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.
Donald Trump has suggested that the Fed should cut interest rates by 1 percentage point and engage in a further round of quantitative easing. He wants to see monetary policy used to give a substantial boost to US economic growth at a time when inflation is below target. In a pair of tweets just before the meeting of the Fed to decide on interest rates, he said:
China is adding great stimulus to its economy while at the same time keeping interest rates low. Our Federal Reserve has incessantly lifted interest rates, even though inflation is very low, and instituted a very big dose of quantitative tightening. We have the potential to go up like a rocket if we did some lowering of rates, like one point, and some quantitative easing. Yes, we are doing very well at 3.2% GDP, but with our wonderfully low inflation, we could be setting major records &, at the same time, make our National Debt start to look small!
But would this be an appropriate policy? The first issue concerns the independence of the Fed.
It is supposed to take decisions removed from the political arena. This means sticking to its inflation target of 2 per cent over the medium term – the target it has officially had since January 2012. To do this, it adjusts the federal funds interest rate and the magnitude of any bond buying programme (quantitative easing) or bond selling programme (quantitative tightening).
The Fed is supposed to assess the evidence concerning the pressures on inflation (e.g. changes in aggregate demand) and what inflation is likely to be over the medium term in the absence of any changes in monetary policy. If the Federal Open Market Committee (FOMC) expects inflation to exceed 2 per cent over the medium term, it will probably raise the federal funds rate; if it expects inflation to be below the target it will probably lower the federal funds rate.
In the case of the economy being in recession, and thus probably considerably undershooting the target, it may also engage in quantitative easing (QE). If the economy is growing strongly, it may sell some of its portfolio of bonds and thus engage in quantitative tightening (QT).
Since December 2015 the Fed has been raising interest rates by 0.25 percentage points at a time in a series of steps, so that the federal funds rate stands at between 2.25% and 2.5% (see chart). And since October 2017, it has also been engaged in quantitative tightening. In recent months it has been selling up to $50 billion of assets per month from its holdings of around $4000 billion and so far has reduced them by around £500 billion. It has, however, announced that the programme of QT will end in the second half of 2019.
This does raise the question of whether the FOMC is succumbing to political pressure to cease QT and put interest rate rises on hold. If so, it is going against its remit to base its policy purely on evidence. The Fed, however, maintains that its caution reflects uncertainty about the global economy.
The second issue is whether Trump’s proposed policy is a wise one.
Caution about further rises in interest rates and further QT is very different from the strongly expansionary monetary policy that President Trump proposes. The economy is already growing at 3.2%, which is above the rate of growth in potential output, of around 1.8% to 2.0%. The output gap (the percentage amount that actual GDP exceeds potential GDP) is positive. The IMF forecasts that the gap will be 1.4% in 2019 and 1.3% in 2020 and 2021. This means that the economy is operating at above normal capacity working and this will eventually start to drive up inflation. Any further stimulus will exacerbate the problem of excess demand. And a large stimulus, as proposed by Donald Trump, will cause serious overheating in the medium term, even if it does stimulate growth in the short term.
For these reasons, the Fed resisted calls for a large cut in interest rates and a return to quantitative easing. Instead it chose to keep interest rates on hold at its meeting on 1 May 2019.
But if the Fed had done as Donald Trump would have liked, the economy would probably be growing very strongly at the time of the next US election in November next year. It would be a good example of the start of a political business cycle – something that is rarer nowadays with the independence of central banks.
- What are the arguments for central bank independence?
- Are there any arguments against central bank independence?
- Explain what is meant by an ‘output gap’? Why is it important to be clear on what is meant by ‘potential output’?
- Would there be any supply-side effects of a strong monetary stimulus to the US economy at the current time? If so, what are they?
- Explain what is meant by the ‘political business’ cycle? Are governments in the UK, USA or the eurozone using macroeconomic policy to take advantage of the electoral cycle?
- The Fed seems to be ending its programme of quantitative tightening (QT). Why might that be so and is it a good idea?
- If inflation is caused by cost-push pressures, should central banks stick rigidly to inflation targets? Explain.
- How are expectations likely to affect the success of a monetary stimulus?
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?
Consumer credit is borrowing by individuals to finance current expenditure on goods and services. Consumer credit is distinct from lending secured on dwellings (referred to more simply as ‘secured lending’). Consumer credit comprises lending on credit cards, lending through overdraft facilities and other loans and advances, for example those financing the purchase of cars. We consider here recent trends in the flows of consumer credit in the UK and discuss their implications.
