDP (per capita) might fall from 75% of US levels to 73%, for instance. That happened (if it happened at all) in the 1990s. Once the Japanese RGDP has fallen by 2% relative to the US, the growth rate from that point forward should be equal to the US RGDP growth rate, even if deflation is continuing to depress the LEVEL of Japanese RGDP by 2%, due to labor market inefficiencies.

I don't know much about Bitcoin. I presume if Bitcoins are being hoarded then the expected appreciation in the value of Bitcoin (its purchasing power) is higher than for cash. That's your assumption. If it's being used for transactions purposes then the EMH implies the expected return should be about equal to cash. What puzzled me is that you seemed to think it was not being hoarded, but also that returns that were expected were consistent with hoarding demand, not transactions demand.

NGDP stability has instrumental value in promoting RGDP stability.

Danny, See my previous answer.

Airman Spry Shark writes:
I also wonder about the claim that Bitcoin holders faced the "prospect of continued long run deflation." How do we know this?
There is a fixed, finite number of Bitcoins that can ever be 'mined', due to the design of its algorithm; the rate of mining is also going to slow over time (again, due to algorithm design).
Scott Sumner writes:

Airman, Yes, but there's also the Efficient Market Hypothesis to consider.

Airman Spry Shark writes:

I suspect that the bulk of its transactional use occurs where cash is not an available alternative (e.g., the late Silk Road) or users are willing to accept higher transaction costs for the anonymity benefits.

Comments for this entry have been closed
Return to top