In focus - Supply and spare capacity

4.1 The amount of spare capacity in the economy

The MPC judges that there is a margin of spare capacity in the economy…

The MPC uses a range of approaches to estimate potential supply and spare capacity. ‘Top-down’ approaches look at the evolution of output as well as indicators of the balance between demand and supply, such as what is happening to prices.

One specific top-down approach is to use statistical filtering techniques. These estimate spare capacity using past observations of output, taking into account indicators of labour market and price developments, for example. Those statistical filters separate actual output into a trend component and a cyclical component. The trend is often interpreted as the measure of the economy’s potential supply capacity and the difference between that and actual output is the margin of slack (see Kuttner (1994) for an example of this type of model).

One important input to the MPC’s potential supply assessment is a statistical filter that uses well-established macroeconomic relationships between GDP, unemployment and domestic inflation (Melolinna and Tóth (2016)). For example, if output growth was weak and there was high unemployment and low inflation, the filter would be likely to signal that there is spare capacity in the economy. This model suggests that there is a degree of spare capacity in the economy — currently around ½% of GDP.

Bank staff have also developed other filters, such as one that uses a broader range of labour market variables including the participation rate and hours worked (see Fleischman and Roberts (2011) for a similar model). This model also suggests that there is spare capacity, although the amount is a little smaller.

An important feature of these models is that potential supply growth is assumed to be less volatile than output growth, consistent with a view that structural changes to the economy tend to happen slowly. That means that when GDP growth slows quickly, the models tend to interpret that as being driven mainly by demand rather than potential supply. Some shocks could affect potential supply growth relatively rapidly, however, and a filter might struggle to pick those changes up. At present, for example, the UK’s potential supply capacity might have been affected by Brexit-related factors.

…and that margin has been wider over the past few years than previously thought.

More generally, over the past year, CPI inflation has fallen and price-based measures of domestically generated inflation have been subdued (Section 2). Core inflation has recently been somewhat weaker than the MPC expected. That has occurred despite the firm growth of wages and labour costs. While some of the relative weakness in inflation could reflect weakness in non‑labour costs or developments in the retail sector (Section 2), the MPC judges that it also signals that spare capacity in the economy has been a little greater over the past few years than previously thought. Actual output is currently judged to be ½% lower than potential GDP.

The MPC also uses a ‘bottom-up’ approach to determine where spare capacity is located, separately considering the amount of spare capacity in the labour market and the amount within firms. As set out below, the evidence suggests that, given the tightness of the labour market, spare capacity is located within companies.

Spare capacity in the labour market

There are few signs of spare capacity in the labour market.

There are various places in which spare capacity within the labour market can be located. People active in the labour market but currently without jobs — summarised by the unemployment rate — could find work. People not active in the labour market currently could enter it, raising the participation rate. Or people already in work could increase their hours — indicated by movements in average hours worked. The MPC judges that across all of these components, there are few signs of spare capacity.

Unemployment remains below the MPC’s estimate of its equilibrium rate…

In February 2019, the MPC judged that the long-term equilibrium rate of unemployment was around 4¼%, similar to its assessment in 2018. That equilibrium rate is determined by the structural characteristics of the labour market such as the tax and benefit system, and the efficiency with which the skills of workers can be matched with the skills that companies are seeking.

One indication that the equilibrium unemployment rate has remained around 4¼% since then is the increased pace of wage growth as the unemployment rate has fallen (Chart 4.2). When unemployment falls below equilibrium it is likely to exert increased upward pressure on pay growth as companies need to pay more to recruit and retain staff with suitable skills. Private sector regular pay growth increased from around 3% over 2018 to a high of 4% in July 2019, although it has fallen back slightly in recent months (Section 3). A statistical filtering technique which assumes a linear relationship between unemployment and wage growth, as in a standard wage Phillips curve,^[1] suggests that unemployment is below its equilibrium rate.

