A general election has been called in the UK for 12 December. Central to the debates between the parties will be their policy on Brexit.
They range from the Liberal Democrats’, Plaid Cymru’s and Sinn Féin’s policy of cancelling Brexit and remaining in the EU, to the Scottish Nationalists’ and Greens’ policy of halting Brexit while a People’s Vote (another referendum) is held, with the parties campaigning to stay in the EU, to the Conservative Party’s policy of supporting the Withdrawal Agreement and Political Declaration negotiated between the Boris Johnson government and the EU, to the DUP which supports Brexit but not a version which creates a border between Great Britain and Northern Ireland, to the Brexit Party and UKIP which support leaving the EU with no deal (what they call a ‘clean break’) and then negotiating individual trade deals on a country-by-country basis.
The Labour Party also supports a People’s Vote, but only after renegotiating the Withdrawal Agreement and Political Declaration, so that if Brexit took place, the UK would have a close relationship with the single market and remain in a customs union. Also, various laws and regulations on environmental protection and workers’ rights would be retained. The referendum would take place within six months of the election and would be a choice between this new deal and remain.
But what are the economic costs and benefits of these various alternatives? Prior to the June 2016 referendum, the Treasury costed various scenarios. After 15 years, a deal would make UK GDP between 3.4% and 7.8% lower than if it remained in the EU, depending on the nature of the deal. No deal would make GDP between 5.4% and 9.5% lower.
Then in November 2018, the Treasury published analysis of the original deal negotiated by Theresa May in July 2018 (the ‘Chequers deal’). It estimated that GDP would be up to 3.9% lower after 15 years than it would have been if the UK had remained in the EU. In the case of a no-deal Brexit, GDP would be up to 9.3% lower after 15 years.
When asked for Treasury forecasts of the effects of Boris Johnson’s deal, the Chancellor, Sajid Javid, said that the Treasury had not been asked to provide forecasts as the deal was “self-evidently in our economic interest“.
Other forecasters, however, have analysed the effects of the Johnson deal. The National Institute for Economic and Social Research (NIESR), the UK’s longest established independent economic research institute, has estimated the costs of various scenarios, including the Johnson deal, the May deal, a no-deal scenario and also a scenario of continuing uncertainty with no agreement over Brexit. The NIESR estimates that, under the Johnson deal, with a successful free-trade agreement with the EU, in 10 years’ time UK GDP will be 3.5% lower than it would be by remaining in the EU. This represents a cost of £70 billion. The costs would arise from less trade with the EU, lower inward investment, slower growth in productivity and labour shortages from lower migration. These would be offset somewhat by savings on budget contributions to the EU.
Under Theresa May’s deal UK GDP would be 3.0% lower (and thus slightly less costly than Boris Johnson’s deal). Continuing in the current situation with chronic uncertainty about whether the UK would leave or remain would leave the UK 2% worse off after 10 years. In other words, uncertainty would be less damaging than leaving. The costs from the various scenarios would be in addition to the costs that have already occurred – the NIESR estimates that GDP is already 2.5% smaller than it would have been as a result of the 2016 Brexit vote.
Another report also costs the various scenarios. In ‘The economic impact of Boris Johnson’s Brexit proposals’, Professors Anand Menon and Jonathan Portes and a team at The UK in a Changing Europe estimate the effects of a decline in trade, migration and productivity from the various scenarios – again, 10 years after new trading arrangements are in place. According to their analysis, UK GDP would be 4.9%, 6.4% and 8.1% lower with the May deal, the Johnson deal and no deal respectively than it would have been from remaining in the EU.
But how much reliance should we put on such forecasts? How realistic are their assumptions? What other factors could they have taken into account? Look at the two reports and at the articles discussing them and then consider the questions below which are concerned with the nature of economic forecasting.
