Poor productivity is only one part of the post-millennial wage disappointment story
The productivity crisis of the last few years is far from over but economic recovery is now well-established and there are at least a few flickers of life in the official data on output per hour. The widely shared assumption, often unspoken, is that when productivity picks up then typical wages will grow at the same rate. But that’s not necessarily a safe bet.
Look closely at the relationship, as a new RF report does (building on an important earlier study for RF by the LSE’s John Van Reenen and João Paulo Pessoa), and we see that since the early 1980s productivity has risen by over 60 per cent while median pay increased by just under 40 per cent: a 23 percentage point gap.
That’s a significant cleavage. It is, of course, a statement of the blindingly obvious that pay can’t outperform productivity over the long term. But it would be a triumph of hope over expectation to believe that once productivity regains its mojo the wages of the typical worker will automatically see commensurate gains.
Look again at the chart and it becomes clear that it wasn’t ever thus. In fact, two-thirds of the overall gap between productivity and pay since the 1980s developed in the decade or so after 2002. In the first part of this period (2002-2007) median pay growth slowed even as productivity continued to perform well. This was followed by the extraordinary post-crisis years in which pay fell sharply while output per worker stagnated. As a result, median pay in 2014 was as low as it was in 2002 – over a decade of lost growth.
Let’s consider the scale of the post-millennial pay disappointment – call it a ‘pay gap’ – arising from this combination of pre-crisis wage slow-down and post-crisis productivity stagnation. Assuming the established pre-2002 trends had continued, typical pay today would be a mighty £2.80 (25%) an hour higher. If we break it down, just over half of this can be attributed to poor productivity performance. That’s hugely important – and the subject of much debate. But it’s worth considering that £1.20 of the gap had nothing to do with productivity.
Economic logic dictates that there are four potential reasons for this decoupling. First, the share of GDP flowing to workers (the ‘labour share’) could fall. Second, the slice of the labour share taking the form of ‘non-wage compensation’ rather than pay could rise. Third, the distribution of pay could change so that less of it reaches the typical worker (the ‘inequality effect’). Finally, consumers might find that the costs of the goods they buy are rising more rapidly than the value of the goods they produce (the ‘deflator effect’).
In the US – where the phenomenon of decoupling is far more clear-cut and longer established – all of these factors have played a role. But the UK experience has been different in a number of ways.
The labour share, a favoured bogeyman for many leftist pundits, has in fact been fairly static over recent decades. (As such, overall employee compensation has – as theory would predict – tended to match productivity quite well). Meanwhile changes in the distribution of pay – by far the most powerful explanation of decoupling over the longer term – have only played a relatively minor role in explaining the trend since 2002.
The deflator effect appears to have been an important driver in the period since the crash, but what really seems to have driven decoupling in the period since 2002 is the rising importance of non-wage compensation (mostly due to pension contributions, though other factors, like hikes in employer NICs by different governments, have also been a factor). Of the £1.20 ‘pay gap’ explained by the recent weakening in the pay-productivity relationship: 10p is due to a falling labour share, 25p is due to changes in the wage distribution, 30p flows from the deflator effect and 55p – almost half – is due to the rise of non-wage compensation.
So much for the recent past. What does any of this suggest about the outlook for the pay-productivity relationship and – more pertinently – for the pace of pay recovery in the coming years?
There are grounds for thinking that some of these pressures – like the deflator effect – were a blip (arising from the 2011 VAT hike) that will fade away. The share of GDP flowing to labour appears structurally stable (though there are no shortage of commentators anticipating that the much hyped rise of the robots will bring a significant shift towards profit).
But when it comes to the distribution of pay, there are two more immediate considerations. At the top, high-end wage inequality was the dog that didn’t bark in the post crisis period. Will this continue? It would be complacent to assume so. If the recovery persists, and underlying patterns of compositional changes in the workforce re-establish themselves, then we might well expect high earners to fare better than the rest. Meanwhile, if we look at the other end of the pay distribution we see that the government has made an unprecedented commitment to cut the gap between the bottom and the middle; that’s exactly what a ‘national living wage’ rising to 60% of the typical wage means. This is obviously good news for low-paid workers; but it may also have the unintended consequence of loosening the link between the growth of productivity and typical pay.
What about the non-wage compensation that has driven so much of the post-millennial decoupling? Here, there is every reason for thinking this will continue to drive a significant wedge between pay and productivity. But in considering how worried we should be about this we need to distinguish between two different types of pressure. One is the result of deliberate and sensible acts of policy: such as auto-enrolment in pensions. The whole point here is to tilt more of labour compensation from today’s wages and into tomorrow’s pensions for the same workers – let’s not feign alarm if it succeeds.
This contrasts with a second driver of non-wage compensation: employers extracting funds that would otherwise have gone in wages to deal with legacy pension deficits. This pressure serves as a direct tax on today’s workers the proceeds of which are transferred to those (baby-boomers) who still work and have access to defined benefit entitlements, along with many retired pensioners. Firms are increasingly the site of an important and almost entirely unpublicised form of redistribution: just not the sort that we would necessarily favour. The implication is that a significant amount of what is termed ‘labour compensation’ isn’t actually flowing to today’s workers at all, but to yesterday’s (look out for a forthcoming report from Oxford’s Brian Bell on this). The scale of these deficits is large; the problem isn’t about to go away.
And this sets up a challenge for the decade ahead. Back in the 1980s, explosive wage inequality was partially offset by a fall in non-wage compensation (we might speculate that this was partly the product of employers failing to allocate enough to fund their pension promises). In essence, the pay pot became more unevenly distributed but less of it was leaking elsewhere. Since the turn of the century non-wage compensation has risen sharply but the distribution of pay has been broadly stable. It’s quite possible that the next decade will be a period in which both of these factors move in the wrong direction: the pay pot might get leakier at the same time as a reduced share of its contents flows to workers in the middle of the distribution.
All this matters. It obviously matters to today’s severely squeezed, poorly pensioned workers – especially the young who have by far the least bargaining power and so are most likely to be at the sharp end of these pressures. It matters as it’s another tear in the already frayed intergenerational thread at a time when the outlook for young workers is already bleak. And it matters in a broader political sense: to the degree that productivity gains aren’t equitably shared, it will only make it harder to sell proven pro-growth policies (infrastructure, new housing and immigration), which all involve a degree of social disruption. Given how risible political leadership has been on these issues, that is a worry.
Without steady productivity growth, pay won’t sustainably rise. That issue is, rightly, already up in neon lights. Yet even with it, we face big challenges in ensuring that these gains translate into typical pay rises. Some of these distributional challenges – particularly the intergenerational ones – are just as vexing as the productivity puzzle that we all dwell upon. It’s time they were illuminated too.