Sign up for our weekly Top of the Charts emails here
Top of the Charts is in a reflective mood this week – what with tomorrow marking the 10th anniversary of the collapse of Lehman Brothers. So, instead of the normal selection of pieces, I’ve written a long read on both what caused the unprecedented post-crash pay squeeze that we are still living with the consequences of today, and what lessons this holds for policy makers. The RF team have edited out the actual flashbacks to financial crisis trauma – so hopefully this makes for an interesting but not scarring read.
And fear not – normal service will resume next week, with a party conference flavour.
All the best,
Director, Resolution Foundation
The cause of, and lessons from, Britain’s pay squeeze
Over a decade on from the start of the financial crisis we are still living with its defining feature: a truly enormous pay squeeze that started in April 2008. The average weekly pay of workers in Britain fell by 6 per cent or £31 between then and the end of the pay squeeze in 2014. So what caused that squeeze1, and what does it mean for policy makers today?
Of course at one level we know the cause was the financial crisis. However, that answer is insufficient. It fails to explain why the pay squeeze was such a central feature of this recession – but nowhere to be seen in previous UK recessions – or why our pay squeeze was so much deeper than that seen in any other major developed country. For context, between 2010 and 2014 real hourly pay declined by 7.1 per cent in the UK, second only to Greece (-15.1 per cent) in the developed world, and much more severe than the US (-2.8 per cent) or Germany (-2.1 per cent).
When (entirely unexpected) pay falls started emerging in 2008, the immediate focus was on the idea that a significantly more flexible labour market was leading to workers negotiating nominal pay cuts as part of an enlightened/Germanic move to avoid bigger increases in unemployment. To economists used to explaining unemployment in recessions as the result of nominal rigidity (ie where workers refuse to accept actual cuts in the pounds and pence of their wages), it made theoretical sense that if unemployment was not rising it must be because nominal rigidities were being overcome. The only problem was it wasn’t true.
Yes, Britain has seen a big rise in the share of workers each year getting a nominal pay cut. But this growth took place in the early 1990s, not in 2009.
As we moved further into the post-crisis period, and the debate around pay moved into the centre of politics, new arguments were put forward for why real wages had continued to fall after economic growth had returned. Two US-inspired inequality arguments gained traction; rising earnings inequality and the idea that profits were being protected, leading to the labour share of national income falling.
First the rising earnings inequality argument hypothesised that typical workers were seeing pay cuts while higher paid workers were gaining. The data offers little support for this argument. Britain’s pay squeeze was pretty widespread – which is why it’s such a big deal politically. The only interesting distributional story on earnings during this period is that lower paid workers did better than most on an hourly basis (due to the minimum wage) but worse on a weekly basis (due to rising underemployment).
The second argument was that the rich were able to increase their share of income going to profits during this period, suppressing pay (aka the labour share). Again, not true. As you normally expect in a recession, labour share initially rose (profits fall before firms get round to firing workers), and has then declined back to around its pre-crisis level.2
So while earnings inequality is too high and workers should have more bargaining power, neither of these arguments have anything to tell us about the pay squeeze. While the divergent paths of GDP and wages in this period was striking, it simply reflected the fact that growth was driven significantly by increases in the size of the workforce – not least because of migration. As others have noted, that’s why we had an economic recovery but not a wage recovery.
Amongst these complex but flawed explanations of pay falls something more obvious has got missed: it isn’t just about what we earn, but what we can buy with it. Between 2010 and 2014 the UK saw inflation of 12 per cent – a third higher than the 9 per cent in the US, and far above 7 per cent in. Germany. Our higher inflation was driven by a simply staggering devaluation. Sterling’s value fell by 27 per cent in the financial crisis – a much bigger fall than the depreciation of the early 1990s as we exited the ERM.
This depreciation meant that the cost of imports rose significantly (we needed more pounds to buy the same German car or French cheese as we had pre-crisis), along with higher VAT3 pushing up inflation (which peaked at over 5 per cent in the autumns of both 2008 and 2011) and eating into real wages. This was a big part of how Britain adjusted to being a poorer country as the financial crisis hit our national income.4
Nominal rigidity in wages didn’t lead to the huge unemployment we had expected because higher inflation delivered real pay falls instead. In marked contrast to the 1970s, British workers now lack the bargaining power to ensure their wages keep pace with prices they paid in the shops. This role of depreciations in driving pay falls was largely missed, with a few honourable exceptions (see here, here and here), until we decided to have a repeat of the same phenomenon as the pound slumped, inflation rose and wages started falling again in the aftermath of the Brexit referendum.
Now of course inflation does not explain the whole story of why Britain’s pay squeeze was second only to Greece amongst the OECD. The sheer scale of our recession is the second (and blindingly obvious) factor.
Indeed in a simple model of the relationship between real pay, inflation and output, the combined role of higher inflation and a deeper recession explain all of the exceptional depth of Britain’s pay squeeze.5 Across 15 of Europe’s major economies and the US the model does a very good job of predicting the decline in real wages. Using this model we can explore what would be the case if the UK had not suffered exceptionally high inflation. Assuming UK inflation had been an average of the countries below, the country’s pay squeeze would not have been particularly abnormal: just -1.4 per cent, rather than the actual -7.1 per cent.
Notes: Estimates based on a simple regression model of the form: Change in real hourly pay (2010 – 2014) = Change in GDP (peak to trough) + Inflation (2010 – 2014) + ε
Understanding the role of lost output alongside inflation in driving the squeeze is crucial. It means it is very dangerous to conclude, as some have, that “growth in average earnings has decoupled from general economic growth”. In fact it’s the lack of growth in output per worker (or productivity) that has given us not only the depth of the pay squeeze but also incredibly weak wage growth ever since.
