Progress | Centre-left Labour politics

Productivity is fine. Let’s count and tax it better

Non-financial productivity gains need measuring through ‘gross domestic utility’, not gross domestic product, otherwise public services and social justice will suffer, argues Peter Kellner

Something weird is going on. We need to do some delving to get to the bottom of it, but its impact on our living standards, government finances, public services – indeed the future of our politics – could be profound. Here are two apparently contradictory truths that illuminate the problem.

First, over the past 10 to 15 years, our lives have been transformed in many ways, large and small. Our smartphones, computers, television sets, cars, light bulbs, and a myriad other things are far better than they were. We buy far fewer newspapers, CDs, DVDs, maps, encyclopaedias, postage stamps et cetera. We talk at length to our friends abroad via Skype or FaceTime. And so on: the list can be extended almost indefinitely. Human ingenuity is generating new and better goods and services as fast as ever.

Second, over the past 10 to 15 years, the productivity data from around the world show a marked slowing down. Note the geography and the timescale. This is not just a British phenomenon; and signs of the slowdown predate the collapse of Lehman Brothers in 2008. The era of austerity and low interest rates may well have intensified the problem but did not cause it. The forces at work go deeper.

There is plainly a conflict. If we had no statistics, our lived experience would tell us that our economies are more productive than ever, thanks to the revolutions in the world of technology. On the other hand, if we had only statistics, and ignored our lived experience, we would deduce that in the early years of the last decade, the post-1945 era of steady productivity growth across the developed world came to an end. How do we reconcile these two observations?

The reason for asking it is not just to resolve a statistical curiosity. As Paul Krugman, the Nobel prize-winning American economist, puts it: ‘Productivity isn’t everything, but in the long run it is almost everything’. It is the main source of economic growth and higher living standards. By generating higher wages and more profitable companies, it ensures higher tax revenues – and thus more money for schools, hospitals, pensions and all the other things the government does. If productivity stops rising, then almost all of us lose out in a host of ways.

This was brought home in the run-up to Philip Hammond’s latest budget. The Office For Budget Responsibility told the chancellor that he would have far less money in the years ahead than it, and he, had expected. And the reason is that it had changed its mind about the United Kingdom’s productivity story. Ever since the OBR had been set up in 2010, it had assumed that productivity would soon bounce back to its historic average of two per cent a year growth. Each year since then, it has reported that the figures were stalled – but, do not worry, the bounce-back was just around the corner; prosperity and buoyant tax revenues would soon return.

Robert Chote, the OBR’s chair, deserves great credit for candour. He did not try to hide or downplay his change of mind. He said this October: ‘Our assumption that productivity growth would return to a more normal rate within a few years reflected a judgement that whatever factors were depressing it in the wake of the financial crisis would fade as it receded further into the past. But as the period of weak performance gets longer, the explanations that people pointed to immediately after the crisis [for an upturn] look less convincing and others seem more plausible.’ In other words, do not expect the gloom to lift any time soon. The OBR’s latest budget day forecast confirms this outlook. Seldom has the science of economics seemed so dismal.

Source: OBR

What, then, is happening? When I was at university, one way we were taught to untangle knotty economic issues was to imagine a country that had just two products. Reviving that approach, let me take you to Nirvania, a land of milk and honey – and nothing else. Until last year, half its workers earned £20,000 a year milking cows and selling their creamy, full-fat output. The other half earned £30,000 a year keeping bees and selling honey. Overall, the average income and output of Nirvania’s workforce was £25,000 a year.

Then one of Nirvania’s brighter sparks returned from a trip abroad. She had learned of a way to give every family their own beehive that generated as much honey as people wanted, automatically and without cost. While this seemed like a great idea, it provoked an upheaval in Nirvania’s economy. Suddenly bee-workers (not to be confused with worker-bees) lost their income – for there was no reason for anyone to buy what they could get for free. However, Nirvania, being a highly efficient and humane society, solved this problem by converting bee-workers, magically overnight, into £20,000 a year milk-providers.

I am delighted to report that Nirvania is now more prosperous than ever. It produces and consumes twice as much milk – and also twice as much honey, as the home beehives have been a stunning success. Nirvania’s overall productivity – and hence economy and living standards – has doubled, for the same number of workers now provide twice as much of both its products as before.

