After years of monetary stimulus, zero interest rates and quantitative easing, the global economy is now experiencing strong, synchronised growth. Finally, some 10 years after the global financial crisis, after the US housing crisis, the Euro zone currency crisis and a challenged time for the financial system, economies are growing strongly almost everywhere, and the leading indicators suggest we are set fair for the medium term. In addition, President Trump has taken the somewhat surprising decision (given where we are in the cycle) to provide additional fiscal stimulus to the US economy, stoking what would appear to be an already hot fire.
Markets have been supported by a variety of factors – zero interest rates and low government bond yields being the most important – but economic and company fundamentals are now taking over as the driver to equity and credit markets. However, with the growth backdrop, all eyes are now on the inflation outlook and the consequent central bank response. With the US Fed now raising rates, markets are potentially finely balanced, and a sharp uptick in rates and bond yields in response to an inflation shock has the potential to challenge market valuations in a way we have not seen for some time. Therefore, the near-term focus has to be on inflation, its drivers and where we go from here.
All things being equal we would expect wage growth to be the main driver of inflation. However, a variety of structural disinflationary forces are likely to overcome any impact from economic activity in the short- to medium-term to dampen inflation. Thus, our own forecasts are for benign inflation levelling out at 2 per cent in the US and around 1.5 per cent in the Euro zone in 2018 – and this is in part due to these disinflationary factors, which are many and varied. Crucially, they are having a meaningful ongoing impact on employment and wage growth, to the extent that the Phillips Curve economic model no longer works.
The Phillips Curve, used by central banks, describes the inverse relationship between the unemployment rate and consumer inflation – the simple rule being the lower the unemployment rate, the greater the inflation, as labour bargaining power increases as labour availability decreases. That relationship held from the pre-war era through to the end of the 1960s, but dislocated in the 1970s era of high unemployment and high inflation, partly caused by oil shocks. In the 1980s this flattened and we now have the current environment of prolonged low inflation alongside very low levels of unemployment. Wage growth has stalled.
One could argue that the Phillips Curve is an over-simplistic way of understanding how wage growth is generated, and one of the factors that may have been underestimated in this model is the importance of inflation expectations. After such an extended period of ultra-low inflation there is an argument that the expectations of workers in their wage bargaining behaviour has reduced. We have seen this in Germany, where there is a very tight labour market and yet wage demands have been lower than expected. After a long period of low inflation this has now arguably been baked in to expectations.
The demise of public sector wage growth is also a factor, especially in Europe where, since the crisis, we would normally have expected a strong fiscal response to low growth, but in fact the trend has been in the other direction with a period of austerity and fiscal restraint after the crisis, which has kept public sector wages low. This has also had a spillover effect on the private sector.
The degree of underemployment in the economy is high, as indicated in Figure 2 showing Euro zone employment and hours worked since the crisis: while employment has been very strong in the Euro zone, the number of hours worked has simply not caught up.
This is partly the rise of part-time work and flexible contracts, as well as labour market reforms in Europe, which now make it easier for companies to hire flexible workers. In Europe and in the US we are seeing a trend of increasing involuntary part-time work; while in Japan non-regular employment occupies a larger share of total work than it did pre-crisis, and increased female participation in the workforce is an accompanying trend. The increasing flexibility of labour markets has resulted in jobs that typically pay less and are less secure – clearly, people in part-time jobs are less likely to demand higher wages because their positions are more precarious.
There is also an understandable economic phenomenon where companies are not hiring full-time workers to the same degree because uncertainty about the future remains higher post-crisis. In the US in particular the participation rate has fallen steeply since 2008-09, a result of the skills of people who became redundant no longer being in demand and the industries in which they worked being drastically reduced. This means the proportion of people out of the labour force for more than a year has risen relative to the rate of unemployment.
While participation is so low, the pent-up supply of labour can respond to an uptick in employment and undermine the wage demands of those already working. Perversely, this might arguably have already happened and been responsible for holding down euro area wages. Average regular pay in the UK, for example, may have risen marginally in recent weeks, but is still below the pre-crisis peak of GBP 473 per week in March 2008.
When the Phillips Curve was being developed in the late-1950s labour markets were overwhelmingly domestic, but in recent decades developed market economies have had access to a much broader global labour supply. Globalisation has allowed corporates to source cheap labour from around the world, enabling companies to increase profit margins through efficiency savings.
Crucially, the multi-decade bull market in bonds has been a by-product of globalisation, with the rise of the global labour force playing a key role in creating a disinflationary environment leading to falling neutral real interest rates, tumbling labour costs and lower bond yields.
Clearly, workers in emerging economies have benefited from this trend – in China alone millions of people have been lifted out of poverty, with China’s per capita income increasing fivefold from $200 to $1,000 between 1990 and 2000, and again from $1,000 to $5,000 between 2000 and 2010.1
There is no doubt that technology and innovation is improving people’s lives. Whether it is making everyday tasks simpler and quicker, freeing up more time for us to spend with our friends and family, or improving healthcare, the benefits are clear. We can order goods and services from wearable devices, the sun and wind are increasingly powering our planet, and the world has never been better connected. But with technology comes an increasing fear that ‘automation’ is already having a negative impact on jobs as robots take paid employment from people.
These fears are not unfounded.
Research conducted by Deloitte and thinktank Reform in October 20162 predicted that 861,000 UK public sector jobs could be lost by 2030 through automation, which would save GBP 17 billion annually in public sector wages compared to 2015.
