Why targeting productivity is a bad idea

Last week I attended a workshop entitled "Enhancing the Bank of England Toolkit," hosted by the Progressive Economy Forum. Presented at the workshop, and underpinning most of the debate, was this report from GFC Economics and Clearpoint Advisers, which was written for the Labour Party and first issued last June. The report was widely criticised at the time, as one of its authors ruefully observed in the introduction to the presentation. Nonetheless, the authors presented it unamended.

The report recommends setting a productivity target for the Bank of England in addition to its existing inflation target:
An additional target will be introduced: productivity growth of 3% per annum. The Bank of England will be required to explain how its policies are impacting upon productivity and, therefore, the potential growth path of the economy.
This target is extremely challenging. A footnote in the report notes that labour productivity growth since 1950 has averaged 2.4%, and describes the proposed uplift of 0.6% above this average as a "small increase." Forgive me, but an increase of 25% in the rate of change is not by any stretch of the imagination "small." It's an absolutely whopping hike, particularly when you take into account the fact that in the 1950s and 60s, the labour force was much smaller due to lower immigration and female participation, and the UK was rebuilding after WWII. In a mature economy which is 80% services and which has nearly full employment of both men and women, that 3% target looks well-nigh impossible.

That said, the report doesn't actually define what it means by "productivity," which makes it somewhat difficult to determine whether the target is at all realistic. So here's the OECD's definition:
Productivity is commonly defined as a ratio between the output volume and the volume of inputs. In other words, it measures how efficiently production inputs, such as labour and capital, are being used in an economy to produce a given level of output.
Simples. Or maybe not:
There are different measures of productivity and the choice between them depends either on the purpose of the productivity measurement and/or data availability. 
The OECD goes on to explain the difference between labour productivity, capital productivity and multi-factor productivity (often misleadingly called "total factor productivity," TFP):
One of the most widely used measures of productivity is Gross Domestic Product (GDP) per hour worked. This measure captures the use of labour inputs better than just output per employee.... 
To take account of the role of capital inputs, an appropriate measure is the flow of productive services that can be drawn from the cumulative stock of past investments (such as machinery and equipment). These services are estimated by the OECD using the rate of change of the ‘productive capital stock’, which takes into account wear and tear, retirements and other sources of reduction in the productive capacity of fixed capital assets. The price of capital services per asset is measured as their rental price.... 
After computing the contributions of labour and capital to output, the so-called multi-factor productivity (MFP) can be derived. It measures the residual growth that cannot be explained by the rate of change in the services of labour, capital and intermediate outputs, and is often interpreted as the contribution to economic growth made by factors such as technical and organisational innovation.
So we have three potential productivity targets: labour productivity, capital productivity, and TFP. Which of these would the Bank of England be expected to target?

If it targeted labour productivity,  there would be implications for environmental sustainability. In the absence of measures to improve efficiency of resource usage, raising GDP or GVA increases the rate at which finite resources are used. It is hard to see how this is remotely compatible with achieving carbon neutrality or sustainable use of resources.

The strong investment focus of this report might make it attractive to target the marginal productivity of capital: after all, if you are going to engineer a massive increase in capital investment, you will want to know how effective it is. The fact that the rate of return on capital has been falling for the best part of forty years could suggest that capital is not being deployed efficiently. If the measures outlined in this report significantly increased the marginal productivity of capital, the rate of return on capital should rise, no doubt accompanied by loud cheers from pension funds. But the Bank of England's responsibility is the risk-free rate: the return on risky assets is determined by the market. It is not clear to me why the Bank, or any other public body for that matter, should be targeting a 3% per annum rise in the marginal return on capital.

Total factor productivity is another potential target. But targeting it would be extremely problematic. In the Cobb-Douglas production function, you calculate the combined output of labour and capital, then multiply it by the number needed to make it agree with your GDP figure. That number is "total factor productivity" (TFP). It is a residual, or perhaps more accurately a fudge. It is usually explained as the unknown effect of technological change and other efficiency gains, but the truth is it arises to a considerable extent from measurement problems. And it suffers from the usual problem with targeting a residual, namely too many moving parts. Targeting TFP is as impractical as targeting the velocity of money.

