Is There Really A Trade-Off Between Inflation And Unemployment?

Courtesy of travelling with my family, I mercifully missed most of the latest eruption of the "Phillips Curve Debate" -- is there a trade-off between unemployment and inflation? I once again have a bit more time for writing, and I wanted to look to see whether there was anything interesting. The NYT article by N. Gregory Mankiw "Yes, There Is a Trade-Off Between Inflation and Unemployment" held the promise of of giving me guidance on what the latest iteration of the debate was about.


Unfortunately, the writing style for op-ed's precluded putting too much wonky details, and so it was a fairly generic overview. If the reader is unfamiliar with the topic, it might be more interesting, but it did not give me what I was looking for.

Rather than attempt to explain what the mainly neoclassical economists are going on about, I want to step back and try to translate their debate into terms that would be understood by people who do not share the same assumptions. I am pretty sure that post-Keynesian economists have a lot to say about the topic as well, but once again, they tend to be discussing wonkish points that would elude an outsider.

"Mankiw Formulation"

I am not a particular fan of Gregory Mankiw's approach to economics, and I will not attempt to summarise his points. However, he does have one useful description that I will start off with. Mankiw:

Today, most economists believe there is a trade-off between inflation and unemployment in the sense that actions taken by a central bank push these variables in opposite directions. As a corollary, they also believe there must be a minimum level of unemployment that the economy can sustain without inflation rising too high. But for various reasons, that level fluctuates and is difficult to determine.

One may immediately note that the statement "that actions by a central bank push these variables in opposite directions" says nothing about fiscal policy (or other forms of policy). At most, it tells us that there are limitations to monetary policy. Someone who has not been indoctrinated by neoclassical theory would just say: well, just use fiscal policy then.

However, one may not that the alleged "corollary" -- "here must be a minimum level of unemployment" -- presumably assumes that monetary policy is the only way to influence the economy. There Is No Alternative (TINA), a topic I will return to later.

You Cannot Trade Without a Exchange Ratio

I have an engineering background, and engineering is largely the science of trade-offs. I have no strong objections to qualitative discussions, but I would argue that we need to at least know the sign of the exchange ratio between two variables in order to say that there is a trade-off between them.

Very simply, if we can have a policy that lowers both the unemployment rate and the inflation rate (or at least leaves inflation unchanged), we cannot pretend there is a meaningful "trade-off" between them.

And this is hardly theoretical: in the United States, we saw a near monotonic decrease in the unemployment rate after the Financial Crisis, yet the inflation rate has done absolutely nothing interesting.

But, But, Counterfactuals!

My previous statement would probably cause umbrage in some quarters. "That's now how it works! The unemployment rate would have fallen anyway!" (or whatever).

We rapidly end up in the fantasy world of counterfactual scenarios: we could have had a different policy that raised inflation by more, and the unemployment rate would have been even lower. Where does this "counterfactual scenario" come from: from assuming that there is a trade-off between unemployment and inflation, and generating the alternative scenario. So there is a trade-off: if we assume there is one in the first place!

Although this sounds crazy, we need to step back and think about conventional economists generate their alternative scenarios. I will return to that after I discuss a scenario of my own.

"Have a Cake and Eat It Scenario"

 Let us imagine a scenario as follows.

• In a hypothetical economy like the United States a few years ago, where the unemployment rate is near where most people think "NAIRU" is, which we will assume is 5%.
• The government launches a programme to hire about 0.4% of the labour force into minimum wage positions. (We assume that there is considerable underemployment in the low wage sectors, as in the United States of a few years ago.)
• The government cuts spending that employs higher-paid employees, with 0.2% of the labour force losing jobs.
• The average wages of jobs lost is 3 times than the hired workers, so government spending falls.
• "All else equal."

I would argue that the "conventional analysis" would just look at the dollar amounts, see that spending is lower, and argue that the net effect is a fiscal contraction. Inflation should be expected to be slightly lower. Meanwhile, the unemployment rate should be expected to fall (in the near run), unless we have a very large multiplier on employment effects of the spending cut. (Private sector employment would need to fall by 0.2% to offset the net gains of government employment.)

(Someone could very reasonably argue: "What about 'real constraints,' Mr. MMT Guy?" My response would be to point out that MMT looks at segmented labour market models, and so the the tightening of the low-wage segment is offset by the high wage, and the net income flows do matter. As a result, I do not see a reason to disagree with the final results of the conventional analysis I set out above. Things would be more difficult to analyse if the programme size was much larger than I set out.)

So we have a policy in which the unemployment rate is expected to fall, as well as inflation (very slightly). The real world might throw us a curve ball, but that's life.

One Thing at a Time

My scenario appears to break the concept of a trade-off, but the secret is that I am changing two things. My argument is that the traditional framing is to do scenario analysis with only only a single variable changing: the government can only scale up or scale down one way in which it intervenes in the economy. Examples include:

• Changing the interest rate (as per Mankiw).
• Increasing/decreasing spending on a single programme.
• Increasing/decreasing a tax rate.

This is how formal modelling is typically done: some form of analysis is used to generate something like an impulse response for a single policy variable. Analysis is done via looking at that single impulse response. The verbal formulations used by conventional economists largely cloud the issues.

If we confine ourselves to these "building block" intervention types, it is probably safe to say that most economists would agree that they are either stimulative or contractionary, and so we expect the unemployment rate and the inflation rate to move in opposite directions (since the variables have opposite correlations with the cycle).

My example breaks the framework by doing two partly offsetting actions: cutting spending in one place, and expanding it elsewhere.

TINA

One reason that discussions between post-Keynesians and neoclassicals often go nowhere is the lack of shared assumptions. If we look at Modern Monetary Theory, pretty much every policy is framed as changing multiple policy variables simultaneously. This allows alleged "trade-offs" to be broken, which appears to make no sense to conventional economists.

As a final note, it partly explains the inherently conservative nature of "neoliberal" economics. If we can only change one policy variable at a time, it looks like we are in the best of all possible worlds: any single policy intervention we can do (except tax cuts) have trade-offs that take us away from a "social optimum," and so the best policy is to make no changes. This bakes in the TINA mindset.

(c) Brian Romanchuk 2019