Don’t underestimate the productivity gap

Ben Chu

Productivity pessimists should be taken seriously. For if they’re correct we should, as a nation, get used to lower growth in average living standards. But if they’re wrong, well, the reverse applies. It also implies that monetary and fiscal policy should be looser.

These are high stakes.

So it follows that we should scrutinise the analysis of the pessimists very carefully too.

Ian McCafferty of the Bank of England’s Monetary Policy Committee probably wouldn’t describe himself as a productivity pessimist. But I think that’s the implication of his recent work suggesting that only 40% of the productivity “gap” is attributable to weak demand and 60% is the result of structural changes in the economy.

Moreover, I think there are some fundamental problems with his research and conclusions.

The first problem is McCafferty’s definition of the productivity gap.

McCafferty defines this shortfall as the difference between productivity in 2007 and today. But this is an unusual definition. Most analysts, including the Bank of England’s staff in a recent report, extrapolate trend growth before the crisis and define the gap as where we would have been in the absence of the 2008-09 recession.

So while McCafferty’s work shows the gap (measured by output per worker) as around 4% of GDP, the conventional analysis puts it at around 13% as this shows:

perworker3 Dont underestimate the productivity gap

McCafferty is implicitly assuming that the trend productivity growth has been lost forever and the only element to be explained is the shortfall relative to 2007. That makes the rest of his analysis, I would argue, inherently pessimistic.

It also spits out some implausible results. Take manufacturing. As this breakdown from McCafferty’s work shows, manufacturing, which accounts for around 10% of GDP, has made no contribution at all to the productivity shortfall:

table14 Dont underestimate the productivity gap

Indeed productivity has grown in the sector (by around 3.8%) since 2007, according to McCafferty’s analysis.

But look at the trend approach broken down by the three main sectors of the economy:

productivity1 Dont underestimate the productivity gap

What this shows is that manufacturing productivity is a full 18 per cent below where it  would have been without the recession. A much more troubling picture than suggested by McCafferty’s analysis.

But is analysing the gap relative to trend really the right approach? To understand why the answer is yes consider what productivity growth means. It means firms and employees are producing more output for each hour of work. They are becoming more efficient. Look at the strong trend growth in the manufacturing sector before the financial crisis. Think of all those annual efficiency improvements in working practices by thousands of firms. Why would that trend have come to a sudden halt in 2007?

Some productivity growth before the crisis was probably a mirage – particularly in the financial sector as the Bank of England’s Andy Haldane has shown. But financial services were not driving overall UK productivity growth before the crisis. Efficiency gains were broad based, as the London School of Economics’ Growth Commission has demonstrated:

LSE1 Dont underestimate the productivity gap

What the productivity pessimists need to explain is why all these sectors unrelated to finance – business services, distribution, retailing, manufacturing etc – suddenly stopped making efficiency gains in the financial crisis and why that trend growth gap cannot be filled now that demand is strong again.

A second problem with McCafferty’s conclusions lies in his treatment of the public sector.

He argues that one should handle the raw eduction productivity figures, which suggest a 9.3% gap relative to 2007 (see Table 1 above), with care because of the difficulty of adjusting for the quality of output.

This is obviously correct. Indeed, the most recent estimates from the Office for National Statistics (which go up to 2011) suggest productivity in the education sector (adjusted for quality) has actually been growing well since the financial crisis. This is perhaps not that surprising given that schools are still educating the same volumes of pupils but with fewer resources as a result of austerity.

The ONS result are here:

ONS education1 Dont underestimate the productivity gap

But McCafferty then seems to ignore his own warning when forming his headline conclusions.

Here is his table showing his narrative explanation of which sectors of the economy, in his view, have suffered a structural supply shock and which are hampered by inadequate demand:

table2 Dont underestimate the productivity gap

As readers can see he seems to attribute the education productivity shortfall entirely to structural factors. That’s strange in light of his own caveats and the ONS’s work.

Notice, also, that he removes social work and residential care activities from his calculations of the overall balance of supply and demand factors.

So what happens if one also removes education, which accounts for 13.1% of McCafferty’s gap, as one probably should?

The answer is that it seems to push the balance of demand vs supply factors to a rough 50-50 split, rather than the 40-60 split which is the bottom line of his work.

And remember that this split is based on McCafferty’s own estimate of the overall productivity shortfall, which as we have seen is already highly pessimistic.


I don’t think McCafferty’s work is entirely without value. Some of his narrative explanations (based on reports by the Bank of England’s regional agents) for why some of the sub-sectors of the economy might have experienced a supply shock are plausible. Indeed, I actually find his narrative approach more useful than the Bank of England’s staff’s attempts to use econometrics to explain the larger trend gap.

But as a big picture explanation of the productivity gap and a guide to policy? Not so much.

Tagged in: , ,
  • bill40

    The productivity problem is 100% that of demand, lack of demand for labour and goods. The problem is the system now advocated is so insane, simple and obvious it gets overlooked.
    The neo liberals are really trying to create a consumer driven economy without consumers. They figure they’ll find them somewhere, it’s called globalisation.

  • Pacificweather

    “Think of all those annual efficiency improvements in working practices by thousands of firms. Why would that trend have come to a sudden halt in 2007?”
    1. Because there is a lag between a sales downturn and redundancies.
    2. It may be necessary to increase sales and marketing staff to improve sales.
    3.It may be necessary to take on more staff for R&D or to reorganise production operations.

    In the best schools productivity is lower than in the worst schools. Class sizes are smaller, teacher pupil ratio is higher and there are more support staff. As secondary schools are reorganised into academies employing more higher paid staff, teach first staff and a reduction in class sizes leading to higher standards. Therefore, productivity should fall to achieve higher stadards. However, ending of modular A levels will have a marginal increase in productivity.

    It takes more people to make a half hour episode of Dr Who now that it did 20 years ago because of changing production values. Similarly, it takes more people to make a modern computer game than it did to make Pacman but the income is not commensurately higher.

    If you are buying labour at the bottom of the market and you are

  • Pacificweather

    Disqus problem.

    Just because you are buying labour at the bottom of the market and paying it less than inflation pay rises it does not follow that that labour is so stupid as to not adjust its productivity to match its pay.

Property search
Browse by area

Latest from Independent journalists on Twitter