Scott Sumner  

What kind of inflation hurts the public?

Borjas, Ideology, and #Foreign... Friday Night Video: Cuba, Sanc...

I recently participated on a panel in London, discussing options for monetary and fiscal policy. At one point I mentioned that if the Japanese didn't want such a large central bank balance sheet, then they ought to set a higher inflation target. (Of course I'd prefer an NGDP target, but it was easier to explain my point using their currently preferred target variable.)

Someone in the audience objected that higher inflation would hurt elderly Japanese on fixed income, and this might reduce aggregate demand. In contrast, I think it would boost consumer spending in Japan. Furthermore, I believe the audience member was reasoning from a price change, whereas I was not. Why?

The audience member made the mistake of thinking about the impact of inflation, without thinking about the cause. If the cause had been disruption to Japanese production from the 2011 tsunami, then his reasoning process would have been correct. (Although in that case the impact would have been too trivial to mention. Even the most massive natural disasters have only a tiny impact on macro variables in large diversified economies.) However, in my hypothetical the inflation was caused by monetary stimulus, not an adverse supply shock.

An increase in inflation caused by an expansionary monetary policy will first shift the AD curve to the right. In the short run output may increase, due to sticky nominal wages. Even if output does not rise, it will certainly not decline as a result of higher AD. It is inflation from adverse supply shocks that causes people to buy less stuff.

Japanese unemployment has recently fallen to 3%, so I would certainly not recommend monetary stimulus in Japan for the purpose of boosting employment. But I might recommend it for three other reasons:

1. Ease the burden of the national debt
2. Reduce the zero bound problem for monetary policy
3. Reduce the role of the BOJ in the Japanese economy.

In normal circumstances, it is not wise to inflate away the debt. It is unfair to lenders, and reduces policy credibility. In the case of Japan, however, previous deflation unfairly favored borrowers, and steeply increased the ratio of public debt to GDP. A modest amount of inflation is justified in that case---but nothing more than 3% in my view. In addition, when the interest rate is zero, the equilibrium rate is often below zero, which means that lenders to the Japanese government are earning an above equilibrium rate of return. A bit more inflation can prevent that.

2. If the Japanese shifted to NGDPLT and/or abandoned the short-term interest rate as a monetary instrument, then the zero bound problem would not be an issue. But as long as they use interest rates to target inflation, monetary policy becomes less effective at very low inflation rates. Thus if they refuse to adopt a sensible monetary policy, then a slightly higher inflation target might make sense as a way of keeping the policy rate above zero, a sort of "second best" policy.

3. The size of the BOJ balance sheet is inversely related to the trend NGDP growth rate. In recent years the balance sheet has become huge, with the BOJ even buying equities. I prefer less government involvement in the economy, and hence a smaller balance sheet. That means a slightly higher inflation target.

As I keep emphasizing to conservatives, it's socialism or inflation.

Comments and Sharing

COMMENTS (14 to date)
John Hall writes:

He was asking about the distributional effects of monetary policy, in particular about the effect of it on elderly people. You only address the distributional point when focusing on borrowers vs. lenders. You conclude that lenders (the elderly) are earning an above average rate of return and higher inflation will cause them to earn less.

Isn't that the whole point arguing that the elderly will be worse off by a higher inflation target? You're just like, but they deserve.

I would actually dispute some of the point about lenders being worse off. I work for an insurance company with bigly exposure to the Japanese market (which makes us a lender in Japan). Our stock price rose yugely when interest rates increased following Trump's win. Even though longer duration securities sold off as interest rates rose, companies like us will be able to invest new funds at higher rates (both at low maturities, but even more so at longer maturities, due to a larger term premium).

Thaomas writes:

There are two issues here. 1) What inflation (price level trend) target optimally trades off reducing real wage-price stickiness to achieve better resource allocation in the short run v increasing uncertainty about relative prices in the long run? 2) What inflation target is best given certain restrictions on monetary policy instruments such as reluctance to use instruments other than changes in ST interest rates or a inflation rate ceiling instead of a price level trend target?

I think the same issues in different forms could be posed about NGDP targeting.