So it is not clear what the Bank of England would be targeting. And it is also not clear how it would meet that target. The report says it would be given new credit guidance tools to enable it to direct bank lending into high value-added sectors. But the Bank would not be allowed to decide which sectors those should be:
...the Strategic Investment Board, an analytical and strategic ‘hub’, must be instrumental in outlining the sectors that should be targeted. It will be in a better position to understand the risk-reward profiles of companies. This would also ensure that any future purchases of corporate bonds by the Bank of England are consistent with the strategic investment required. 
The Bank would be tasked with directing credit to industrial sectors determined by the Government, and would be held accountable for ensuring that the sectors selected by the Government delivered the desired productivity improvement. I'm really not sure how this is compatible with the report's recommendation that the Bank should continue to have operational independence. The "new tools" would come with very tight strings attached.

The idea that a productivity target can be met entirely through credit guidance assumes that there is a simple linear relationship between the amount of bank credit provided to "productive sectors" and productivity in the economy as a whole. I don't think such a relationship exists. The last decade has shown conclusively that throwing money at businesses doesn't necessarily make them borrow. Yes, many SMEs face high nominal interest rates on borrowing. But SME lending is risky, and managing it is expensive. Lending against fixed assets, especially property, is considerably safer and doesn't have the management overhead, so it is hardly surprising that banks prefer it. It's really not good enough to beat up banks who don't lend enough to SMEs (or rather, to the "right" SMEs). If the Government wants interest rates on loans to its preferred SMEs to reduce significantly, it will have to provide guarantees.

But anyway, what does "productive sectors" mean? I found the chapter on "industrial strategy" deeply disturbing. For me, it was far too narrowly focused on science and technology, and on manufacturing.There was barely a mention of service industries in the entire report, except for financial services which the authors want cut back hard. At one point Nesta's report on the importance of creative arts was cited, but there was no recognition of the need to invest in creativity. Even more seriously, there was no mention of the the desperate need to ramp up and professionalise the care sector in the light of the approaching demographic timebomb.

The idea seems to be that the 3% productivity target could be met solely by directing credit to highly productive sectors. But this could easily result in a dual labour market - a relatively small number of highly paid people (mostly men) working in high value-added export industries, and everyone else working in low-paid, domestically-focused jobs. Do we really want our manufacturing and exports fetish to result in a German-style labour market?

Arguably we already have a dual labour market, but the high-low wage split is finance/not-finance. I don't really see how replacing one group of highly paid, mostly male workers with another, while doing nothing to raise the wages of women & less skilled men, is progress. Where jobs are concerned, services are the future for most people. The challenge is how to raise productivity and wages in service industries.

Productivity is notoriously difficult to raise in service industries. William Baumol's "cost disease" theory arose from his study of the economics of the performing arts: he observed that although a violinist in 1965 earned a lot more than a violinist in 1865, he didn't produce any more music. And in many service industries, raising productivity has unfortunate welfare consequences. Amazon's "human robots", for example. Or cleaners paid less than the minimum wage because they can't meet productivity targets. Or 15-minute care slots for frail elderly. Is this really what we want?

Realistically, productivity in service industries is not going to rise by anything like 3%, not least because the quality cuts needed to achieve it would be political dynamite. So to have any chance of achieving that demanding 3% target, either there must be a considerable shift away from services or a much smaller workforce. A large increase in capital investment could deliver a sizeable expansion of manufacturing, but there wouldn't necessarily be an associated increase in well-paid jobs. It is all too likely that the capital investment would go into robots and automation.

Furthermore, there would be a considerable temptation for the Strategic Investment Board - stuffed as it would be with people from the STEM sectors - to define "productive industries" so narrowly that funding for important service industries such as the creative arts would be  squeezed to death. This is utter folly. Creative people may not be hugely productive in themselves, but without them, manufacturing dies.

As an example of how narrow the focus is, a footnote to the report envisages that the productivity increase would come from one particular industry:
The pace of technological change suggest that governments should be aiming for a higher rate of increase in productivity than recent historic averages....The current rapid advances in the global semiconductor industry hold the key to faster productivity growth... 
Semiconductors alone are to deliver a 3% increase in productivity, apparently. That is one heck of an enormous responsibility for the semiconductor industry.

I suspect this narrow focus on high value-added manufacturing is partly driven by the Left's resurgent obsession with the trade balance, as if the UK were still on a gold standard. Indeed, the report makes some reference to "ending persistent current account deficits." Seriously, guys - a current account deficit is not a big deal for a reserve currency issuer with a floating exchange rate and a centuries-long record of meeting its obligations, and it's mostly outside our control anyway. We really don't need to return to the mercantilism of the past. Let's focus on things that make a real difference to the lives of people in the U.K., such as having a properly funded care system. And let's also focus on the vibrant creative industries in which the U.K. leads the world.