- UK’s new Brexit deal worse than continued uncertainty – NIESR
Reuters, David Milliken (30/10/19)
- Brexit deal means ‘£70bn hit to UK by 2029′
BBC News, Faisal Islam (30/10/19)
- Boris Johnson’s Brexit deal worse for economy than Theresa May’s, new analysis shows
Politics Home, Matt Honeycombe-Foster (30/10/19)
- Boris Johnson’s Brexit deal ‘would cost UK economy £70bn’
The Guardian, Richard Partington (30/10/19)
- UK economy suffers ‘slow puncture’ as general election is called
ITV News, Joel Hills (30/10/19)
- Boris Johnson’s Brexit deal ‘would deliver £70bn hit to economy by 2029’
Sky News, Ed Conway (30/10/19)
- Boris Johnson’s Brexit deal won’t cost Britain £70bn by 2029
The Spectator, Ross Clark (30/10/19)
- Boris Johnson’s Brexit deal would make people worse off than Theresa May’s
The Guardian, Anand Menon and Jonathan Portes (13/10/19)
- How Boris Johnson’s hard Brexit would hit the UK economy
Financial Times, Chris Giles (13/10/19)
- Boris Johnson’s Brexit deal is worse for the UK economy than Theresa May’s, research suggests
CNBC, Elliot Smith (19/10/19)
- What are the arguments in favour of the assumptions and analysis of the two recent reports considered in this blog?
- What are the arguments against the assumptions and analysis of the two reports?
- How useful are forecasts like these, given the inevitable uncertainty surrounding (a) the outcome of negotiations post Brexit and (b) the strength of the global economy?
- If it could be demonstrated beyond doubt to everyone that each of the Brexit scenarios meant that UK GDP would be lower than if it remained in the EU, would this prove that the UK should remain in the EU? Explain.
- If economic forecasts turn out to be inaccurate, does this mean that economists should abandon forecasting?
It’s been a while since I last blogged about labour markets and, in particular, about the effect of automation on wages and employment. My most recent post on this topic was on the 14th of April 2018 and it was mostly a reflection on some interesting findings that had been reported by Acemoglu et al (2017). More specifically, Acemoglu and Restrepo (2017) developed a theoretical framework to evaluate the effect of AI on employment and wages. They concluded that the effect was negative and potentially sizeable (for a more detailed discussion see my blog).
Using a model in which robots compete against human labor in the production of different tasks, we show that robots may reduce employment and wages … According to our estimates, one more robot per thousand workers reduces the employment to population ratio by about 0.18–0.34 percentage points and wages by 0.25–0.5 percent.
Since then, I have seen a constant stream of news on my news feed about the development of ever more advanced industrial robots and artificial intelligence. And this was not because of some spooky coincidence (or worse). It has been merely a reflection of the speed at which technology has been progressing in this field.
There are now robots that can run, jump, hold conversations with humans, do gymnastics (and even sweat for it!) and more. It is really impressive how fast change has been happening recently in this field – and, unsurprisingly, it has stimulated the interest of labour economists!
A paper that has recently come to my attention on this subject is by Graetz and Michaels (2018). The authors put together a panel dataset on robot adoption within seventeen countries from 1993 to 2007 and use advanced econometric techniques to evaluate the effect of these technologies on employment and productivity growth. Their analysis focuses exclusively on developed economies (due to data limitations, as they explain) – but their results are nevertheless intriguing:
We study here for the first time the relationship between industrial robots and economic outcomes across much of the developed world. Using a panel of industries in seventeen countries from 1993 to 2007, we find that increased use of industrial robots is associated with increases in labor productivity. We find that the contribution of increased use of robots to productivity growth is substantial and calculate using conservative estimates that it comes to 0.36 percentage points, accounting for 15% of the aggregate economy-wide productivity growth.
The pattern that we document is robust to including various controls for country trends and changes in the composition of labor and other capital inputs. We also find that robot densification is associated with increases in both total factor productivity and wages, and reductions in output prices. We find no significant relationship between the increased use of industrial robots and overall employment, although we find that robots may be reducing the employment of low-skilled workers.
This is very positive news for most – except, of course, for low-skilled workers. Indeed, like Acemoglu and Restrepo (2017) and many others, this study shows that the effect of automation on employment and labour market outcomes is unlikely to be uniform across all types of workers. Low-skilled workers are found again to be likely to lose out and be significantly displaced by these technologies.
And if you are wondering which sectors are likely to be disrupted most/first by automation, the rankings developed by McKinsey and Company (see chart below) would give you an idea of where the disruption is likely to start. Unsurprisingly, the sectors that seem to be the most vulnerable, are the ones that use the highest share of low-skilled labour.