So what are the lessons for policy makers from the trauma of a wage squeeze we’d all rather forget, other than obviously avoiding financial crises in the first place? Three stand out.
First, we all need to pay more attention to the exchange rate. We used to worry about it when we went abroad but we now know it matters for what we can afford, not just in Paris or Berlin, but in Portsmouth or Burnley. As Chris Giles argues, rather than championing big devaluations as some sort of route to a manufacturing renaissance, we should be avoiding them because they tend instead to be a signal that we are poorer as a nation. The exchange rate is now a mechanism for changes in our national income to feed through very quickly indeed to changes in household incomes, in certain circumstances.
Second, you could also conclude from this that big exchange rate movements in a crisis are a bad thing. But wait – be careful what you wish for. The alternative to an exchange-rate driven pay squeeze would have been a surge in unemployment.6 Yes, pay squeezes are painful. But that pain is at least shared. A quick comparison of the US and UK financial crises experiences bears this out.
The UK had a deeper initial recession than the US and a much worse productivity (and wage) performance during this period, with the US actually seeing fast productivity growth between 2008 and 2011 as unemployment rose. But this doesn’t mean the typical US household got the better deal – the performance of incomes in the UK and US was almost identical despite lower output per capita in the UK And crucially the poorest did slightly better in the UK (partly offset by slightly higher price rises than the middle and top faced).7 This reflects the UK’s much stronger employment performance – which disproportionately benefitted poorer households. In 2008 both the UK and US had employment rates of 72 per cent, but by 2014 UK employment had reached 73 per cent while the US employment rate remained under 69 per cent. Sharing the economic pain of the crisis via the pay squeeze in 2010s hasn’t made for a happier country, but it’s a fairer alternative to the 1980s, when more than three million people were unemployed in each and every month for a full five years.
The third lesson is that we should not simply assume the next crisis will be like this one and that pay recessions are here to stay. Britain avoided the huge surge in unemployment in part because our exchange rate was able to weaken by a staggering amount. That reflected the fact that the UK was harder hit by the financial crisis than almost anyone else. If the next recession is more synchronised, or one in which other countries were hardest hit, our exchange rate would not fall and we might once again see an unemployment recession of the kind we expected back in 2008.
This matters for how we think about being prepared for a future recession – here are two examples.
First, in the practical business of assisting the unemployed, Jobcentres are the frontline in the UK. Those very Jobcentres are however currently having their role expanded to deliver support (or conditionality) to those receiving social security who are in work. If that (very resource intensive) new role is rolled-out before the next recession hits we need a plan for rolling it back away again overnight and refocusing on helping get people back to work.
Second, whether we have a pay or an unemployment recession matters for our macroeconomic policy response to it. Our automatic stabilisers (fiscal instruments that help counter swings in the economic cycle without intervention by policymakers) will behave differently in each case. Our tax credits system cushioned earnings losses for poorer households in the financial crisis, but far less support would be delivered by the meagre £73.10 a week we give people that lose their job entirely. Weaker automatic stabilisers in a different crisis would require more active moves by the Treasury and Bank of England to compensate.
The pay squeeze was the most surprising and in many ways long lasting effect of the financial crisis – the defining economic event of any of our lifetimes. Our horror at its effect on our living standards should not get in the way of our attempts to understand why it happened and what lessons that has for us in the years ahead. The UK’s pay squeeze was so exceptional compared to other countries because of the combination of the sheer depth of our recession (an unalloyed bad thing) with a new (and much more ambiguous) feature of our economy – a devaluation led inflation spike that we have now experienced twice in this decade. So let’s pay attention to the exchange rate. And plan for the next crisis.
1 Note this is a different (albeit related) question to the one dominating economic discussions at present: why is pay growth today so weak when unemployment has reached record lows? This is a crucial question, and one we will be returning to with new research shortly. A further related question is the cause of the pre-crisis earnings slowdown – covered here.
2 Recent ONS data revisions have undermined some previous arguments about declining labour share post-crisis
3 Higher VAT is the clearest direct effect of austerity on the pay squeeze, leaving aside arguments about the macroeconomics. Public sector pay restraint has also played an austerity related role in weak pay over the past decade. But while pay caps of 2 per cent did begin during the course of the pay squeeze they were not its drivers and public sector pay outperformed private sector pay for the initial post-crisis period. The later move to public sector pay freezes and then a 1 per cent cap has left public sector pay falling further than private sector pay more recently.
4 The effect can be seen clearly in the fact that while our consumer price inflation (measuring inflation amongst the goods and services domestically consumed) was exceptionally high, the UK’s GDP deflator (measuring inflation amongst the goods and services domestically produced) was largely identical to that in Germany and the US – so what we are able to consume domestically given our incomes (ie what we produce domestically) fell.
5 The model has a good fit (R2 = 0.5) and does a good job of predicting the UK’s pay squeeze. Model estimate is 6.7 per cent compared to actual 7.1 per cent. We then estimate the equation again, substituting UK inflation (11.9 per cent) with an average of the other countries in the model (7.4 per cent) and predict the real pay squeeze.
6 This is of course an oversimplification with wider policy choices, not least regarding the nature of a country’s labour market having a big role to play.
7 Lower income households experienced higher inflation than average in the crisis (unlike in the post-Brexit inflation spike) but over the time period as a whole the effect of differential inflation will have been relatively small in comparison to the living standards gains from the UK’s stronger employment performance and particularly the way in which these employment increases in the UK were concentrated among lower income households.