Here is the rub. The ONS – the Office of Nirvanian Statistics – shows something very different. On their calculations, per capita income has been reduced from £25,000 to £20,000 a year. Far from doubling, the measured economy, and productivity, have contracted by 20 per cent.

The key to this contradiction is the term ‘measured economy’. Economic statistics count only transactions where money changes hands. Anything that is provided free does not count. It follows that whenever we can access for free something we used to pay for, then – in the absence of any other change – the measured economy will contract. This is so, even if we consume far more of what we obtain for free, precisely because we no longer have to pay for it.

Let us now return from Nirvania to the real world. In the past whenever we sent a letter through the post, we added to gross domestic product by buying a stamp. Today when we send an email, GDP remains unchanged. When we took photographs and sent the film off to be printed and sent pictures to our friends, the measured economy expanded. Today we can use our smartphones, download the pictures and share them as widely as we wish, all without troubling the economic statistics. Likewise when we view a news site rather than buying a paper, use Spotify to add to our playlists, view an old programme on iPlayer rather than buy a DVD or make a call via Skype. We are able to send many more written messages, take vastly more pictures, digest more news stories, have bigger music libraries, watch more movies and speak to our friends more often, because the marginal cost of each activity is now zero. Our actual activity is up, but our measured activity is down. In these areas, technology is enhancing our lives but contracting our economies.

There is a second layer to this argument. Some measured activity is still taking place. We still have to buy our computers, tablets and smartphones and pay subscriptions to, say, Vodafone, Sky and Spotify. These transactions do add to the economy. But even here, the statistics do not fully capture what is going on. The typical computer costs a little less than it did 10 or 15 years ago, but it is far more powerful: faster and with a bigger memory. Unless you are unlucky in where you live, your broadband speed at home allows you to download weblinks, games and movies in far less time. Television sets cost much the same as they did, but they are thinner and have bigger screens than before, and give far sharper pictures. Cars are smarter, more reliable and more fuel-efficient (especially so if they are one of the new hybrids or electric cars). And some bits of kit are no longer needed at all. Free apps on our phones have massively reduced the demand for maps and satnavs. iPod sales have collapsed.

How do these advances show up in our growth and productivity statistics? The short answer is, on the whole they do not. If 100 people used to produce 1,000 television sets a week, and now make 1,200 sets of the same quality, then the calculation is simple. Productivity and output both rise by 20 per cent. But suppose the technical advances mean that the same 100 workers continue to make 1,000 television sets a week, but they are better than they were, how do we measure that? If we buy a television with a 50 per cent bigger screen than our last one, should that count as a 50 per cent bigger purchase?

The problem here is technical, and also philosophical. What is the fundamental ‘thing’ we are trying to measure? Is it the number of square inches of screen? Or the number of pixels we can view (a high definition picture has roughly four times as many as a traditional, standard definition picture)? Or, more subjectively, the increased pleasure (in economists’ parlance, ‘utility’) we derive from viewing a larger, sharper picture?

Whatever your answer, the statisticians almost certainly think differently. They know they have a problem, but they are struggling to solve it. Statisticians in the UK and other countries have been trying to develop ‘hedonic price indices’ that capture changes in product quality, but with very limited success. It is a complex and expensive task. For these reasons, a recent Organisation for Economic Co-operation and Development report noted that ‘the systematic uptake of hedonic methods is national price statistics has so far been confined to a small number of countries and products’.

As it happens, the UK’s Office of National Statistics is one of the world’s most advanced statistical agencies in addressing the impact of technology on the economy. It is a public institution that we should cherish. Yet still it struggles. Its challenges were laid bare last year by Sir Charles Bean, the former deputy governor of the Bank of England. In a report, commissioned by the government, into the quality of UK economic statistics, he questioned whether the current framework ‘is flexible enough to capture the full extent of the transformation brought about by the digital revolution … The great challenge for economic measurement stems from the fact that the consumption of digital products often does not involve a monetary transaction’. Therefore, he concluded, ‘official statistics may be missing an important aspect of the contemporary economy’.

I had reason to observe this first hand when I was president of YouGov.he company had to complete regular forms for the ONS about our turnover, staff levels and so on. These fed into the ONS’s statistics on the state of the one thing we were not asked to do was say how many surveys we had conducted, or how many questions we had put to how many respondents. In other words, we were not asked to provide any figures about the volume of work we did, as distinct from the money we earned.

