Unemployment remains at historically low levels and the MPC judges that very little slack remains in the economy. Notwithstanding a projected rise in structural productivity growth, potential supply growth is expected to be subdued. As a result, the pace at which output can grow without generating inflationary pressures is likely to remain modest.
In the medium term, the pace at which output can grow without generating inflationary pressures — known as the potential growth rate of the economy — is determined by the economy's supply capacity. In turn, that depends on structural features of the economy such as population growth and growth in productivity. In the short term, however, there may be scope for output to grow more quickly than this if resources are underutilised such that there is 'slack' in the economy. Thus, in judging how fast the economy can grow without generating inflationary pressures, the MPC assesses both the degree of slack and the outlook for potential supply.
During the financial crisis, output fell, unemployment rose substantially and a significant degree of slack opened up. In the years since then, output growth has risen (Chart 3.1). But much of that recovery in output growth has been accounted for by a rise in total hours worked as slack in the labour market has been absorbed. The unemployment rate, for example, fell from 8.5% in 2011 to 4.3% in the three months to November (Chart 3.2), its lowest level since 1975. At the same time, the potential growth rate of the economy has remained subdued due to persistent weakness in productivity growth.
In the run-up to this Report, the MPC conducted its annual reassessment of supply-side conditions. Based on the range of evidence set out in this section, the MPC judges that very little slack remains overall (Section 3.1). Furthermore, the rate of potential growth is projected to be subdued at around 1½% (Section 3.2). As such, the pace of demand growth consistent with balanced domestic inflationary pressures is likely to be modest (Section 5).
The degree of slack in the economy — or the gap between demand and potential supply — is a key determinant of domestic inflationary pressures. Judging the degree of slack is difficult, however, since the level of potential supply cannot be directly observed.
The MPC's best collective judgement is that the overall margin of slack is currently very small, at just under ¼% of GDP. That is broadly consistent with the results from top-down statistical filters that estimate potential supply using past observations of GDP, inflation and unemployment, as well as with detailed evidence on the degree of slack within individual components of supply. Those individual components encompass spare resources within the labour market — reflected in unemployment and inactivity — and within firms. While the overall margin of slack is judged to be only slightly narrower than in November, its composition is now judged to be different, with greater slack remaining within average hours worked but less within companies' capital utilisation.
The degree of slack in the labour market reflects the balance between companies' labour demand and the amount of labour supplied by households. Growth in labour demand is likely to have remained robust in recent quarters, with most indicators of employment intentions above their historical averages and the number of vacancies relative to the size of the labour force continuing to increase (Table 3.A). As a result, employment growth has generally been solid. Although employment fell in Q3, these data tend to be volatile and employment growth rebounded in the three months to November.
Robust growth in employment over 2017 has resulted in a further tightening of the labour market. Survey measures of recruitment difficulties are above their past averages and most picked up further in Q4 (Table 3.B). The unemployment rate fell from 4.7% at the start of the year to 4.3% in the three months to November and is expected to remain at that level in coming months (Chart 3.2).
As discussed in Box 4, declines in unemployment beyond a certain point, known as the equilibrium rate, will put upward pressure on wages and inflation as jobs become increasingly difficult to fill at prevailing wage rates. The equilibrium rate is unobservable and hard to estimate with precision. Based on a range of evidence, the MPC judges that the long-term equilibrium unemployment rate is around 4¼%, a little lower than judged a year ago and broadly in line with the current headline rate of unemployment.
When judging the overall degree of slack in the labour market, it is important to consider broader measures than just the unemployment rate. Increases in the number of people in work, for example, can be associated not just with a fall in unemployment, but also with people entering employment who previously said they were not actively looking for work. The 'marginal attachment' ratio — the proportion of the population who report that they would like a job but are not currently seeking work — has fallen sharply in recent years (Chart 3.3). That suggests that the scope for the employment rate to increase further as such people enter the labour market is likely to be limited.
There may be scope for companies to expand output by utilising their existing capital or labour more intensively. According to the Labour Force Survey (LFS), for example, average hours worked in Q3 were a little below the hours that households said they would like to work. And, while the share of part-time workers who report that they would prefer a full-time job has fallen, it remains above its pre-crisis average (Chart 3.4). As such, there is likely to be some scope for companies to increase output by raising the number of hours their employees work. There is likely to be little scope for companies to use their existing capital more intensively, however. Survey indicators of companies' capacity pressures, for example, are currently around normal (Chart 3.5).