David R. Henderson writes:

Scott, How would an expansionary monetary policy, which presumably means having Japan’s central bank buy more assets, reduce, rather than increase, the central bank’s balance sheet?

Thaomas writes:

On the "distributional" effects of different inflation targets, this just means that different prices are deferentially sticky.

Borrowers/lenders does not get at it either as debt contracts can be indexed.

LK Beland writes:

Professor Henderson,
"How would an expansionary monetary policy, which presumably means having Japan’s central bank buy more assets, reduce, rather than increase, the central bank’s balance sheet?"

I believe that Prof Sumner discusses the issue at length in this post:

baconbacon writes:
1. Ease the burden of the national debt

How does this work? According to the Wicksellian interest rate theory the rise in inflation would increase interest rates. Since Japan owes >100% of its GDP the interest rate increase would eventually outstrip the decline in debt/NGDP.

Scott Sumner writes:

John, You said:

"He was asking about the distributional effects of monetary policy, in particular about the effect of it on elderly people."

No he was not. I ought to know as I was there.

He suggested that a higher inflation target would fail to boost AD because if the cost of living increased then people could not buy as much stuff.

David, I don't agree that an expansionary monetary policy implies buying more assets, at least not in the long run. The size of a central bank balance sheet is negatively correlated to the average inflation rate, and the correlation is very strong.

Now it's possible that they might have to buy more assets for a brief period, to establish credibility, but over time they'd hold fewer assets. Compare the balance sheets of the BOJ (over 80% of GDP) and the Reserve Bank of Australia (about 4% of GDP).

Thanks LK.

Bacon, At positive interest rates you are right. But at the zero bound the nominal rates rises by less than the increase in inflation, hence real rates fall.

Chris H writes:

@David Henderson, Scott gave a much less long winded explanation but I had already written this before I saw it and I wanted to give a shot at answering your question from my understanding of MM views. The first point is that for most responsible, low inflation countries, the primary way that the inflation or NGDP growth rate is changed is through expectations rather than direct money creation. Of course if a central bank just opened the flood gates and started printing $1 trillion notes like mad then money creation would start dominating expectations (or perhaps better stated is that expectations would almost entirely be following the trend of money creation). This then means that at low inflation levels, credible policy targets will impact NGDP or inflation more than the amounts of money creation that central banks will actually be willing to engage in within these nations.

To be sure, actual steps such as the central bank buying more assets will have an impact; it will just be less than a change in expectations based on a change in ultimate policy goals. So once you’ve set a certain policy goal, you’ve mostly (though not entirely since credibility matters, it’s just most central banks in low inflation countries have a fairly high amount of credibility to at least come close to targets) shot your expectations bolt. So something like QE will be expansionary within a policy target. However a higher policy target resets expectations. As such fewer steps are needed within that given regime to get expectations up to a higher level. To give sort of an illustrative example, suppose a central bank currently has a target of 2% inflation, but the market expectation is that the bank may undershoot by up to 1%. So long as they keep that target, the bank will need to use things like cutting interest rates or engaging in QE to either raise market expectations or to just expand supply to the point of hitting their target. Otherwise businesses won’t raise labor compensation/prices and people will have lower incentive to circulate money rather than keeping it in some sort of way that’s equivalent to stuffing it under the mattress. Now suppose that the bank raises their target to 3% inflation and the market still believes that they may undershoot by up to 1%. Businesses and people will now be anticipating 2% inflation and change hoarding, pricing, and employee compensation behavior in anticipation of the change. Thus even without specific actions beyond setting the target, the central bank has raised the inflation rate. Furthermore they can now maintain it without the same level of asset purchases, tending to lower their balance sheet.

