Scott Sumner  

Deficits always matter

Hugh Hewitt on the Interest De... Monetary offset isn't really a...

Paul Krugman has a new post entitled "Deficits Matter Again":

Not long ago prominent Republicans like Paul Ryan, the speaker of the House, liked to warn in apocalyptic terms about the dangers of budget deficits, declaring that a Greek-style crisis was just around the corner. But now, suddenly, those very same politicians are perfectly happy with the prospect of deficits swollen by tax cuts; the budget resolution they're considering would, according to their own estimates, add $9 trillion in debt over the next decade. Hey, no problem.
He's right about the GOP hypocrisy, but I can't help noticing that the GOP is not the only group that changed their tune after the election.

Unlike Krugman, I've consistently argued that deficits are not a good way of stimulating the economy, and that monetary policy should be used to assure the appropriate level of aggregate spending.

Those apocalyptic warnings are still foolish: America, which borrows in its own currency and therefore can't run out of cash, isn't at all like Greece. But running big deficits is no longer harmless, let alone desirable.

The way it was: Eight years ago, with the economy in free fall, I wrote that we had entered an era of "depression economics," in which the usual rules of economic policy no longer applied, in which virtue was vice and prudence was folly. In particular, deficit spending was essential to support the economy, and attempts to balance the budget would be destructive.

This diagnosis -- shared by most professional economists -- didn't come out of thin air; it was based on well-established macroeconomic principles. Furthermore, the predictions that came out of those principles held up very well. In the depressed economy that prevailed for years after the financial crisis, government borrowing didn't drive up interest rates, money creation by the Fed didn't cause inflation, and nations that tried to slash budget deficits experienced severe recessions.

That last sentence is incorrect. If you look at the group of nations with independent monetary policy, there is no correlation between (cyclically-adjusted) deficit reduction and growth. At the beginning of 2013, a letter signed by 350 Keynesians warned that fiscal austerity risked pushing the US into recession. Instead the deficit fell by nearly half, and economic growth actually sped up in calendar year 2013 (Q4 to Q4). The reason is simple---monetary offset.

But these predictions were always conditional, applying only to an economy far from full employment. That was the kind of economy President Obama inherited; but the Trump-Putin administration will, instead, come into power at a time when full employment has been more or less restored.

How do we know that we're close to full employment? The low official unemployment rate is just one indicator. What I find more compelling are two facts: Wages are finally rising reasonably fast, showing that workers have bargaining power again, and the rate at which workers are quitting their jobs, an indication of how confident they are of finding new jobs, is back to pre-crisis levels.

What changes once we're close to full employment? Basically, government borrowing once again competes with the private sector for a limited amount of money. This means that deficit spending no longer provides much if any economic boost, because it drives up interest rates and "crowds out" private investment.

This isn't really the issue. To the extent that the Keynesian model allows fiscal stimulus to work, it is based on the zero bound issue, not the economy being deeply depressed. Thus in 1982 the economy was deeply depressed, but nowhere near the zero bound. Fiscal stimulus would have been almost completely ineffective at boosting demand (even in the New Keynesian model) because Paul Volcker would have tightened monetary policy enough to keep the economy on track for his low inflation goals. (Of course, there may have been supply-side effects.)

One area where I do agree with Krugman is that economic slack is almost gone. In this recent MoneyIllusion post I also cited the acceleration in hourly wage growth. James Alexander contested this argument, pointing to more ambiguous data for weekly wage growth. But I like hourly wages best, because they are a particularly sticky variable. When the growth rate of hourly wages changes, it usually signals a disequilibrium in the labor market. (In fairness, James points to one case (2008) where they briefly gave a false signal. They probably are not the best forward looking indicator, but I do think they are one of the best indicators of the current condition of the labor market--perhaps with a slight lag.

Krugman links to an earlier post that provided one argument for stimulus, even without much economic slack:

Does this mean that the case for easy monetary and fiscal policies is over? No, but it's subtler now: it hinges mainly on the precautionary motive. Right now the economy looks OK, but things may change. Of course they could get better, but they could also get worse -- and the costs of weakness are much greater than those of unexpected strength, because we won't have a good policy response if it happens.

What I mean is that because interest rates are still near zero, a bout of economic weakness can't be met with strong monetary expansion; and discretionary fiscal stimulus is politically hard, especially given who'll be running things. This strongly suggests that you want to build up some momentum, get further away from a lee shore, pick your metaphor; that means letting the economy build strength, inflation rise modestly.

In this post I criticized this view. The Fed may want to raise its inflation target (although I prefer other reforms) but it should not overshoot an unchanged inflation target to build up ammunition against a potential zero bound problem. Doing so would make inflation even more procyclical, and make another recession more likely. Indeed excessive monetary stimulus in 2006 was one of the causes of the 2008 recession. It pushed inflation above target, and a couple year slater the Fed tightened to reduce inflation at the worst possible time.