- “The effect of automation on wages and employment is likely to be positive overall”. Discuss.
- Using examples and anecdotal evidence, do you agree with these findings?
- Using Google Scholar, put together a list of 5 recent (i.e. 2015 or later) articles and working papers on labour markets and automation. Compare and discuss their findings.
Latest resesarch from the independent American think tank The Conference Board paints a worrying picture about the growth of UK labour productivity. While global growth in labour productivity has weakened following the financial crisis, its weakness in the UK is singled out in the Board’s 2019 Productivity Brief. It finds that amongst large mature economies the decline in labour productivity growth rates has been greatest in the UK. This has important implications for the country’s longer-term well-being and, specifically, it peoples’ living standards.
The UK saw the growth in real GDP (national output) fall from 1.8 per cent in 2017 to 1.4 per cent in 2018. The Conference Board predicts that this will fall further to 0.8 per cent in 2019. In the context of living standards, the growth in real GDP per capita is particularly important. An increase in the population will, other things being equal, lower living standards because more people will be sharing a given amount of real national income. The growth in real GDP per capita fell from 1.1 per cent in 2017 to 0.7 per cent in 2018 and is predicted to fall to just 0.1 per cent in 2019.
Chart 1 shows the annual rates of growth in real GDP and real GDP per capita from the 1950s. The average growth rates are 2.4 and 1.9 per cent respectively. The other series shown is the annual growth in real GDP per person employed. This is a measure of the growth in labour productivity. Its average annual growth rate is also 1.9 per cent. This illustrates the intrinsic long-run relationship between labour productivity growth and the growth rate of GDP per capita and hence in general living stanadards. (Click here to download a PowerPoint copy of the chart.)
In the short term, rates of growth in output per worker (labour productivity) and GDP per capita (general living standards) can be less similar. For example, when unemployment rates rise labour productivity rates may be little affected despite GDP per capita falling. Nonetheless, the important point here is the close long-run relationship between the growth in labour productivity and GDP per capita. This then raises an important question: what factors contribute to the growth in output and labour productivity?
An approach known as growth accounting helps to identify four key contributors to the growth of total output. The first is the quantity of labour, commonly measured in labour hours. The second is the quality of labour, also known as labour composition. Third is capital services which are physical inputs into production and include machinery, structures and IT capital. Capital services are affected by quantity and quality, but, unlike labour, it is practically more difficult to separate out these dimensions. Fourth, is Total Factor Productivity (TFP).
TFP it is essentially the residual contribution to output growth that cannot be explained by changes in the quantity and quality of the individual inputs. Hence, in principle, it is capturing changes in how effectively the labour and capital inputs are being employed and combined in production. The Conference Board’s Productivity Brief describes the growth in TFP as providing ‘a more accurate picture of the overall efficiency by which capital, labour and skills are combined in the production process’.
Chart 2 shows Conference Board estimates of the percentage point contribution of these four sources of growth since 1990. Over this period, output growth averaged 2 per cent per year. The contribution of capital services and, hence, what is known as capital accumulation is particularly significant at 1.5 percentage points per year. This has been significantly larger than the contribution of labour hours which averaged only 0.3 percentage points per year since 1990. This evidences the importance played by capital deepening for output growth in the UK. (Click here to download a PowerPoint copy of the chart.)
Capital deepening captures the growth in capital services relative to the growth in the labour input. It takes on even greater significance when we think about the growth in labour productivity since, after all, this is the growth in output relative to the quantity of labour. It is significant though that since 2015 the growth of capital services has contributed only 1 percentage point to output growth while the growth of labour hours has contributed an average of 0.7 percentage points. This points to a slowdown in capital deepening and hence in the growth of labour productivity.
Chart 2 also illustrates the importance of TFP growth to overall output growth. It is also important (along with capital deepening and the growth in labour quality) for the growth in labour productivity. Interestingly, we observe significant fluctuations in the growth of TFP. This is thought to reflect fluctuations in the utilisation of inputs. For example, if the utilisation of inputs falls (rises) when output falls (increases) this will be mirrored by a disproportionately large fall (increase) in TFP. In the longer-term, however, changes in TFP capture aspects of technological progress and advancement that enable more effective production methods and techniques to be deployed. In other words, the growth of TFP captures the ability of production to benefit from the advancement in ideas, products, processes and know-how.