Since very little slack is judged to remain in the economy, the speed at which output can expand without generating inflationary pressures will largely depend on growth in potential supply. That growth can be driven either by increases in labour supply — the size of the labour force and the number of hours that people are willing to work — or by growth in productivity — the amount produced for each of those hours worked.
In recent decades, growth in the size of the UK workforce has tended to come mainly from population growth. Growth in the working-age population — those aged over 16 — has slowed in recent quarters, in part due to a slowing in net migration, which fell to 230,000 in the year to 2017 Q2 from 336,000 in the previous year.
In the MPC's latest forecasts, population growth is assumed to evolve in line with the ONS's principal projection published in October 2017. In that projection, growth in the working-age population remains slower than on average over the past decade (Chart 3.6). Within that, the ONS projects a further slight fall in annual net migration, to 211,000 by mid-2020. Box 5 describes the implications of lower net migration for overall UK labour supply.
The extent to which changes in the population affect the size of the workforce, however, depends on the proportion of the population who are active in the labour market versus, for instance, those in retirement or education. As explained in Box 5, that proportion — the participation rate — is expected to remain broadly flat in coming years (Chart 3.6).
Finally, overall labour supply will depend on the average number of hours those employed are willing to work. This is projected to fall very slightly in coming years (Chart 3.6), in part reflecting the increasing proportion of the population in older age groups, who tend to work fewer hours (see the box on pages 22–23 of the February 2016 Report).
Taking all these factors together, potential labour supply is projected to grow by 0.4% per year on average over the next three years, much lower than its average of 1.2% in the years following the crisis (Chart 3.6).
Given the relatively subdued outlook for labour supply growth and very little slack within the economy (Section 3.1), growth in the economy's overall supply capacity will rely in large part on trends in structural productivity growth.
Four-quarter hourly productivity growth stalled in the first half of 2017 and, while it picked up slightly in the second half of the year, it nevertheless remains subdued. One reason for the recent weakness may have been a shift in the composition of employment growth relative to past norms. Employment growth in the year to Q3 was concentrated in people and jobs with characteristics typically associated with lower-than-average wages (Section 4). To the extent that these characteristics are associated with lower levels of productivity, this shift in employment composition is likely also to have reduced aggregate productivity growth. These compositional effects will reduce productivity growth only for as long as such shifts continue, however.
More generally, productivity growth has been so weak since the financial crisis that the level of productivity is barely above its pre-crisis peak (Chart 3.7). A standard growth-accounting framework suggests that around half of the weakness since 2010 has been associated with slow growth in the amount of capital — the resources and equipment available to produce output — per hour worked (Chart 3.8). In turn, that reflects subdued business investment over much of that period (Section 2). The remainder of the weakness in productivity growth is accounted for by weak growth in the efficiency with which labour and capital are put to use, known as total factor productivity.
Sectoral data may provide further information about the drivers of the shortfall in productivity growth. These data suggest that over half of the shortfall relative to pre-crisis rates has been concentrated in the financial and insurance services and manufacturing sectors (Chart 3.9). To some degree, the slowdown in financial services is likely to reflect unusually high growth in measured productivity prior to the crisis, driven by increased leverage and risk-taking within financial firms over that period. In addition, mismeasurement of financial services output may have played a role in overemphasising the effects on productivity growth of higher financial sector leverage prior to the crisis and the deleveraging since then.(1) Although financial services productivity growth may pick up relative to the period following the financial crisis, the pace of growth seen in the 2000s is unlikely to return.
In the manufacturing sector, it is possible that the process of offshoring could have boosted measured productivity growth during the early 2000s.(2) In addition, productivity growth is likely to have been affected by trends in world trade flows. Growth in world trade tends to be associated with productivity gains through greater economies of scale and increased competition,(3) and so — since manufacturing firms tend to be highly integrated within global supply chains — their productivity growth is likely to have been affected by the weakness in trade growth since the crisis.
The shortfall in productivity growth can also be analysed at the firm level. Productivity growth varies widely between companies, and has tended to come from those companies at the frontier of the productivity distribution.(4) Although that is still the case, the slowing in productivity growth since the crisis also reflects slower growth of the top end of that distribution, while the lower end of the distribution has, on average, experienced stronger growth than in the past (Chart 3.10).