A better solution would be to switch to 4% NGDP level targeting. This would generate countercyclical inflation.

Comments and Sharing

COMMENTS (24 to date)
Market Fiscalist writes:

Probably a pedantic point but on "Unlike Krugman, I've consistently argued that deficits are not a good way of stimulating the economy, and that monetary policy should be used to assure the appropriate level of aggregate spending.".

But doesn't monetary stimulus that is carried out via bond purchases also lead to greater deficit spending when the interest payments come in and are passed back to the govt who will spend it (or reduce taxes) ? Its just more spread out over time than the other kind.

Scott Sumner writes:

Market, I'd rather say that monetary policy has fiscal effects, rather than suggest it's a type of fiscal policy. If that's your view then we agree.

Double correction: first, Greece like America borrows in its own currency, and second, the American Treasury's privileged access to the vault of a subservient central bank isn't guarantee that it "can't run out of cash." Just ask any Latin-American treasury officer that served during the 80s.
I have a printer at home but having access to a printer doesn't mean that I "can't run out of money." The same logic applies to national fiat money: The Treasury will run out of money it's no longer accepted by anyone.

Sorry for the typo:
The Treasury will run out of money once it's no longer accepted by anyone.

AntiSchiff writes:


For what it's worth, many Latin American countries that went bankrupt chose to go bankrupt rather than solely using inflation to deal with debt crises. They chose a combination of inflation and default. As many hyperinflationary episodes demonstrate, especially the rather recent extreme case of Zimbabwe, inflation expectations typically get enormously high before currency abandonment occurs. The US will not come clise to currency abandonment.

Scott Sumner writes:

Pedro, Didn't a lot of the Latin American bankruptcies involve dollar denominated debts?

Andrew_FL writes:

Krugman said in August and tweet the same article in October that President Clinton should deficit finance years infrastructure spending

And we've been at NAIRU since the beginning of last year! And perhaps more importantly, unless Krugman thinks Hillary's election would've triggered a recession Trump's election averted (LOL) it doesn't make sense for him to think we wouldn't be at or below NAIRU for the bulk of President Clinton's hypothetical first term.

Scott, dollar denominated debt was only part of the problem, maybe at some point a trigger, but the really significant problem was chronic deficit and unsustainable domestic debt leading to lack of trust in the currency and then hyperinflation.
AntiSchiff, it's my opinion that the process is the opposite of what you describe: it's not hyperinflation that leads to lack of trust in the national currency, it's lack of trust in the national currency that leads eventually to hyperinflation. You can even have chronic lack of trust without reaching the phase of hyperinflation, but severely limiting the amount of *real* currency that can be printed by a government.
Regarding the US, when I consider my own portfolio strategies, I can tell you that my implicit trust in the dollar has been severely impacted during the last 10 years. Any fiat currency is just a step away from mismanagement if the political environment is conducive to it.

bill writes:

If memory serves, Krugman was calling for fiscal stimulus as recently as October (when he was probably certain of a Democratic President and a Republican House)

To make it even more clear, it's theoretically possible to default on foreign obligations while maintaining monetary policy and domestic debt under control. The problem is that usually when a country defaults on its foreign obligations it's done exactly because its monetary policy and domestic debt is out of control. The somewhat widespread notion that countries fail because of their foreign indebtedness are in my opinion more about ideology and political face-saving than about solid monetary and public economics.

Andrew_FL writes:

@bill-You're right, in October, Krugman tweeted an article NYT published in August, and called for years of deficit financed infrastructure spending by Hillary Clinton, as I mentioned above-we were thinking of the same article/tweet.

Scott Sumner writes:

Pedro, I understand all that. I thought you were claiming that countries could not avoid default with inflation. They can unless the debt is denominated in foreign currencies.

James Alexander writes:

Mmm. "One case" in 2008. A modestly important one I would have thought: and in economics where macroeconomics great failing is having so few reliable data points vs lots of competing theories.

Also, oddly enough, AHE looks to have given EXACTLY THE SAME wrong reading in the previous case too, in 2000.

In neither case did AWE give a false signal. In fact, falling growth in AWE before the recession hit was a really GOOD indicator that monetary policy was quite tight enough, too tight, in fact.

foosion writes:

Is there a precise and generally accepted definition of "austerity"? It can't just be shrinking deficits; for example, how many were claiming Clinton was practicing austerity?