Another indicator of the equilibrium rate is the flows into and out of unemployment. Those flows are partly determined by the structural features of the economy. For example, the flow out of unemployment into employment — the job-finding rate — reflects in part the efficiency with which people are matched to new jobs, and could increase due to technological progress, increased educational attainment or improved skills in the workforce (Elsby, Michaels and Ratner (2015)). The job destruction rate — the flow of people out of employment into unemployment — can be affected by factors such as the flexibility of employment contracts. If the job-finding rate improves or the job destruction rate falls due to such structural factors, then the equilibrium unemployment rate would be expected to fall.

Over the past year, the job-finding rate has risen and the job destruction rate has remained low (Chart 4.3). If those rates were to persist, unemployment would settle at a rate lower than 4¼%. But those rates are likely to reflect in part strong labour demand rather than structural factors. Despite the slowing in GDP growth and some signs of softening demand for labour, the labour market remains very tight (Section 3).

Chart 4.2 Wage growth has picked up as the unemployment rate has fallen

Wage Phillips curve: wage growth and unemployment

(a) Whole-economy regular pay. Three-month average growth on the same period a year earlier.
(b) Diamond for 2019 Q4 shows Bank staff’s projections, based on data to November.

Chart 4.3 The job-finding rate has risen while the job destruction rate has remained low

Flows between employment and unemployment ^(a)

Sources: ONS and Bank calculations.

(a) Two-quarter flows. Flows between employment and unemployment are based on total employment and unemployment of people aged 16–64. Dashed lines are averages from 2002 to 2007.

The MPC judges that the long-term equilibrium unemployment rate has remained at around 4¼%. Unemployment is projected to remain below its equilibrium rate over the forecast period, which is expected to exert continued upward pressure on wage growth. But the extent of upward pressure may be reduced by the composition of unemployment. Recently, the number of people who have been unemployed for more than six months has fallen (Chart 3.11). Those people in longer-term unemployment tend to put less upward pressure on wages and inflation, as they tend to be less likely to find a job than those who have been out of work for shorter periods of time (Gordon (2013)). So wage growth is likely to be boosted less by a reduction in unemployment driven by lower long-term unemployment than short-term unemployment.

…and spare capacity appears to have been largely absorbed elsewhere in the labour market.

Other indicators of labour market slack also suggest little spare capacity. The participation rate remains high, at 64.1%, as the proportion of people who want to work has increased within certain demographic groups in recent years. For example, rises in the state pension age, as well as improved health and longevity, have raised the participation rates of older workers (Saunders (2018)). Female participation rates have also risen in recent years. The ‘marginal attachment ratio’ — the proportion of the working-age population who are not currently in work or seeking employment but report that they would like a job — has stabilised during 2019 at a historically low level (Chart 4.4). That suggests that the scope for more people to enter the labour market is limited.

The average number of hours worked per person has been stable over the past year, according to official data. The number of hours that people say they would like to work, over and above those they are currently working, has also been close to zero in recent quarters (Chart 4.5). That suggests that average hours are around their ‘desired’ or equilibrium level.

Chart 4.4 The proportion of people not currently looking for work, but who would like a job, has been stable over 2019 at a low level

Marginal attachment ratio ^(a)

Sources: ONS and Bank calculations.

(a) Number of those aged 16–64 who say they are not in work or not actively looking for work but would like a job, as a percentage of the 16–64 population.

Chart 4.5 People no longer wish to work longer hours, on average

Net additional desired hours ^(a)

Sources: Labour Force Survey and Bank calculations.

(a) Number of net additional hours that the currently employed report they would like to work, on average, per week.

Spare capacity within companies

Productivity growth has been weak since the crisis and has slowed further in recent years.

Within companies, spare capacity reflects the extent to which output produced by each employee is below its potential level. In other words, the extent to which companies’ capital equipment, such as vehicles or computers, is being underutilised. Spare capacity within companies will mean that labour productivity is below its potential level.

Labour productivity growth has been persistently weak since the financial crisis, and has been generally lower than the MPC’s forecasts. Over the past couple of years, it has slowed further. It is expected to have been below zero on both an output per-head and per-hour basis in the year to 2019 Q4 (Chart 4.6), weaker than expected a year ago.