Now, the point about YouGov is that it employs modern technology. As an online research company, we could generally do the same amount of research more cheaply than our traditional rivals, who asked questions over the phone or face-to-face. Suppose we won business by charging £80,000 for a piece of research for which another company had previously charged £100,000. You might think this would show up as the same ‘volume’ contribution to Britain’s GDP, with a lower price and higher productivity. You would be wrong. In fact, because the ONS measured revenue not volume, it would show up as a reduction in GDP, with no change in price or productivity.

This is a small example of a much wider point about the way we measure services. By definition, services are intangible. You can count things (even if we do not really know how to adjust the way we count things that evolve), but counting services, that is non-things, is altogether harder. Comparing the amount of milk sold by Tesco and Sainsbury is easy; comparing the number of haircuts sold by an upmarket stylist with a downmarket barber is a different matter. If the former charges 10 times the latter, is it 10 times as good – and so, worth 10 times as much to the national economy?

For this reason, statisticians across the world generally equate the volume of services sold with the turnover of each company. To some extent price changes can be caught when, say, a given hairdresser raises its prices. Until fairly recently this approach did no great violence to the larger economic picture. Technological changes to service provision tended to be fairly slow. But today, when change is more rapid, then the published statistics are liable to lead us astray. More generally, the implied equation between turnover and volume means that technological improvements, such as those offered by YouGov, show up as a decline in the economy, not an increase in productivity. And the faster the change, the greater the reported contraction in GDP for any given piece of work.

To recap, there three distinct ways in which recent technical changes have caused problems for our economic statistics: we do some things free that we used to pay for (such as send emails rather than letters); we buy products which are greatly superior to their predecessors of yesteryear (such as television sets); and new technology is shaking up some service industries (such as opinion polls).

However, we must go one stage further to see why these statistical problems end up making a mess of the public finances. A few months ago, before the OBR abandoned its expectation of an early recovery in the rate of productivity growth, one of Britain’s best-known economists challenged my logic. She said that whenever we save money – by not buying stamps or newspapers or CDs or iPods or satnavs – we have more money to spend on other goods and services. Indirectly, technical change will still cause jobs to be created, profits to be earned and government tax revenues to rise. The economic statistics will pick this up. Funding for public services will revive. There is no need to panic.

There is some truth in the more-money-for-other things argument; but how much? Here are four reasons why it will not restore the era of high, productivity-driven growth or buoyant tax revenues.

First, the benefits of switching any given product or service from paid to free boosts spending power on a one-off basis. When the last stamp or CD or TomTom has been bought, the boost has happened and cannot be repeated. It helps to raise the level of spending on other things, but does nothing to lift the rate of growth of spending in subsequent years. In economists’ parlance, there is a substitution effect, leading to a one-off income effect.

Second, there is no increase in spending power when technology leads to an improved rather than cheaper product. As we have seen this is broadly the story of television sets over the past few years. It also applies to cars and smartphones, among other products.

Third, even when technology does help consumers spend more on other things, the overall impact is not necessarily good for productivity (as conventionally measured) or tax revenue. New spending creates new jobs – but at the expense of old jobs. The past decade has been fine if you want to work in an Amazon warehouse or upmarket London hotel or restaurant, but not if your job used to be printing newspapers or staffing a high street bank branch or developing 35mm films. When both sides of the ledger are taken into account, the result may be a larger number of jobs in aggregate, but fewer full-time, decently-paid and secure jobs. Indeed, recent employment numbers, together with the controversies surrounding zero-hours jobs, suggests that this is precisely what is happening.

Fourth, past technical advances have been driven by revolutions in transport, energy and manufacturing processes. They led to, and transformed, the great factories that churned out the new products that changed our parents’ and grandparents’ lives in the 20th century. Today’s revolutions are driven by data. But whereas factories existed in places – usually near their markets – today’s giants, such as Apple, Microsoft, Google and Facebook are far freer to choose where they go. They can organise their affairs to operate where costs and taxes are lowest. Even companies that have to employ large numbers of people close to their customers, such as Amazon, are able to a large extent to choose where their profits are taxed. They are often not choosing Britain. The strained relationship between geographically-bound nations and geographically footloose multinational companies is one of the great business dramas of our times.

What, then, is to be done? For a laissez-faire liberal who wants small government and does not mind if the new economy leaves millions of workers with insecure, low-paid jobs, the answer is simple: not much. But if we want a fairer society with decent public services, the challenges are severe. Here are some provisional thoughts.