As well as these pre-existing trends, the outlook for productivity growth is likely to be affected by changes in trading arrangements as a result of Brexit. Any reduction or reorientation of trade and supply chains, for example, is likely to weigh on productivity growth for a period.(5) In addition, uncertainty around the eventual shape of the post-Brexit trading arrangements has been weighing on business investment (see Box 3), and consequently growth in the capital stock.
Overall, the MPC judges that productivity growth is likely to pick up in coming quarters, as the weakness in productivity in early 2017 unwinds and business investment increases (Section 2). Growth in structural productivity is expected to remain subdued, however, as the factors weighing on growth since the crisis persist, and the impact of Brexit continues. Furthermore, since there is judged to be little scope for companies to use their existing capital more intensively (Section 3.1), growth in output per hour is projected to be broadly in line with this structural rate, and hence a little slower than projected in November (see Box 6).
(1) Part of the output produced by financial institutions is known as FISIM (financial intermediation services indirectly measured). It is possible that the growth rate of FISIM — which relates closely to the stock of loans and deposits — can overstate changes in the services provided by these institutions to households and companies. For further details see Tenreyro, S (2018), 'The fall in productivity growth: causes and implications'. See also Burgess, S (2011), 'Measuring financial sector output and its contribution to UK GDP', Bank of England Quarterly Bulletin, 2011 Q3.
(2) For further details see Tenreyro, S (2018), ibid.
(3) See the box on page 29 of the August 2016 Inflation Report for more details. Many studies find that competition leads more productive firms to expand and less productive firms to exit markets, raising aggregate productivity, for instance
Bloom, N, Draca, M and Van Reenen, J (2016), 'Trade induced technical change?
The impact of Chinese imports on innovation, IT and productivity', The Review of Economic Studies, No. 83(1), pages 87–117.
(4) For more details, see Haldane, A (2017), 'Productivity puzzles'.
(5) For more details, see Carney, M (2017), '[De]Globalisation and inflation'.
The MPC expects growth in the economy's potential supply capacity to be subdued
Decomposition of estimated potential output growth (a)
Unemployment is projected to remain at its current low level in Q1
Unemployment rate and Bank staff's near-term projection (a)
Employment growth has remained robust over 2017
Changes in employment, vacancies, redundancies and survey indicators of employment intentions
Recruitment difficulties have continued to intensify
Survey indicators of recruitment difficulties
The proportion of people not currently looking for work, but who would like a job, has continued to fall
Marginal attachment ratio (a)
The proportion of part-time workers unable to find a full-time job remains slightly elevated
People working part-time who could not find a full-time job, as a proportion of total employment (a)
Survey measures suggest capacity pressures are around normal
Survey indicators of capacity pressures (a)
Monitoring the MPC's key judgements
Potential labour supply growth is expected to be subdued relative to the past decade
Contributions to estimated annual growth in potential hours worked (a)
Productivity has barely risen since the financial crisis
Whole-economy hourly labour productivity (a)
Productivity growth remains subdued
Contributions to four-quarter growth in whole-economy hourly labour productivity (a)
Finance and manufacturing account for over half of the recent weakness in productivity growth
Contributions to hourly labour productivity growth (a)
Much of the productivity growth shortfall has been concentrated at the top of the productivity distribution
Productivity per worker by productivity percentile (a)
The unemployment rate has fallen sharply in recent years, from 8.5% in late 2011 to 4.3% in the three months to November, its lowest level since 1975. When unemployment is low, that tends to put upward pressure on wage growth and inflation as companies need to pay more in order to recruit suitably skilled staff. And when unemployment is high, wage and inflationary pressures tend to be subdued as companies find it relatively easy to recruit and retain the right people. A key judgement for the MPC is where the 'equilibrium rate' of unemployment is — the rate consistent with meeting the inflation target in the medium term. This box explores the concept of the equilibrium unemployment rate in more detail and the evidence for where it currently lies.
The equilibrium unemployment rate can vary over time. Over the longer term, it represents the rate of unemployment that the economy is capable of achieving sustainably over many years. This long-term rate is determined by the structural features of the economy that affect the time it takes for people to find the right jobs, for example the extent to which potential workers are a good match for the jobs companies want to fill. It will also be influenced by the tax and benefit regime, which affects the incentives for people to move between employment and unemployment. Since these factors tend to be slow-moving, the long-term equilibrium rate is usually assumed to change only slowly over time.
