Well, the Phillips Curve is very many things. There’s the actual “curve of Phillips,” which was drawn and statistically estimated in a famous 1958 paper by the New Zealand economist A. W. Phillips. That curve plotted the relationship between the rate of wage increase and the unemployment rate in England from the early 1860s to the late 1950s, and it appeared to show that high wage increases were associated with low unemployment, and vice versa.

Although Phillips didn’t greatly expand upon what he was thinking — I suppose he thought that people would know what he had in mind — looking at the curve, one would naturally think that the explanation for the relationship would be a demand-side explanation: when unemployment is low, the labor market is tight, so wages rise quickly. That’s just a commonsense piece of economics. And insofar as wage increases are passed on to price increases, this would make the explanation of inflation predominantly one of demand.

Now, that’s interesting because it was a commonplace of 1950s Keynesianism that inflation was predominantly caused by so-called autonomous “cost-push” factors rather than demand. The idea was that union wage bargaining could drive up wages more or less irrespective of the unemployment rate — that it was a supply-side phenomenon driven by the monopoly power of the trade unions (or in the United States, sometimes the oligopoly power of firms).

So what Phillips’s Phillips Curve seemed to show was the authenticity of “demand-pull” rather than the more Keynesian “cost-push” theory of inflation.

Now, that’s the beginning of the story. But there were other economists, following Phillips, who drew the same sort of shape and associated it with entirely different theories, in all sorts of ways. So throughout the 1960s, there are other statistical estimates of the Phillips Curve that introduce additional explanatory variables, besides unemployment, some of which are more characteristic of cost-push schools.

For example, one thing they would put in was some measure of profitability, because the cost-push people always argued that unions bargained harder for wage increases when profits were high. So to the extent that profit is found to statistically explain wage increase and hence inflation, that could look like a cost-push explanation.

But the point is, the literature in the ’60s was mixed. All sorts of different Phillips Curves appear. And this is part of the problem. If you say the Phillips Curve is something or other, whatever it is you say, you can probably find something in the literature that supports that. And equally well something entirely different than that.

So what is the Phillips Curve? It’s a very problematic matter, I’m afraid.

Yes, right. The Phillips Curve Myth is a collection of stories, or variations on a story, that says that there was once a widespread, or consensus, opinion — especially typical of Keynesian economists, especially in the 1960s into the 1970s — that lower unemployment could be bought at the price of somewhat higher inflation, and that this had been demonstrated by Phillips’s Curve.

So, Keynesian economists (or “so-called Keynesian economists,” if you don’t think they were real Keynesians) are accused of two things. First, they are accused of believing that high inflation reduces unemployment, or anyway that policies that reduce unemployment also increase inflation, and that that’s a sustainable long-term outcome. Of course, that’s said to be an error.

But the myth is the idea that lots of people ever believed this, or that it was the consensus of the time. That’s the first part of it.

Secondly there’s the proposition that policy in that era was framed around this erroneous Keynesian belief, and that the resulting quest for lower unemployment through inflationary policy explains the rising inflation of the ’60s and the inflation crisis of the ’70s — certainly in Britain and probably in the United States. And maybe other countries too, but I focus on those two.

So, that’s a myth because it’s completely untrue. Practically nobody believed anything like that.

Right. So the second story, the conventional truth of today, as you put it, would be one in which the appearance of high wage increases going with low unemployment is seen as an outcome of disequilibrium. So when the labor market is unusually tight and unemployment is unusually low, wages will tend to rise fast and pull up inflation with them.

What happens next in the modern story is that the wage bargainers will adjust to the realization that there’s now persistent inflation. So, let’s say current labor-market conditions — the tightness of the market — are such that, if there were no expectation of any inflation, the result would be a wage bargain providing for, say, a 3 percent wage increase. But if you have those same labor-market conditions in circumstances where inflation has for some time been about 5 percent, then wage bargainers will expect those circumstances to continue, they will take it into account in their wage bargaining, and the result will be a wage increase of not 3 percent but three plus five, that is, 8 percent.

So if one were to set policy in the manner of the myth, to maintain low unemployment at the cost of a little more inflation, that will work temporarily, but the higher rate of inflation will soon come to be expected. And when it is expected, it will come to be incorporated in the wage bargain and inflation will increase. And if the policy maker persists in maintaining that low level of unemployment, the process will repeat itself over and over, and you would not get “a little extra inflation” — you would get ever-accelerating inflation, which would continue as long as unemployment remains too low.

So this low rate of unemployment — which people in the ’60s supposedly (according to the myth) believed could be maintained indefinitely by allowing some higher but steady rate of inflation — can, according to today’s orthodoxy, in fact be maintained only by allowing an accelerating rate of inflation, and accelerating ultimately without limit.

And pretty well nobody thinks that allowing inflation to accelerate without limit is a good idea. So this amounts to the view that macroeconomic policy should not be seeking to maintain that low level of unemployment in the first place. It won’t work, since it will result in ever-accelerating inflation.

That’s the sense of the modern view and it clearly makes the mythical version of the Phillips Curve out to be a very foolish theory.

It sounds mad, but the understanding before Friedman’s 1967 presidential address was that if you ran an inflationary policy with excessively low unemployment, expectations would adjust and the relationship between inflation and unemployment would shift. In other words, that little story Friedman tells about expectations in his paper, the story about —in Friedman’s terminology — the “shifting Phillips Curve,” that theoretical story is utterly commonplace before 1968. It’s not put in the language of the Phillips Curve, which can make it a little harder to find. But it’s very much there. In the end, I found thirty-plus perfectly clear statements of it by very prominent economists.