We need a different approach to company taxation. Different ideas have been mooted for extracting more money from those companies that have done exceptionally well in recent years and could reasonably be expected to contribute more to government revenues. One example is a ‘Tobin tax’ on financial transactions. Another is forcing multinational companies to come clean about where they really earn their profits and tax them accordingly. Put to one side the arguments about specific proposals. The larger point is that in today’s world, the only way to tackle data-driven businesses that cross borders at will is to ensure that politics does so, too. The UK is too small to do much on its own. It needs to develop a common taxation strategy with other countries. Here is an idea for starters. We should join forces with other European countries and pool our sovereignty to make sure we receive our fair share of corporate taxes. We could call this brave new venture the European Union.

We need to tax the benefits of ‘free’ technology. Productivity gains through the twentieth century generally boosted wages and profits – and hence tax revenues. Put another way: it could be measured financially, therefore it could be taxed. When gains in utility arise from the proliferation of free services, their absence from conventional economic data leads to their absence from the range of taxable activities. That list of activities needs to be expanded, so it is more closely aligned to what former Treasury official Julian Laite calls ‘gross domestic utility’, or GDU, which is buoyant, not just gross domestic product, which is not.

This means taxing the flow of data. The rate need not be high. Suppose it were designed to cost the average smartphone, tablet and computer user a total of £1 a day. People who used their phones and computers mainly for calls, emails, social media and the most commonly-used games and apps would have to pay just a few pence; only those who downloaded the most complex apps, games, videos et cetera would pay more. The government could expect to gather up to £20bn a year in new tax revenues. It could even be designed to be a progressive tax, with the levy applying only to data consumption above a certain level, so that lower-income, lower-use smartphone and computer owners pay nothing. Again, such a tax probably needs to be co-ordinated internationally to be most effective.

We need to make sure that the benefits of new technology are shared more widely. In theory free services are more democratic than paid for services. People on average incomes no longer need to worry about the cost of a newspaper or long-distance phone call or encyclopaedia for their children – as long as they can afford a smartphone and a monthly subscription. However, those on low incomes, or rely on state benefits, or are older and tech-averse, are often excluded from these benefits. Likewise with online shopping and banking – fine if you have a credit card and a basic understanding of handling personal finances, but treacherous if you do not. We need an active strategy for giving people access to the online world at every public institution – schools, post offices, doctors surgeries, council offices et cetera – as well as making financial and computer literacy a central element of every secondary school. We now have a free telephone helpline for people seeking to apply for Universal Credit; why not a free helpline for people needing human help mastering the wonders of the online world?

Finally, our public services need to accelerate the pace at which they bring modern technology into their day-to-day operations. Much, of course, has already been done. We can access many services online. Good schools have plenty of computers; hospitals and doctors surgeries have moved from appointments and patient data being written down to being recorded on computer systems. But the essential organisation of many services – a big and obvious example is school classrooms – have changed far less than many private-sector companies seeking to maintain their profits in the face of technical change. Progressive parties and their trade union allies understandably seek to protect public sector workers from disruptive change. But in the long run, it will be better for public sector jobs, as well as citizens who rely on the services they provide, to embrace change in an orderly way – and by ‘orderly’ I mean as fast as sensibly possible, not as slow as the defenders of the status quo can get away with.

Those ideas will not cure all our present ills. Nor does this analysis cover every cause of our stalled productivity figures. Investment levels are hopelessly inadequate, and the era of low interest rates has made it easy for lazy companies to use cheap money to avoid hard decisions. But the growing gulf between GDU and GDP, and the contrast between the growing benefits of modern technology, especially to the better-off, and the growing difficulties in measuring and taxing them, present a challenge that progressive politics, in particular, needs to address. 


Peter Kellner is former president of YouGov


9. Productivity: nothing to fear?

Is productivity really a problem for thew British economy – or is it simply the way we look at it that is the issue? Peter Kellner thinks so, and Richard Angell interviews him on his new essay for Progress magazine, while Alison McGovernStephanie Lloyd and Conor Pope discuss podcasts for popstars, feminists, football fans and wrestlers.



Progressive centre-ground Labour politics does not come for free.

It takes time, commitment and money to build a fight against the forces of conservatism. If you value the work Progress does, please support us by becoming a member, subscriber or donating.

Our work depends on you.

Print Friendly, PDF & Email

Peter Kellner

is president of YouGov

Add comment

Sign up to our daily roundup email