In the shorter term, cyclical factors, such as changes in the mix of unemployment, can also affect the unemployment rate consistent with stable wage pressures. For example, people who have been out of work for over a year tend to be less likely to find employment than those who have been out of work for a shorter period of time, and so tend to exert less downward pressure on wages and inflation.
Following the financial crisis, that shorter-term equilibrium unemployment rate probably rose as the proportion of people out of work for over a year increased sharply (Chart A). But that effect has largely unwound, with the proportion unemployed for over 12 months back at its pre-crisis average. That suggests that the recession was not associated with any structural rise in long-term unemployment, in contrast to previous UK experience. It also suggests that the shorter-term equilibrium unemployment rate is likely to be close to its long-term structural rate.
The equilibrium unemployment rate is unobservable and so difficult to estimate with precision. One way of assessing its level, however, is to make use of the relationship between unemployment and wages. After accounting for factors other than unemployment that are likely to be influencing wage growth — for example growth in productivity — it is possible to infer what the rate of equilibrium unemployment would need to be in order to be consistent with current wage growth.
In February 2017, the MPC lowered its estimate of the long-term equilibrium rate from 5% to 4½%, following a period when wage growth had been below its projections over successive quarters. A lower equilibrium rate helped explain those forecast errors. Over the past year, annual pay growth has remained subdued (Section 4). Although that partly reflects continued weakness in productivity growth, including that stemming from recent shifts in the composition of employment growth, it would, all else equal, be consistent with a long-run equilibrium unemployment rate somewhat below 4½%.
A more formal way of using the relationship between unemployment and wages to estimate the equilibrium rate is to use statistical filtering techniques.1 These techniques impose an assumption about the relationship between unemployment and wage growth, for example that it is linear. The estimated equilibrium rate is then allowed to vary over time in order to capture persistent structural changes in the labour market. Chart B shows that the equilibrium unemployment rate estimated using one such statistical filter has fallen since 2010. Although there is considerable uncertainty around these statistical estimates, over the past year they suggest that the equilibrium rate has fallen very slightly further and remained close to the headline unemployment rate.
Another way to estimate the level of equilibrium unemployment is to examine the structural features of the economy that determine the time it takes people to find the right jobs. When the efficiency with which employees are matched to new job vacancies improves, or when the rate at which existing jobs are destroyed falls, then, for a given level of labour demand, the long-run equilibrium unemployment rate will fall.
One factor likely to have improved the efficiency with which employees are matched to job vacancies is a rise in the average educational attainment of the workforce. More highly skilled workers are likely to be better-suited on average to the jobs on offer. In addition, technological progress — for example the increasing use of online vacancy sites — is likely to have improved job matching by reducing the cost to companies of advertising vacancies and to workers of searching for new jobs. Despite these developments, however, the rate at which the unemployed move into employment — which depends in part on the efficiency of job matching — remains no higher than prior to the crisis (Chart C).
The rate at which jobs are destroyed is also likely to have been affected by structural changes. Increased flexibility within the labour market may have meant that employment contracts can be adjusted more easily, such that firms are able to reduce employee hours to lower levels for a period without making their workers redundant. This will tend to lower the equilibrium unemployment rate, although it may also lead to periods of 'underemployment' of those in work if hours are reduced below the level of those that employees wish to work.
Last year, the MPC judged that such structural changes had reduced the job destruction rate and hence the rate of equilibrium unemployment. Since then, the job destruction rate has fallen slightly further (Chart C), largely due to a fall in the number of people moving from temporary work into unemployment. Should these developments persist, they would suggest that the equilibrium unemployment rate is lower than previously thought.
Taking all the evidence together, the MPC judges that the long-term equilibrium rate of unemployment is slightly lower than judged a year ago, at around 4¼%. That is broadly in line with the headline rate of unemployment. Taken together with other evidence (Section 3.1), overall slack within the economy is likely to be very small at just under ¼% of GDP.
1 For more details, see Berry, S, Corder, M, Duffy, C, Hackworth, C and Speigner, B (2015), 'Trends in UK labour supply', Bank of England Quarterly Bulletin, 2015 Q4.
Long-term unemployment has fallen back to its past average rate
Unemployment rates by duration (a)
Estimated equilibrium unemployment from a filter model has fallen a little further over the past year
Unemployment and estimated shorter-term equilibrium rate from a statistical filtering mode
The job destruction rate has fallen slightly further over the past year
Flows between employment and unemployment
At a basic level, output in the economy can expand because either there are more people producing it or gains in productivity enable more output to be produced by the same workforce. Increases in the size of the workforce have accounted for nearly all of UK output growth over the past decade. Those increases have in turn resulted mainly from population growth, although changes in labour market participation rates have also been significant (Table 1).