In September 2013,3 a study by Carl Benedikt Frey and Michael A. Osborne for the University of Oxford looked at 702 occupations in the US and used a unique methodology to work out which jobs would be most at threat from ‘computerisation’. The authors found that 47 per cent of total US employment is under threat, with the most at-risk occupations being telemarketers, insurance underwriters, watch repairers, library technicians, loan officers and credit analysts.
Thus far, automation has arguably increased employment, as the amount of activity that automation generates is enough to create net positive jobs growth, but it hasn’t thus far generated productivity or wage growth. Why? In order to meet labour demand, the ‘robot army’ workforce is merely being added to the existing pent-up labour supply, which means corporates are not currently having to concede wage growth. This may change, but not in the short term.
Also, productivity, while slow in general, has increased mainly in areas that are not employment intensive, such as technology and information, which have been the real productivity winners.
By contrast, areas that require larger numbers of employees, such as personal care, manufacturing, services, leisure and restaurants, have seen lower wage growth. This has also created a way to disenfranchise labour – job creation in technology, for example, is rarely in the technology itself but in ‘gig economy’ roles and short-hours contracts. It is labour selling itself not by the hour or the week or month, but by the gig – and that is clearly less beneficial to labour and more beneficial to capital.
This ultimately means wage growth is seen mainly at the top end, while the skills gap means that mobility from lower-paid jobs to higher-paid jobs is ever-widening. This may change over time as digital natives become a bigger part of the economy, but again this is unlikely in the short term.
Unionisation across the Euro zone has declined from around 30 per cent to 15 per cent, further eroding the bargaining power of labour. Moreover, immigration in Germany recently and in the UK since the mid-2000s has added to labour supply, while demographic effects such as people working longer and retiring later has also had an impact on wages. We have also had three decades of positive shocks to the labour force (baby boomers in the 1960s, the emergence of the Soviet Union in the 1970s and an increase in workers in China in the 1980s), which has added to the labour market.
All of this suggests that the labour market is much better at absorbing employment growth than it has been in the past, creating less friction and less upward pressure on wages.
Will any of these factors flip?
If globalisation, automation, underemployment, unionisation and other factors have created a dynamic where wage growth has stalled despite rising numbers of people in work globally, we must ask how long these factors might continue? In our opinion, it is likely the above trends will persist and the Phillips Curve will remain subject to these overwhelming forces in the medium term.
Globalisation became something of a touchpoint globally with the UK referendum on EU membership, the election of Donald Trump, and the Italian referendum. There is a sense that progress at any cost has benefited wealthy innovators who have introduced new technology.
and created new sectors of industry at the expense of ‘ordinary’ people who have seen wages stagnate. Brexit and Trump have shown us that the public are not beyond rejecting globalisation in their millions, and if globalisation is stopped, the forces of disinflation that supported the 35-year bond bull market could be undermined, threatening the valuations of all financial assets.
However, deglobalisation pressures appear to have ebbed away, with less geo-political risk now than we had 12-18 months ago, and labour is still relatively cheap versus capital and automation. But the threat of trade protectionism remains, and even has renewed momentum as President Trump seeks to introduce tariffs – if left unchecked these could make emerging market workers more expensive to developed market companies and enhance developed market labour bargaining power. But this is likely to be a medium-term impact, if at all.
That said, there is a strong argument to suggest we might still currently reside in the normal lag you would expect between economic growth and wage growth. Figure 3 indicates that the gap between unemployment and wages in the US is not unusual in historical terms.
We also expect some transitory factors to drop out of inflation, such as the steeply lower contribution from the telephones services area of CPI that occurred in the US last year, the depression in US used car prices, and lower rents. Also, as the housing market in the US heals we may see mortgage supply becoming less constrained, and this might ease job mobility over the medium term, which could add to labour bargaining power. Meanwhile, immigration has arguably peaked in the UK and Europe, which would also remove exogenous supply shocks from the labour market.
We have also seen a couple of warning signs about inflation – such as a slight uptick in wage inflation in the US in January – which have raised the prospect of interest rates rising more aggressively. This led to the recent short-term spikes in volatility we saw in global markets in February. Clearly, there are heightened tensions following a nine-year bull market that few participants seem to have enjoyed.
While we do not expect to have to revise our inflation expectations upwards, we are alert to any signals that would challenge this view.
If inflation is to break out of a comfortable range, that might be best indicated by US 10-year or five-year break-evens. Since 209, US 10-year break-even inflation has been in a range between 1.4 per cent and 2.5 per cent, and is currently in the middle of that. Pre-crisis it was at a level of around 2.5 per cent, so in a scenario where breakeven inflation rises to 2.5 per cent the market could either decide this is within the normal range or that inflation is poised to break out. We are watching these and other signals closely.
We have had a world of low growth, low inflation and low rates, prone to volatility and headwinds. Markets have largely been driven by monetary policy, against which risk assets have performed well. With this as a backdrop, somewhat unusually after nine years of economic expansion the current political regime in the US has decided to loosen the fiscal reins and give the economy a boost, which has the potential to be inflationary. However, the deflationary forces we have examined are still widespread and are likely to continue to depress wage growth. So while the risk of inflation has risen modestly, wage growth in the short- to medium-term should remain at around two per cent to three per cent, helping to keep inflation in cheque and within our forecast range.