Analysing consumer credit data is important because the growth of consumer credit has implications for the financial wellbeing or financial health of individuals and, of course, for financial institutions. As we shall see shortly, the data on consumer credit is consistent with the existence of credit cycles. Cycles in consumer credit have the potential to be not only financially harmful but economically destabilising. After all, consumer credit is lending to finance spending and therefore the amount of lending can have significant effects on aggregate demand and economic activity.
Data on consumer credit are available monthly and so provide an early indication of movements in economic activity. Furthermore, because lending flows are likely to be sensitive to changes in the confidence of both borrowers and lenders, changes in the growth of consumer credit can indicate turning points in the economy and, hence, in the macroeconomic environment.
Chart 1 shows the annual flows of net consumer credit since 2000 – the figures are in £ billions. Net flows are gross flows less repayments. (Click here to download a PowerPoint copy of the chart.) In January 2005 the annual flow of net consumer credit peaked at £23 billion, the equivalent of just over 2.5 per cent of annual disposable income. This helped to fuel spending and by the final quarter of the year, the economy’s annual growth rate had reached 4.8 per cent, significantly about its long-run average of 2.5 per cent.
By 2009 net consumer credit flows had become negative. This meant that repayments were greater than additional flows of credit. It was not until 2012 that the annual flow of net consumer credit was again positive. Yet by November 2016, the annual flow of net consumer credit had rebounded to over £19 billion, the equivalent of just shy of 1.5 per cent of annual disposable income. This was the largest annual flow of consumer credit since September 2005.
Although the strength of consumer credit in 2016 was providing the economy with a timely boost to growth in the immediate aftermath of the referendum on the UK’s membership of the EU, it nonetheless raised concerns about its sustainability. Specifically, given the short amount of time that had elapsed since the financial crisis and the extreme levels of financial distress that had been experienced by many sectors of the economy, how susceptible would people and organisations be to a future economic slowdown and/or rise in interest rates?
The extent to which the economy experiences consumer credit cycles can be seen even more readily by looking at the 12-month growth rate in the net consumer credit. In essence, this mirrors the growth rate in the stock of consumer credit. Chart 2 evidences the double-digit growth rates in net consumer credit lending experienced during the first half of the 2000s. Growth rates then eased but, as the financial crisis unfolded, they plunged sharply. (Click here to download a PowerPoint copy of the chart.)
Yet, as Chart 2 shows, consumer credit growth began to recover quickly from 2013 so that by 2016 the annual growth rate of net consumer credit was again in double figures. In November 2016 the 12-month growth rate of net consumer credit peaked at 10.9 per cent. Thereafter, the growth rate has continually eased. In January 2019 the annual growth rate of net consumer credit had fallen back to 6.5 per cent, the lowest rate since October 2014.
The easing of consumer credit is likely to have been influenced, in part, by the resumption in the growth of real earnings from 2018 (see Getting real with pay). Yet, it is hard to look past the economic uncertainties around Brexit.
Uncertainty tends to cause people to be more cautious. With the heightened uncertainty that has has characterised recent times, it is likely that for many people and businesses prudence has dominated impatience. Therefore, in summary, it appears that prudence is helping to steer borrowing along a downswing in the credit cycle. As it does, it helps to put a further brake on spending and economic growth.
- What is the difference between gross and net lending?
- Consider the argument that we should be worried more by excessive growth in consumer credit than on lending secured on dwellings?
- How could we measure whether different sectors of the economy had become financially distressed?
- What might explain why an economy experiences credit cycles?
- Explain how the growth in net consumer credit can affect economic activity?
- If people are consumption smoothers, how can credit cycles arise?
- What are the potential policy implications of credit cycles?
- It is said that when making financial decisions people face an inter-temporal choice. Explain what you understand this by this concept.
- If economic uncertainty is perceived to have increased how could this affect the consumption, saving and borrowing decisions of people?
Growth in the eurozone has slowed. The European Central Bank (ECB) now expects it to be 1.1% this year; in December, it had forecast a rate of 1.7% for 2019. Mario Draghi, president of the ECB, in his press conference, said that ‘the weakening in economic data points to a sizeable moderation in the pace of the economic expansion that will extend into the current year’. Faced with a slowing eurozone economy, the ECB has announced further measures to stimulate economic growth.