In the MPC's projections, the size and composition of the population are assumed to evolve in line with the ONS's latest principal population projections, published in October 2017. Under those projections, growth in the workforce is subdued relative to the past decade, as the average age of the population continues to rise and the level of net migration falls. This box discusses the implications of these developments.
The proportion of the UK population aged over 65 has been rising steadily, from 20% of the 16+ population in 1997 to 22% in 2016, and it is projected to rise to 23% by 2020, accounting for most of the growth in the 16+ population in the ONS projections (Chart A). A rising average age, all else equal, tends to reduce labour supply growth, since it reduces the proportion of people participating in the labour market relative to those in retirement. Currently just over 10% of those aged over 65 are in work or seeking work, compared with 79% of those aged 16 to 64.
Despite this effect from population ageing, UK labour market participation has been stable in recent years. This is in part because the average participation rates of older people have increased. As explained in the box on pages 30–31 of the November 2014 Report, a number of factors are likely to have contributed to that increase, including better health and improved longevity, and rises in the state pension age. The stability in aggregate participation also reflects other developments, including the continued rise in the proportion of women in or seeking work. The MPC judges that the participation rate is likely to remain broadly flat over the forecast period, as the factors supporting participation continue to offset the effect of demographic shifts.
Another feature of UK demographics over the past decade has been the significant contribution of net migration to growth in the UK workforce (Table 1). Since the EU referendum, however, levels of long-term migration have fallen and the ONS projects a further gradual fall in coming years (Chart B).1 All else equal, that will reduce the pace of growth in UK labour supply slightly.
There is a chance that net migration could fall more sharply than the gradual decline implied by the ONS projections. There tends to be a positive relationship between migration flows to the United Kingdom and economic conditions in the United Kingdom relative to those in migrants' home countries. Bank staff analysis suggests that the subdued outlook for UK GDP per capita, combined with stronger growth prospects in other countries (Section 1), would, on its own, reduce net migration by a little more than implied by the ONS projections over the next three years. In addition, net migration is likely to be affected by any changes to institutional arrangements for the movement of labour, or uncertainty around those arrangements.
It is possible that lower net migration from the European Union could have effects on potential supply over and above those that arise simply from their effect on population growth. Data from the LFS, for example, suggest that migrants from the European Economic Area tend to be more likely to participate in the labour market than those from the domestic population, in part because these migrants also tend to be younger. They are also more likely to hold degrees than people in the domestic population. To the extent that these degrees are associated with higher levels of skills and productivity, a fall in net inward migration could affect overall UK productivity growth. Given the number of migrants relative to the size of the existing UK population, however, combined with the fact that — despite their higher qualifications — migrants tend not to be disproportionately represented in higher-skilled occupations, these effects are likely to be small.
What matters for inflation is not only the impact of changes in the population on potential supply, but also the effect on aggregate demand. As explained in the box on pages 30–31 of the May 2015 Report, the impact of a change in labour supply on GDP growth and inflation will depend in part on the reason for that change. In general, a rise in labour supply caused by higher net inward migration tends to have only a small impact on aggregate wage growth and inflation, since it results in a contemporaneous increase in aggregate demand as migrants begin to spend straight away.2
Particularly abrupt falls in migration as a consequence of the Brexit vote could, however, result in labour shortages in sectors that have become reliant on migrant labour, and hence greater pricing pressures within those sectors. Shortfalls of seasonal foreign workers — who may not be captured within the official long-term migration statistics — have been widely reported by contacts of the Bank's Agents within sectors such as logistics and food processing.
1 The official long-term migration statistics shown in Chart B define a long-term migrant as 'a person who moves to a country other than that of his or her usual residence for a period of at least a year, so that the country of destination effectively becomes his or her new country of usual residence'.
2 See, for example, Nickell, S and Saleheen, J (2015), 'The impact of immigration on occupational wages: evidence from Britain', Bank of England Staff Working Paper No. 574.
Most of the growth in the UK workforce over the past decade has been accounted for by population growth
Contributions to changes in size of the UK workforce
Older people account for most of 16+ population growth in the ONS's projections
Contributions to annual 16+ population growth (a)
Net inward migration continues to fall slightly in the ONS's projections
Decomposition of net inward migration by nationality (a)