A decline in the growth in TFP growth following the financial crisis is found quite widely in mature economies. The annual rate of growth of TFP across mature economies fell from 0.5 per cent year in 2000-2007 to 0.2 per cent in 2010-2017. In the UK this fall was from 0.5 per cent to -0.1 per cent. Hence, the decline in TFP growth of 0.6 percentage points between 2010 and 2017 was double the 0.3 percentage point fall across all mature economies. In 2018 the Conference Board estimate that TFP in the UK fell by 0.1 percent further exacerbating the downward pressure on labour productivity.
As our final chart shows, it is the magnitude to which labour productivity has eased following the financial crisis that sets the UK apart. While across all mature economies the growth of output per labour hour (another measure of labour productivity growth) fell from an average of 2.3 per cent per year in 2000-2007 to 1.2 per cent in 2010-2017, in the UK the fall was from 2.2 per cent to 0.5 per cent per year. (Click here to download a PowerPoint copy of the chart.)
While the productivity problem facing the UK is not new, the latest figures comes as a very timely reminder of the extent of the problem. To some extent the uncertainty around Brexit and the negative impact on capital accumulation has only helped to exacerbate the problem. But, this may mask a more systemic problem facing the UK. Getting to the root of this problem matters. It matters most significantly for our long-term wellbeing and prosperity. The productivity gap with our major industrial competitors is a gap that policymakers need not only to be mindful of but one that needs closing.
- What do you understand by the term labour productivity. How could we measure it?
- Why is it important to look at the growth of output per capita when assessing the benefits of long-term growth?
- Why is labour productivity important for the long-term well-being of a country?
- What do you understand by the method of growth accounting?
- What is the distinction between capital accumulation and capital deepening?
- What might explain why the growth of labour productivity has been lower in the years following the post-financial crisis?
- What do you understand by Total Factor Productivity (TFP)?
- What does the long-term growth of TFP attempt to capture?
- If you were an economic advisor to the government, what types of policy initiatives might you recommend for a government concerned about low rates of growth of labour productivity?
How would your life be without the internet? For many of you, this is a question that may be difficult to answer – as the internet has probably been an integral part of your life, probably since a very young age. We use internet infrastructure (broadband, 4G, 5G) to communicate, to shop, to educate ourselves, to keep in touch with each other, to buy and sell goods and services. We use it to seek and find new information, to learn how to cook, to download music, to watch movies. We also use the internet to make fast payments, transfer money between accounts, manage our ISA or our pension fund, set up direct debits and pay our credit-card bills.
I could spend hours writing about all the things that we do over the internet these days, and I would probably never manage to come up with a complete list. Just think about how many hours you spend online every day. Most likely, much of your waking time is spent using internet-based services one way or another (including apps on your phone, streaming on your phone, tablet or your smart TV and similar). If your access to the internet was disrupted, you would certainly feel the difference. What if you just couldn’t afford to have computer or internet access? What effect would that have on your education, your ability to find a job, and your income?
Martin Jenkins, a former homeless man, now entrepreneur, thinks that the magnitude of this effect is rather significant. In fact, he is so convinced about the importance of bringing the internet to poorer households, that he recently founded a company, Neptune, offering low-income households in the Bronx district of New York free access to online education, healthcare and finance portals. His venture was mentioned in a recent (and very interesting) BBC article – a link to which can be found at the end of this blog. But is internet connectivity really that important when it comes to economic and labour market outcomes? And is there a systematic link between economic growth and internet penetration rates?
These are all questions that have been the subject of intensive debate over the last few years, in the context of both developed and developing economies. Indeed, the ‘digital divide’ as it is known (the economic gap between the internet haves and have nots) is not something that concerns only developing countries. According to a recent policy brief published by the New York City Comptroller:
More than one-third (34 percent) of households in the Bronx lack broadband at home, compared to 30 percent in Brooklyn, 26 percent in Queens, 22 percent in Staten Island, and 21 percent in Manhattan.
The report goes on to present data on the percentage of households with internet connection at home by NYC district, and it does not take advanced econometric skills for one to notice that there is a clear link between median district income and broadband access. Wealthier districts (e.g. Manhattan Community District 1 & 2 – Battery Park City, Greenwich Village & Soho PUMA), tend to have a significantly higher share of households with broadband access, than less affluent ones (e.g. NYC-Brooklyn Community District 13 – Brighton Beach & Coney Island PUMA) – 88% of total households compared with 58%.