As the workshop was supposedly about the Bank of England's toolkit, I was also concerned by the lack of interest in aggregate demand management tools such as helicopter money. When someone suggested this at the workshop, one of the panellists said "Oh I don't think we will need that." Apparently there will never again be any need to support aggregate demand. I don't believe it.

And finally - a warning. The Bank of England's primary job is to manage aggregate demand, and it should have freedom to use whatever tools it sees fit to do that job. It does not, and should not have, any responsibility to deliver, or help to deliver, political promises. But the recommendations in this report would force the Bank to be instrumental in delivering a Labour Government's industrial strategy as outlined in its manifesto. This would amount to politicisation of the central bank. I think it would be a terrible error.

Related reading:

Keynes and the death of capitalism
Bifurcation in the labour market
The silent gender divide
Productivity and employment, a cautionary tale
We need to talk about productivity

Image from Hyacinth at Wikipedia.


  1. This comment has been removed by the author.

  2. Workshop should be rephrased - "New Obligations for the Bank of England".

    Sounds like outsourcing politicians duty to serve the people. But beware that EU does the same. An entity not fast enough accessible through politicians and therefore Brexit-Worthy. Finally, you can blame the central bank for even more. Outsourcing is awesome neglect play.

    I do think helicopter money isn't in their toolkit, because of media discouragement. Some economist blame central banks and fear of ZIRP that there isn't enough room or toolkits if a new recession hits to lower rates. Switzerland has negative rates. Japan has low rate for decades. But beware of helicopter money or negative rates.

    Friedman and one of his simplified thoughts that you can solve everything with the interest rate isn't really helpful. He even talked about helicopter money and maybe even more toolkits for a central bank. Biased economists, journalist accentuate only their favored friedman quotes.

  3. This is outsourcing responsibility to the Bank that should be in the hands of the government, who can then pass the buck when the Bank ends up missing an impossible target.

    If I were being cynical (and conspiracist) I would interpret this as an attempt by the Corbynite left to undermine the case for independent central banks. Instead, it is probably just a terrible idea that has caught wind amongst the circlejerk of "progressives" that dominate the left today. I don't know what's worse...

    Frances - Grace Blakely recently wrote about a HPI target for the FPC. The idea was hilariously bad but I would be interested to hear your view.

    1. Yes, I remember that. I really should pick it apart. It doesn't originate with Grace - it has been around for much longer than that. Apart from being pretty impractical, it is also enormously regressive. I pointed this out to her but she didn't seem all that bothered. Perhaps she is comfortable with telling the plebs that they shouldn't aspire to own their own homes. I am not.

  4. Peter Rice, Clearpoint Advisors17 May 2019 at 15:11

    We wanted to follow up on a few inaccuracies in the blog. It is a feature of a long report that readers miss some of the details. To ensure that there is an accurate recollection of the report we set out the following points and clarifications. We are setting out the response in a set of comments due to the minimum character count for each comment.

    Statements from the blog are numbered with our comments following:

    1. "That said, the report doesn't actually define what it means by "productivity," which makes it somewhat difficult to determine whether the target is at all realistic."

    Perhaps we should have been clearer. The standard definition of productivity is output per hour worked. This is labour productivity. In footnote 1, we compare our target with the labour productivity growth rate since 1950s. So it is a clear assumption that the ‘productivity target’ would be labour productivity, given this is what we are using for the purpose of our argument. No mention is made elsewhere in the report of any other definition of productivity.
    2. "The fact that the rate of return on capital has been falling for the best part of forty years could suggest that capital is not being deployed efficiently."

    For the record, rate of return on capital has not been significantly falling. See:
    The service sector has been seeing increases in rates of return. Manufacturing has had a reduction – Globally rates of return on capital have not been falling.
    (See also: “Why have interest rates fallen far below the return on capital” Marx, Mojon and Velde, January 2018, Federal Reserve Bank of Chicago)

    Other than a small dip post financial crisis, real rates of return have increased in the US. To that extent we agree that capital in the UK by comparison to the US is not being efficiently deployed hence our focus on an industrial strategy for the UK.

  5. Peter Rice, Clearpoint Advisors17 May 2019 at 15:14

    hence our focus on an industrial strategy for the UK.

