The Great Demographic Reversal: Ageing Societies, Waning Inequality, and an Inflation Revival

by Charles Goodhart

While there is continuous discussion about the potentially changing levels of the equilibrating (neutral) level of the real interest rate (r*) and of the underlying, sustainable, rate of growth (g*), the common assumption is that the NRU (u*) is fixed and constant. But the evidence is contrary. u* appears to have been more variable over time than either r* or g*. Apart from the Asian countries (China and Japan), the sustainable rate of growth in advanced economies has seemed to vary in recent decades between about 3.5 and 1.5% per annum, and the equilibrating real interest rate between 2 and 0% per annum. In contrast, u* has varied between about 5 and 2%. After the unhappy experience of the 1930s, economists such as Beveridge and Keynes initially thought that achieving a level of 4–5% would be a good outcome. But the success of full employment policies bringing UE down to around 1.5% after World War II, initially with relatively little inflation, led to the aim of holding UE below 2%.4 Demand management, and full employment policies, greatly strengthened the bargaining power of workers and trade unions. The alternative to not agreeing to the employer’s wage offer would be another job elsewhere, rather than unemployment. Also, as noted in Chapters 3 and 5, demography led to an improving dependency ratio. Union membership generally rose until about 1980

Another way of describing the NRU is that it marks that level of UE that makes workers content with the real wage growth that their enhanced productivity makes possible. The growing bargaining power (relative to employers) of labour between 1945 and 1980 meant that the underlying NRU was increasing commensurately, perhaps to as much as 5.5%. It is deeply ironic that Keynesian demand management led inexorably to a much higher NRU.

There has been recently, in the period 2007–2020, yet another twist in this tale. Whereas the focus in the (unhappy) 1970s was on a (vertical) Phillips curve, the opposite has occurred in the (equally unhappy) years of the GFC and its aftermath. Over the last decade, or so, the Phillips curve, rather than being vertical, has appeared to become almost horizontal, see Diagram 8.5. Unemployment has varied quite a lot, from a peak in 2009 to a low point in 2019, whereas inflation has remained quite low and stable. First, inflation failed to fall (as much as predicted, given the high and rising UE in 2009/2010), and then, after the recovery to much lower levels of UE, it failed to rise from the prior doldrums. The Phillips curve combinations of UE and price inflation are shown, over the years 2006–2018, separately for the USA, UK, Germany and Japan.

The Phillips curve is defunct; ii. Expectations are all that matter; iii. Successful monetary policies; iv. A changing structure of employment; v. Growing weight on global factors; vi. A shifting NRU.

Once every potential worker is employed a further increase in demand has to be reflected in higher inflation. And, because of the heterogeneity of industries/labour, bottlenecks and inflation are almost bound to start rising well before total full employment is achieved. What ‘well anchored’ inflationary expectations do cause is to make people assume that deviations from target inflation will be temporary, not permanent.

Lindé and Trabandt (2019) argue that a non-linear Phillips curve could explain both the ‘missing deflation’ in 2008–2010 and the slow recovery subsequently. ‘Put differently, even though economic growth may resume after a deep recession, price and wage inflation will only increase modestly until economic slack has subsided sufficiently’, p. 3. But, given the current historically low levels of unemployment in the USA, UK and Japan, this cannot be a satisfactory explanation, unless the NRU had previously shifted,

If the shocks affecting the economy were supply shocks, such as oil price variations and changes in indirect taxes, tariffs and (externally driven) exchange rate changes, then policy would have to balance the loss from disequilibrium output against the loss from failing to hit the inflation target. In order to identify the Phillips curve in such conditions of simultaneity, one would need to be able to assess and quantify occasions of supply shocks and/or monetary policy errors.

during the Great Moderation, 1992–2005, when everything macroeconomic seemed to remain stable. Thus the fact that the slope of the calculated relationship between unemployment (or the output gap) and wage (price) inflation appeared to become more horizontal in these later decades may be just an artefact of better monetary policies and fewer supply shocks, rather than representing any change to the underlying structural relationship. If this was so, then calculations of Phillips curve relationships for separate segments of a single monetary area, e.g. where monetary policy could not offset regional diversities, should show a greater (negative) slope.6 This does seem to be so, as evidenced in some current empirical work,

Whereas workers in the prime age range, 20–55 years old, mainly move between employment and unemployment, the elderly, 55–75 years old, tend to switch between employment and non-participation in the work force, e.g. in the guise of (early) retirement.

As already demonstrated in Chapter 3, participation rates in many countries have been increasing rapidly in recent years, especially elderly women, as job availability rises; (though the rise has been more limited in the USA). It appears that they can be tempted back to work quite elastically in response to reasonable opportunities regarding openings and pay. Indeed, as Mojon and Ragot, (2019), show, if you strip out the elderly from the calculated relationship, then the adjusted Phillips curve tends to fit much better.