Now what if the market just doesn’t buy the proclamation of a changed target? In that case yes some expansionary steps will be needed in the short term to make the policy credible. But once that credibility is established further actions will not be needed. Think of it like establishing credibility for deterrence. Say you’re in a society that tends to be violent towards those who are perceived as weaker and strength is established by how well you can beat the guys higher up in the hierarchy. Say you have the capability to beat up anyone in this society, but are interested in reducing the amount of fights you will actually have to engage in. If you never fight anyone, everyone will go after you on the assumption of your weakness, and you’ll have to fight a lot despite never trying or wanting to. But say instead you have a very targeted fight with the person at the very top of this society’s hierarchy. Yes that’s going to be a bigger battle than any individual fight you’d have with lower end people, but once it’s done you’re at the top and everyone now respects your combat abilities. Thus almost no one will try to fight you and in the long run you live a much more peaceful existence. In that same way a bank that establishes credibility to inflate needs to take far fewer actions to do so going forward (see the Federal Reserve being able to unwind QE vs the ECB having to expand it due to more timid previous monetary policy).

I’m sure I’m off in some details and I welcome any corrections or nuance I missed, but that is the basics for changing regimes in monetary policy as I understand it.

Majromax writes:


I think you might be putting the cart before the horse with the impact on seniors' fixed incomes. Why do seniors have fixed income?

Are their incomes predominantly in the form of privately-held nominal bonds, which would be negatively impacted from higher inflation, or are they in the form of inflation-indexed pensions and government supports? (This is an honest question; I don't know what Japanese retirement supports look like.)

If 'fixed income' does mean 'nominally fixed income', then inflation would indeed hurt the buying power of seniors.

In fact, a loss like that is guaranteed somewhere in the Japanese system, in the event of increased inflation expectations. Japan has issued long bonds, and they will go down in NPV commensurate with the increase in long-term rates from higher inflation expectations; increasing inflation levels will lead to a lump sum drop in the value of the Japanese national debt stock.

Scott Sumner writes:

Majromax, I think you misunderstood my point. Yes, it might reduce purchases by old people, but it would boost overall AD. That was my point.

Jose writes:

"As I keep emphasizing to conservatives, it's socialism or inflation."

Prof. Sumner, what scenario is more "socialist"?

1. rules based central bank, that buys a diversified basket of assets (equities, mortgages, bonds of all kinds) following a NGDP futures based rule

2. Discretionary, bureau run monetary policy, that only buys treasury bonds.

Your defense of NGDP futures based monetary policy is brilliant. A second-best approach, higher inflation target under discretionary policy making seems at odds with a lot of what you write.

And your three point defense is incredibly weak: 1. let's help incompetent government,
2. let's make technocrats' lives easier,
3. smaller balance sheet = smaller role for the monetary authority (??).

Not a single word about stable environment that allows faster productivity growth therefore higher real incomes ...

Scott Sumner writes:

Jose, I agree with your criticism of my second best policy. But that does not address my claim that it is better than current policy.

Of course a rules based approach of NGDP targeting is best, as you say.

Here's another way of making my point. Volcker's 1982-87 policy (4% inflation) is better than Bernanke's 2006-14 policy (1.5% inflation).

Matthias Görgens writes:

I was at the panel in London.

Scott, you mentioned that the welfare cost of undershooting an inflation target are actually better avoided by hitting an ngdp target than an minflation target but mentioned that menu costs are an exception.

Selgin's piece "Less than zero" about constant ngdp (and thus productivity improvements lead to secular deflation) argues that menu costs are actually better mitigated that way instead of with a constant price level target.

Another topic, I remember a few years ago that people were really concerned about tovernments getting too much into debt. Now we have some people complaining central banks would run out of debt to buy.

The flip side of QE without clear price level or ngdp level target is that you can do a lot of it without raising inflation. I wonder whether countries could exploit that as a one time fix to monetise their public debt, and then adopt a slow growing ngdp level target? Or are there any other tradeoffs between which ngdp level growth rate to choose apart from size of the monetary base? Is nonetary policy supposed to be neutral here?

Ankita Goyal writes:

In order to ease the burden on the national debt of Japan, the fed should buy government bonds. This would help the monetary base to expand, while helping interest rates to fall. Inflation would not be the best solution for the national debt because it can affect the consumer price index adversely. If the elderly and unemployed are not able to afford the inflated prices of goods, then it would decrease private spending amongst the public. In order to prevent that, it is better for the fed to buy government bonds.

Comments for this entry have been closed
Return to top