First it has indicated that interest rates will not rise until next year at the earliest ‘and in any case for as long as necessary to ensure the continued sustained convergence of inflation to levels that are below, but close to, 2% over the medium term’. The ECB currently expects HIPC inflation to be 1.2% in 2019. It was expected to raise interest rates later this year – probably by the end of the summer. The ECB’s main refinancing interest rate, at which it provides liquidity to banks, has been zero since March 2016, and so there was no scope for lowering it.
Second, although quantitative easing (the asset purchase programme) is coming to an end, there will be no ‘quantitative tightening’. Instead, the ECB will purchase additional assets to replace any assets that mature, thereby leaving the stock of assets held the same. This would continue ‘for an extended period of time past the date when we start raising the key ECB interest rates, and in any case for as long as necessary to maintain favourable liquidity conditions and an ample degree of monetary accommodation’.
Third, the ECB is launching a new series of ‘quarterly targeted longer-term refinancing operations (TLTRO-III), starting in September 2019 and ending in March 2021, each with a maturity of two years’. These are low-interest loans to banks in the eurozone for use for specific lending to businesses and households (other than for mortgages) at below-market rates. Banks will be able to borrow up to 30% of their eligible assets (yet to be fully defined). These, as their acronym suggests, are the third round of such loans. The second round was relatively successful. As the Barron’s article linked below states:
Banks boosted their long-term borrowing from the ECB by 70% over the course of the program, although they did not manage to increase their holdings of business loans until after TLTRO II had finished disbursing funds in March 2017.
Whether these measures will be enough to raise growth rates in the eurozone depends on a range of external factors affecting aggregate demand. Draghi identified three factors which could have a negative effect.
- Brexit. The forecasts assume an orderly Brexit in accordance with the withdrawal deal agreed between the European Commission and the UK government. With the House of Commons having rejected this deal twice, even though it has agreed that there should not be a ‘no-deal Brexit’, this might happen as it is the legal default position. This could have a negative effect on the eurozone economy (as well as a significant one on the UK economy). Even an extension of Article 50 could create uncertainty, which would also have a negative effect
- Trade wars. If President Trump persists with his protectionist policy, this will have a negative effect on growth in the eurozone and elsewhere.
- China. Chinese growth has slowed and this dampens global growth. What is more, China is a major trading partner of the eurozone countries and hence slowing Chinese growth impacts on the eurozone through the international trade multiplier. The ECB has taken this into account, but if Chinese growth slows more than anticipated, this will further push down eurozone growth.
Then there are internal uncertainties in the eurozone, such as the political and economic uncertainty in Italy, which in December 2018 entered a recession (2 quarters of negative economic growth). Its budget deficit is rising and this is creating conflict with the European Commission. Also, there are likely to be growing tensions within Italy as the government raises taxes.
Faced with these and other uncertainties, the measures announced by Mario Draghi may turn out not to be enough. Perhaps in a few months’ there may have to be a further round of quantitative easing.
- ECB statement following policy meeting
Reuters, Larry King (7/3/19)
- European Central Bank acts to boost struggling eurozone
BBC News, Andrew Walker (7/3/19)
- The European Central Bank Tries to Avoid Repeating Past Mistakes
Barron’s, Matthew C. Klein (8/3/19)
- ECB pushes back rate hike plans, announces fresh funding for banks
CNBC, Silvia Amaro (7/3/19)
- Why the ECB Followed the Fed’s Flip-Flopping
Bloomberg, Mohamed A. El-Erian (7/3/19)
- Central Banks Don’t Have the Answer and Markets Know It
Bloomberg, Robert Burgess (7/3/19)
- Missing out on monetary normalisation
OMFIF, David Marsh (12/4/19)
- The ECB is attempting to get ahead of event
Financial Times, The editorial board (8/3/19)
- Explainer: What is the fuss about European Central Bank TLTRO loans?
Reuters, Balazs Koranyi (4/3/19)
- Investigate the history of quantitative easing and its use by the Fed, the Bank of England and the ECB. What is the current position of the three central banks on ‘quantitative tightening’, whereby central banks sell some of the stock of assets they have purchased during the process of quantitative easing or not replace them when they mature?
- What are TLTROs and what use of them has been made by the ECB? Do they involve the creation of new money?
- What will determine the success of the proposed TLTRO III scheme?
- If the remit of central banks is to keep inflation on target, which in the ECB’s case means below 2% HIPC inflation but close to it over the medium term, why do people talk about central banks using monetary policy to revive a flagging economy?
- What is ‘forward guidance’ by central banks and what determines its affect on aggregate demand?