    3. "If it targeted labour productivity, there would be implications for environmental sustainability. In the absence of measures to improve efficiency of resource usage, raising GDP or GVA increases the rate at which finite resources are used. It is hard to see how this is remotely compatible with achieving carbon neutrality or sustainable use of resources."

    We clearly disagree with your analysis.
    a) increasingly services and the intangible economy is what drives the economy. We discussed this at length at chapter 6 in the report and highlight the extent of the services economy value add in Chapters 7 and 9. We see an increase in the ability to combat climate change issues with advances in technologies to support decarbonisation.
    b) cost of renewable energy falling sharply. There is an inflection point at which is becomes more profitable for companies to invest in and use renewable energy sources.
    You can have as much productivity growth (remember, that's got nothing to do with the size of the workforce, as it's a per hour worked measure) as you want, and still be environmentally sustainable. The two things are not incompatible.

    In our view this then leads on to the other confusion below, where the difference between GDP per capita and GDP are confused:

  6. Peter Rice, Clearpoint Advisors17 May 2019 at 15:15

    4. “It's an absolutely whopping hike, particularly when you take into account the fact that in the 1950s and 60s, the labour force was much smaller due to lower immigration and female participation, and the UK was rebuilding after WWII. In a mature economy which is 80% services and which has nearly full employment of both men and women, that 3% target looks well-nigh impossible.”

    Obviously, the size of the labour force has nothing to do with productivity growth or GDP per capita. And it does not matter whether you are at full employment. The question about shift to service industries is more interesting. In fact the gains from technology are significant. We spent all of chapter 6 on this very important point.
    As was stated in the report:

    “In truth, quantum computers, robotics, 5G and blockchain will all to varying degrees offer the potential for significant productivity gains. Recent developments in the semiconductor industry suggest that the pace of innovation is, if anything, likely to accelerate too.
    The challenge for policymakers is clear. Better data will be needed to ensure that investment, both public and private, produces the best returns. The fourth industrial revolution is not being measured accurately.”

    We go on the explain what is happening with the technology revolution. It is more than just efficiency gains which are themselves significant. It is also about new services that can be offered because of for example 5G. these very services that are productivity enhancing are also offering ways of doing business that are sustainable.
    With the internet of things “IoT” and 5G, we can in real-time monitor pollution, business waste, power management, home heating efficiency, control traffic flows – that data will inform government and private sector about where to put their efforts to aid in decarbonisation and business sustainability. In the energy sector alone AI is allowing huge efficiency gains in how it manages systems for energy use. We are talking about significant factor gains for saving power. The plans we outlined in Chapter 5 outline how an industrial strategy would support productivity, sustainability and decarbonisation. For example in the UK, enhancing energy systems will drive forward energy efficiency and economic growth. We can have decarbonisation and growth.

  7. Peter Rice, Clearpoint Advisors17 May 2019 at 15:16

    5. “There was barely a mention of service industries in the entire report.”

    We refer you to the above. We spent the bulk of our time discussing technology and its impact on services. In particular we gave emphasis to professional, scientific & technical services throughout, as well as many other service sectors.

    6. “Realistically, productivity in service industries is not going to rise by anything like 3%, not least because the quality cuts needed to achieve it would be political dynamite.”

    You get productivity improvements through quality enhancements, not cuts. If you improve the quality of a service, and pay the same in nominal terms, real output increases. This is a productivity improvement. Of course, there are issues in measuring quality, which is touched on in the paper, but for even small businesses the advances in the likes of Xero accounting services, the apps on phones that speed up administrative tasks, google for research, public and private sector provision of information for free - cloud storage with high security that even small businesses can afford, the use of for example Microsoft 365 for all your applications – which are automatically updated and managed at such low cost. This has allowed small businesses in any sector of the economy to be more efficient and at significantly lower costs. It is not about cost cutting - its about enhancing work - making work more meaningful and improving your services or manufactures.

    7. “And in many service industries, raising productivity has unfortunate welfare consequences.”

    The gig economy and low pay are not a feature of raising productivity – it is a business model of some companies we do not support. In fact ‘gig’ economy companies tend to reduce productivity.
    The promotion of low wage jobs does not support productivity gains. Low wages are in low output sectors of the economy.

    8. “Semiconductors alone are to deliver a 3% increase in productivity, apparently. That is one heck of an enormous responsibility for the semiconductor industry.”