And in fact, Friedman’s famous paper, which people refer back to, as you say, as if he were bringing the tablets down from the mountain, as if it was a great insight — if you actually read it, this little argument doesn’t get that kind of emphasis in the paper. It’s just part of a story he’s telling. But there’s no ringing the bell to say, this is the big point I want to make. Nothing like that. The story doesn’t feature in the introduction of the paper. It doesn’t feature in the conclusion. There’s no hint of it in the title. It’s a couple of paragraphs buried away in the middle.

So the answer to the question, “What did people actually believe back when they were supposedly captured by the mythical story?” is that they actually believed exactly the same story Friedman said they should have believed all along. Nobody thought it was possible to keep the labor market persistently out of equilibrium. It’s very, very hard to find one person who says that.

Well, the emergence of the myth can be dated quite precisely. It’s 1975. That’s the beginning of it. Milton Friedman gave a lecture in London in 1975 where he more or less told this story. He said, you know, silly old Phillips believed a lot foolish things and the Keynesians were very enthusiastic about them. And then he repeated a very similar story in his Nobel Prize lecture, which was published 1977.

And it’s already in the undergraduate textbooks by 1978 — extraordinarily quickly. I have a paper where I looked at runs of undergraduate textbooks. I took textbooks that had several editions on either side of 1978, and the myth doesn’t appear before 1978.

Not all the textbooks adopt the myth. Some stick to the truth, some adopt the story. And some of them adopt the story — that all these foolish people used to believe that the Phillips Curve illustrated a stable trade-off — even when you can see that in their own earlier editions, there’s no sign of such a belief. So they make up the myth and essentially accuse themselves, in earlier editions, of having made this mistake, when they hadn’t. It’s extraordinary.

But the dating turns out to be very precise. It’s in the textbooks in 1978 and not before. So it must be Friedman’s 1977 Nobel lecture that did that.

I’m sure it’s also true that the timing was just right for what used to be called the New Right — Thatcher- Reagan — to focus on the control of inflation as the policy priority. I don’t think Reagan and Thatcher themselves did this, but their supporters found it convenient to disparage past policy makers because this was supposed to be a new start for economic policy. So surely there’s political value in the story there.

If anything, though, I think it’s more pronounced a bit later, in the movement for central-bank independence in the 1990s. There’s an extraordinary number of facile little pieces written by central-bank research department people, mostly in central-bank publications, that ran through the myth as straightforward historical fact, in a sort of tone — and sometimes explicitly saying — that this is why we need an independent central bank.

There’s a bit of mud there, of course, because if it was a mistake that everybody in the ’60s shared, including central bankers, it wouldn’t matter whether you had an independent bank or not. But I guess the idea is that it’s a mistake made by politicians because they’re the ones who wanted low unemployment — that’s sort of the implication to it.

So I certainly think there’s political value in the myth for advocating central bank independence. (Which is starting to look a bit silly at the moment.)

And then it’s a very attractive story educationally, of course. Because it has the appearance of being a story of scientific progress when one says there was just great puzzlement before 1958 — nobody knew anything much — and then Phillips discovered something with an element of truth but misunderstood it.

And then Friedman came along and said, well, that’s right, except that you haven’t considered inflation expectations: once you run an inflationary policy, people will come to anticipate it, and that will affect the relationship between unemployment and wage bargaining with the consequence that inflation accelerates.

This looks like progressive steps in scientific development, so one could see why it would be attractive to textbook authors, to enthrall their readers with the progress of economic science. And one could see, therefore, why it would stick in the textbooks and keep on being repeated.

And then it doesn’t take long — since it’s an attractive story that people remember — for people who were undergraduates reading the textbook in 1978 to be lecturing first-year undergraduates ten years later. If that’s what they learned as undergraduates, why would they question it? And so it stays in the textbook, stays in the teaching, and it’s great for economic science.

Well, you would’ve thought, wouldn’t you? That’s the way inflation targeting is described, but that’s not what [Bank of England governor] Andrew Bailey says about the current inflation. He says that we’ve got to blame the Russians and greedy businessmen, and that sort of thing. Inflation has increased and the central bankers strangely don’t accept that it’s their fault. And they might have a case, too, at least up to a point.

I think I’m probably in sympathy with what you were saying. I think, setting aside COVID or the war in Ukraine, in normal circumstances, at least, there has been an excess of caution, based on fear of rising inflation, on the basis that it will quickly get its way into expectations, and once it does that, it’ll be very hard to control it again. And I think that probably resulted in long periods of excessively tight policy in Europe. And I think that’s ultimately traceable to this idea that, back in the past, we made this mistake terribly badly and we mustn’t make it again.

So there’s that. But I also think that it had the practical effect of making the Phillips Curve far too central in macroeconomic policy, so that everything appears to be about the unemployment-inflation relationship, to the neglect of other macroeconomically important outcomes such as asset bubbles. The fixation on inflation, meaning consumer price inflation of one sort or another, is surely part of the reason that asset bubbles have been allowed to develop, including before the financial crisis.

And the logic, or pseudo-logic, is that, since we’ve got this terrible problem with the temptation to exploit the Phillips Curve, we must insist to the central banks that they not do that. And the mechanism of insisting is inflation targets.

But in telling them the inflation target is everything, that that’s what you ought to worry about — well, that’s actually an instruction not to worry about asset bubbles or currency bubbles or financial crisis. And of course, if you instruct the central bank not to worry about these other things, you’ll increase the danger that things will go wrong.