But, do these large variations in internet connectivity matter? The evidence is mixed. On the one hand, there are several studies that find a clear, strong link between internet penetration and economic growth. Czernich et al (2011), for instance, using data on OECD countries over the period 1996–2007, find that “a 10 percentage point increase in broadband penetration raised annual per capita growth by 0.9–1.5 percentage points”.
Another study by Koutroumpis (2018) examined the effect of rolling out broadband in the UK.
For the UK, the speed increase contributed 1.71% to GDP in total and 0.12% annually. Combining the effect of the adoption and speed changes increased UK GDP by 6.99% cumulatively and 0.49% annually on average”. (pp.10–11)
The evidence is less clear, however, when one tries to estimate the benefits between different types of workers – low and high skilled. In a recent paper, Atasoy (2013) finds that:
gaining access to broadband services in a county is associated with approximately a 1.8 percentage point increase in the employment rate, with larger effects in rural and isolated areas.
But then he adds:
most of the employment gains result from existing firms increasing the scale of their labor demand and from growth in the labor force. These results are consistent with a theoretical model in which broadband technology is complementary to skilled workers, with larger effects among college-educated workers and in industries and occupations that employ more college-educated workers.
Similarly, Forman et al (2009) analyse the effect of business use of advanced internet technology and local variation in US wage growth, over the period 1995–2000. Their findings show that:
Advanced internet technology is associated with larger wage growth in places that were already well off. These are places with highly educated and large urban populations, and concentration of IT-intensive industry. Overall, advanced internet explains over half of the difference in wage growth between these counties and all others.
How important then is internet access as a determinant of growth and economic activity and what role does it have in bridging economic disparities between communities? The answer to this question is most likely ‘very important’ – but less straightforward than one might have assumed.
- Comptroller, New York City, Internet Inequality
- Czernich, N., Falck, O., Kretschmer, T. and Woessmann, L., 2011, Broadband infrastructure and economic growth, The Economic Journal, 121(552), pp.505–32
- Koutroumpis, P., 2018, The economic impact of broadband: evidence from OECD countries, Ofcom
- Atasoy, H., 2013, The effects of broadband internet expansion on labor market outcomes, ILR Review, 66(2), pp.315–45
- Forman, C., Goldfarb, A. and Greenstein, S., 2009, The Internet and Local Wages: Convergence or Divergence? (No. w14750), National Bureau of Economic Research
- Is there a link between economic growth and internet access? Discuss, using examples.
- Explain the arguments for and against government intervention to subsidise internet access of poorer households.
- How important is the internet to you and your day to day life? Take a day offline (yes, really – a whole day). Then come back and write about it.
Workers in the UK and USA work much longer hours per year than those in France and Germany. This has partly to do with the number of days paid holiday per year, partly with the number of hours worked per day and partly with the number of days worked per week.
According to the latest OECD figures, in 2017 average hours worked per year ranged from 2257 in Mexico (the OECD’s highest) to 1780 in the USA, 1710 in Japan, 1681 in the UK, 1514 in France, 1408 in Denmark and 1356 in Germany (the OECD’s lowest). Annual working hours have been falling in most countries across the decades, as the chart shows. However, in most countries the process has slowed in recent years and in the UK, the USA and France working hours have begun to rise. (Click here for a PowerPoint of the chart.)
But why do working hours differ so much from country to country? How do they relate to productivity? How do they relate to human happiness and welfare more generally?
Causes of the differences
There are various reasons for the differences in hours worked between countries.
In a situation where individual workers can choose how many hours to work, they have to decide the best trade off for them between income and leisure. As wages rise over time, there will be substitution and income effects of these extra hourly wages. Higher wages make work more valuable in terms of what people can buy from an extra hour’s work. There is thus an incentive to substitute work for leisure and hence work longer. This is the substitution effect. On the other hand, higher wages allow people to work fewer hours for a given income. This is the income effect.