    Semi-conductors are very important, but the industrial strategy proposals and our report in general gave a much wider view of how we could see sustainable and productivity growth in the economy. The industrial strategy would be a key driver in the support of productivity growth as we transform infrastructure and support public and private sector sustainability and decarbonisation – often driven by technological advances.

    You miss the incredible flow-on effects of semiconductors. Our reason for focussing on semiconductors is because they are a general-purpose technology “GPT”. Another example of a GPT is electricity.

    From 1974 until 2013, computing (i.e. semiconductors) contributed around 1/3 of annual labour productivity growth (See: “Is the Information Technology Revolution Over?” Byrne, Oliner, and Sichel (2013)). In fact they view the advances in semi-conductors as likely to mean that the pace of labour productivity growth will exceed the long-run average. Since their paper was written, we are seeing that semi-conductor advances have again speed up, enabling another technology, AI to also be a GPT. The combination is hugely advantageous - a Labour government would want ensure that the benefits of such transformations are more widely shared.

    1. I shall respond in detail to this long series of comments in a second blogpost. However,
      I must here rebut your accusation that I am "confused" about the nature of productivity. I assure you I am not.

      Firstly, let me correct your mathematics. I refer you to the way in which labour productivity is generally calculated. For the whole economy, labour productivity is GDP/total hours worked. This is not the same as GDP per capita, which is GDP/total population.

      If labour productivity is calculated as GDP/total hours worked, then it must depend on the size of the population, since there are only a limited number of hours in a working day. Obviously "population" includes inactive people (children, old, ill), but unless the proportion of inactive is very high, a large population will work more hours in total than a small one. Unless the GDP of the larger population is proportionately larger than that of the small population, labour productivity must be lower.

      For example, imagine we have a population of 100,000 which has 20% inactive, and another population of 50,000. Both populations have 20% inactive, and the working proportion of both populations does a standard 40-hour working week for 50 weeks of the year (this isn't terribly realistic but it keeps the numbers simple). The first group has total hours worked per annum 80,000*40*50 = 160,000,000. The second group has 40,000*40*50 = 80,000. Let's imagine both groups have output per annum of £5bn. Labour productivity of the first group is 5,000,000,000/160,000,000 = £31.25 per hour. Labour productivity of the second group is 5,000,000,000/80,000 = £62.5 per hour. The labour productivity of the smaller group is therefore twice that of the larger group. GDP per capita is also twice as much, but that is irrelevant as far as productivity is concerned.


    2. /2.....

      Now, to your point that the size of the labour force does not affect productivity GROWTH. Up to a point this is true: adding extra workers shouldn't change the rate of output of each worker. However, the management overhead becomes larger as the size of the workforce increases, which can slow the rate of productivity growth. And it is possible to reach a point where increasing the size of the labour force actually reduces productivity. As the old adage says: "Too many cooks spoil the broth".

      Although the absolute size of the population doesn't necessarily affect productivity growth, its composition does. The re-entry of women to the workforce not only changed the size of the working population, it also changed its composition. As a direct result of women entering the workforce, services such as child care and domestic cleaning started to make a significant contribution to GDP. Furthermore, other compositional effects (eg rising proportion of elderly) result in rising demand for services such as healthcare and social care. Much of the immigration since WWII has been driven by growing demand for labour in service industries. Compositional effects matter.

      I also said that in a services-dominated economy which has nearly full employment of both men and women, productivity growth might be hard to come by. This is because of the nature of services, as I explained in the post. You dismiss the "gig" economy and low-wage service sectors as "things you don't support". This will not do, I'm afraid. These sectors rely on large numbers of cheap labourers, and they are an increasingly important part of GDP. And you rightly point out that they tend to reduce productivity. Productivity growth in high-tech services would have to be absolutely astronomical to offset the drag that these labour-intensive services exert on overall productivity growth. I think you are expecting far too much.

      The reason I mentioned nearly full employment of women is that historically, increasing output in service industries by raising the technical skills required to do them has almost always disproportionately driven out women (nursing is an honourable exception). You need a strategy to raise productivity in labour-intensive low-wage sectors without adverse consequences for the employment of women and minorities. It is very sad that a report which aims to solve the UK's productivity problem has no answer to the proliferation of low-skill, low-wage, low-productivity industries in respond to rising demand for cheap personal services, which in my view is a large part of the productivity puzzle.

      In summary, I have not confused GDP with GDP per capita, and nor have I confused labour productivity with workforce size.


Post a Comment

Popular posts from this blog

WASPI Campaign's legal action is morally wrong

The foolish Samaritan