Just how low can unemployment go without leading to inflation?

Economic View

Hundreds of workers punch buttons out of sheet metal in a Birmingham factory in the early 1900s. Unemployment averaged below 5 per cent in 22 of the years between 1881 and 1914 (Getty Images)

How low can we go? The nice round figure of 5 per cent is generally assumed to be the lowest that unemployment can go in the UK without generating excessive wage-push inflation – a crude definition of “full employment”. We are now within a whisker of that level, with the Labour Force Survey level of 5.3 per cent for the three-month average, and in the final month, September, unemployment being 5.2 per cent.

But suppose the floor for unemployment is lower – 4 per cent, or even 3 per cent – what then? It would mean that the economy could carry on growing for quite a while yet without excessive strains in the labour market, or at least without those strains showing through in higher wage demands.

So what is the floor? The Beveridge Report defined full employment as 3 per cent. In the 1950s and 60s unemployment was rather lower than this. On the claimant count measure it ranged between 1.2 per cent and 2.6 per cent, and in 1970 it was only 2.7 per cent. Then it shot upwards, as you can see from the top graph, and was in double digits for the first half of the 1980s. Maybe the immediate postwar period was an aberration and those low levels, coupled with overly expansive monetary policies, did lead to the double-digit inflation of the late 1970s and early 80s. But if you go back further, to the late 19th century, unemployment ranged from 9.8 per cent in 1886 to 2 per cent in 1899. Indeed, in the 33 years from 1881 to 1914, unemployment averaged below 5 per cent in 22 of those years. (These figures come from a great article in Labour Market Trends, January 1996.)

Of course labour market conditions of today are vastly different from those of the late Victorian and Edwardian eras. There was, until 1911, only limited support for the unemployed, and there was a lot of casual employment, particularly in the docks. In addition those figures are claimant count ones, which are somewhat lower than those of the Labour Force Survey, as you can also see in the top graph. But without wishing to push the parallels too far, there are some similarities between the casual labour force then and our prevalence of self-employment, part-time working, and zero hours contracts. There is also the ready supply of labour coming into the job market from the rest of Europe, just as there was throughout the 19th century a ready supply from the land. Put this together and the obvious question is this: if it were possible to sustain unemployment below 5 per cent then without generating inflation, why not now?

There is a further parallel, not with the 19th century but with the 1970s. This is the high level of labour participation, which is shown in the bottom graph. We now have an even higher proportion of people of working age in work than then, the highest ever, at least in peacetime. It would figure that if the job market were sucking more people into jobs than ever before, that the sustainable unemployment rate might be lower than it had previously been.

The awkward acronym for describing the sustainable unemployment rate is the Nairu (the non-accelerating inflation rate of unemployment). This clearly has varied over time. The OECD did some calculations for the UK back in 1997, which suggested that in the late 1960s the Nairu was around 2 per cent, around 6 per cent in the late 1970s, and somewhere between 5 and 10 per cent, a pretty wide band, in the 1980s. The Office for Budget Responsibility thinks that the Nairu is 5.25 per cent, so we are already there.

The problem with all such calculations is that you don’t know how low unemployment can go without generating inflation until it reaches the level where it starts to cause problems. Then you have to act more violently than you would have, had you acted earlier. At the moment there does not seem to be much overall pressure on pay rates, and there are some sectors where pay may be falling, such as investment banking. But it is rising fast in construction, and the Bank of England agents, who talk to businesses around the country, do report that a lot of companies are finding it hard to recruit staff.

My own guess is that the Nairu for the UK now may well be below 5 per cent, but it is very hard to know by how much. There will almost certainly be a lag, in that it will take a while for tightness in the job market to show through in higher wages, but we don’t know how long that lag is either. We also don’t know to what extent pressure in the labour market can be absorbed by pushing up productivity, rather than hiring more staff. But what is most likely is that we will have to test this out in practical terms. That is, unemployment will indeed dip below 5 per cent and we will have to see then what happens to pay.

Meanwhile in the short term there is the practical decision facing the Monetary Policy Committee: when do you start increasing rates? There has been so much guidance that has proved misguided that there seems no point in commenting here, except to observe that what the Federal Reserve does in the US will probably determine what we do. We will be fast followers, not pioneers.

In the longer term, I think it is worth pondering whether there has been a sea-change in attitudes towards work that have not yet been fully appreciated by economic commentators. The big surprise after the downturn was the way unemployment rose by much less than in previous recessions, with the available work being, so to speak, “shared out” between the available work-force. That suggests there has been a shift in preferences, with people accepting lower pay in exchange for higher employment. That was a reaction to an emergency, but maybe this shift in preferences will persist, as so far it seems to have done. If that is right, the recovery can also persist for quite a while yet.