Shane L writes:

It seems that left and right have switched places on multiple issues in 2016, which suggests tribal identification as a driving force in politics. This is illustrated starkly in a Yougov/Economist poll in December 2016 showing that Republicans and Democrats flipped on attitudes towards Vladimir Putin in the summer of 2016:

Broadly speaking, there seems to be flipping on attitudes towards:

- The EU
- Globalisation/free trade
- Russia/Putin
- Deficit spending
- Infrastructure spending
- The health of the American electoral system
- The CIA
- The poor, the masses
- Wikileaks

I'm being very loose with definitions here and one could nitpick, but I have definitely seen some individuals personally changing opinion, perfectly in step with others of their political identity, to embrace ideas that a few months ago they stated were abhorrent.

AntiSchiff writes:


Yes, on Dr. Sumner's point, inflation got incredibly high in Weimar Germany, post WW2 Hungary, and of course, more recently Zimbabwe, to mention just a few examples. People are very reluctant to abandon currencies, even in such extreme cases of inflation, though it does happen. For one thing, taxes often must be paid in the official currency.

Hence, domestic debt can normally be inflated away, without default.

roystgnr writes:

I'll second the "define austerity" question. I'll happily accept multiple definitions if different economists disagree; inconsistent definitions would still be a big step up from no definition.

You contrast "fiscal austerity" with "the deficit fell by nearly half", which suggests that the distinction between "austerity" and "extravagance" hinges on whether the second derivative of debt with respect to time is negative or positive.

Is there some way to prove that this is indeed the function of debt which might be expected to have a contractionary or expansionary effect, rather than, say, the first derivative (as I would have naively assumed), or the third derivative?

Scott Sumner writes:

foosion and roystgnr, I think it's usually defined as a significant fall in the cyclically adjusted deficit. It's a term most often used by Keynesians, so I think it makes sense to go with their definition, even if one doesn't regard it as sensible. Since I am often criticizing Keynesian claims, I think it's only fair to rely on their definition, not mine.

Shane, I see exactly the same thing.

Market Fiscalist writes:

On 'Market, I'd rather say that monetary policy has fiscal effects, rather than suggest it's a type of fiscal policy. If that's your view then we agree.'

I think we agree. I believe that the only way the authorities can vary the money supply over the long term is via deficits. It just turns out that monetary policy is the optimal way of making any required short term adjustments to the money supply because it minimizes the short-term effect on the deficit by allowing the effects to be spread out over time.

foosion writes:

"The cyclically adjusted budget balance, sometimes known as the full employment budget balance, is the budget balance that would obtain when GDP is at potential."

"A projection of the government's budget deficit assuming the economy is at a normal level of activity. This is done by presupposing that consumer spending and taxes remain unchanged."

These definitions leave some room for interpretation, as it's hard to know what GDP would be at potential or at a normal level of activity. The later definition seems less open to interpretation, but I rarely see it.

Any corrections would be appreciated.

Thaomas writes:

I think it's closer to the truth to say "deficits NEVER matter." If a government has the opportunity to invest in an activity that will bring future benefits greater than present costs, taking into account the borrowing rate and opportunity costs, not necessarily the market prices of the inputs into the activity, it should engage in that investment even if it increases the deficit.

This would mean that if the central bank does not succeed in maintaining NGDP on target and the economy falls into a recession in which borrowing rates are low and some inputs have market prices above their temporarily unemployed opportunity cost, application of this rule will result in deficits rising during recessions and falling during booms, but in neither part of the business cycle should the "deficit" be part of the decision function.

Thaomas writes:
To the extent that the Keynesian model allows fiscal stimulus to work, it is based on the zero bound issue, not the economy being deeply depressed.

What is meant by "fiscal stimulus" here?

The economy being deeply depressed (marginal costs being less than market prices for many project inputs) with low real borrowing rates would mean that the deficit would ceteris paribus go up if the government were following a prudent NPV rule for investing. Personally, I would not consider that "fiscal stimulus." Investing beyond what an NPV rule would yield would be.

If the depression occurred because the central bank is constrained in keeping NGPD on target by, say, an inhibition on buying "whatever it takes" worth of government or private domestic currency or foreign currency denominated assets and the government can supply the central bank with the kind of assets it is NOT constrained from buying the "fiscal stimulus" will raise incomes by allowing the central bank to come closer to its target if the NPV rule is not violated "too much."

foosion writes:

@Thaomas, presumably fiscal stimulus means a significant rise in the cyclically adjusted budget deficit. See the prior few posts on the definition of austerity.

Craig writes:

How are you modeling current fed policy:

1. (countercyclical): above trend inflation during recessions, and below trend during booms / tight labor markets.
2. (procyclical): below trend during recessions, and above trend during booms.
3. (half pro cyclical): below trend during recessions, and approaching "trend" (target) during booms.

Obviously 1 > 2 and 1 > 3. Do you expect the Fed to follow through on 1 next recession? If not, do you prefer 2 or 3, and is that what you think the fed is following?

Comments for this entry have been closed
Return to top