As incomes rise, generally the substitution effect will tend to decline relative to the income effect. This is because of the diminishing marginal utility of income. Richer people will tend to value a given rise in income less than poorer people and therefore will value the income from extra work less than poorer people. Richer people will prefer to work fewer hours than poorer people. Generally workers in richer OECD countries work fewer hours than those in poorer OECD countries.
But this does not explain why people in the USA, Canada, Japan and the UK work longer hours than people in Germany, Denmark, Norway, The Netherlands and France.
One possible explanation for these differences is the role of trade unions. These tend to be stronger in countries with lower working hours. Reducing the working week or obtaining longer holidays is one of the key objectives of unions.
Another is income distribution. The USA, despite its high average (mean) income, has a relatively unequal distribution of income compared with Germany or France. The post-tax-and-benefits Gini coefficient in the USA is around 0.39, whereas in Germany it is 0.29, meaning that Germany has a more equal distribution of disposable income than the USA. In fact, rises in real incomes in the USA over the past 10 years have gone almost exclusively to the top 10 per cent of earners, leaving the median income little changed. In fact median household income only rose above its 2007 (pre-recession) level in 2016.
Social and cultural explanations may also be important. People in countries with higher working hours relative to hourly wages may put a greater store on consumption relative to leisure. The desire to shop may be very strong. The ‘Anglo-Saxon’ economic model pursued by right-of-centre governments in English-speaking countries, such as the USA, Canada, Australia and the UK puts emphasis on low taxes, low regulation, low public expenditure and self-advancement. Such a model encourages a more individualistic approach to work, with more emphasis on earning money.
Then there is the attitude to hours worked generally. There is a saying that in the UK the last one to leave the office is seen as the hardest working, whereas in Germany the last one to leave is seen as the least efficient. Social pressures, from colleagues, family, friends and society more generally can have a major effect on people’s choices between work and leisure.
Productivity, in terms of output per hour worked, tends to decline as workers work longer hours. People get tired and possibly bored and demotivated towards the end of a long day or week. If workers are paid by the output they produce and if productivity declines towards the end of the day, then the hourly wage would fall as the day progresses. This would act as a disincentive to work long hours. In practice, most workers are normally paid a constant rate per hour for normal-time working. For overtime, they may even be paid a higher rate, despite their likely lower productivity. This encourages them to work longer hours than if they were paid according to their marginal productivity.
Linking pay more closely to productivity could encourage people to opt for fewer hours (if they had the choice). Indeed some companies are now encouraging workers to choose their hours – which may mean fewer hours as people seek a better work–life balance. (See the BBC article below about PwC’s employment strategy.) Alternatively, some other employers adopt the system of giving workers a set amount of work to do and then they can leave work when it is finished. This acts as an incentive to work more efficiently.
It is interesting that countries where workers work more hours per year tend to have a lower output per hour worked relative to output per worker than countries where workers work fewer hours. This is illustrated in the chart opposite. The USA, with its longer working hours, has higher output per person employed than France and Germany but very similar output per hour worked.
Hours and happiness
So are people who choose to work longer hours and take home more money likely to be happier than those who choose to work fewer hours and take home less money? If people were rational and had perfect knowledge, then they would choose the balance between work and leisure that best suited them.
In practice, labour markets are highly imperfect. People often do not have choices about the amount they work; they work the hours they are told. Even if they do have a choice, they are unlikely to have perfect knowledge about the impact of long hours on their health and happiness over their lifetime. They may not even be good judges of the shorter-term effects of more work and more pay. They may believe that more money will buy them more happiness only to find soon afterwards that they are wrong.
- What factors are likely to encourage workers to work longer hours?
- Give some examples of jobs where workers have flexibility in the amount of hours they work per week and jobs where the working week is of a fixed length.
- For what reasons are annual working hours longer in the USA than in Germany?
- Would it be in employers’ interests if the government legislated so as to reduce the maximum permitted working week? Explain.
- What is meant by ‘efficiency wages’? How relevant is the concept to the issue of the average number of hours worked per year from country to country?
- Explain why people in poorer countries tend to work more hours per year than people in richer countries.
- If workers’ wages equalled their marginal revenue product, why might some workers choose to work more and others choose to work less (assuming they had a choice)?
- Are jobs in the gig economy and zero-hour contract jobs in the interests of workers?
- Is South Korea wise to cut its work limit from 68 hours a week to 52?