Some of the weakness is likely to reflect spare capacity…

The MPC judges that some of the recent weakness in productivity growth is likely to be cyclical, reflecting the emergence of spare capacity within companies. That is consistent with results from survey measures of capacity utilisation, some of which have fallen sharply over 2019 to below average (Chart 4.7). In addition, reports from the Bank’s Agents suggest some firms have hoarded labour in case the slowdown in output growth is temporary.

As well as slowing domestic demand growth (Section 2), the cyclical slowing in productivity growth might reflect global factors. Increased trade protectionism and the associated rise in uncertainty have contributed to a weakening in global demand growth. That is likely to have weighed on productivity growth, particularly in the highly traded manufacturing sector. Productivity growth in manufacturing has slowed in all G7 countries (Chart 4.8), especially in some euro-area countries with which the UK has close supply-chain links.

Chart 4.6 Productivity growth is expected to be below zero in the year to 2019 Q4

Measures of labour productivity ^(a)

Sources: ONS and Bank calculations.

(a) Output is based on the backcast for the final estimate of GDP. Diamonds show Bank staff’s projections for 2019 Q4, based on data to November.

Chart 4.7 Survey measures of capacity utilisation have fallen over 2019

Survey indicators of capacity utilisation ^(a)

Sources: Bank of England, BCC, CBI, CBI/PwC, IHS Markit/CIPS, ONS and Bank calculations.

(a) Differences from averages between 2000 and 2007. Measures are from the Bank’s Agents, the BCC (non-services and services), the CBI (manufacturing — capacity; financial services, business/consumer/professional services and distributive trade — business relative to normal) and IHS Markit/CIPS (manufacturing — backlogs; services — outstanding business). Sectors are weighted using shares in gross value added. The BCC data are not seasonally adjusted.

…but some of the weakness is likely to reflect structural factors.

Some of the recent weakness in productivity growth is likely to reflect structural factors, rather than weaker demand. In other words, underlying productivity growth has slowed.

While changes in the pace of underlying productivity growth tend to emerge slowly, one factor that may have affected it more rapidly is Brexit. Evidence from the Decision Maker Panel (DMP) Survey suggests that Brexit may have reduced the level of actual UK productivity by as much as 2% since mid-2016.^[2]

Chart 4.8 Productivity growth in the manufacturing sector has fallen in the G7 economies

Manufacturing sector hourly labour productivity in the G7 countries ^(a)

Sources: Eikon from Refinitiv, Eurostat, OECD, ONS and Bank calculations.

(a) Manufacturing productivity is calculated as gross value added divided by hours worked.

Chart 4.9 Productivity growth has been weak in manufacturing and finance

Contributions to hourly labour productivity growth ^(a)

Sources: ONS and Bank calculations.

(a) Sectoral output per hour is calculated as gross value added (GVA) divided by hours worked. Figures in parentheses are weights in GVA. Other is calculated as a residual and includes other sectors and allocation effects.

One way Brexit is likely to have weighed on productivity growth is through heightened uncertainty. Heightened uncertainty has weighed heavily on business investment, as it has incentivised firms to delay spending until they have more clarity around the future trading relationship between the UK and the EU (Section 4, November Report). That has caused capital per worker to grow more slowly, lowering labour productivity growth.

Another way Brexit is likely to have weighed on productivity growth is through the time and effort spent preparing for Brexit to help ensure a smooth transition. Companies in the DMP have reported spending money — a little above 1% of gross value added, on average — on Brexit planning, which might otherwise have been invested in other activities. Companies also reported that they were spending senior management time on Brexit planning. That could have diverted time away from other productive activities.

Looking across sectors, the time and money spent on planning for Brexit has been correlated with weaker productivity growth. Since the referendum, productivity growth in the manufacturing and finance sectors has remained weak (Chart 4.9), and these sectors have also spent the most time on Brexit planning (Chart 4.10).

Based on data from the DMP, Bank staff estimate that the time and resources spent on Brexit planning have reduced productivity growth since the referendum, with the largest impact in 2019 (Chart 4.11). Consistent with that, a decomposition of productivity growth derived from a structural economic model suggests that negative supply shocks have dragged on productivity growth over the past few years.

Four-quarter underlying productivity growth is judged to have slowed to around ⅓% at the end of 2019, from around 1% two years ago, reflecting in part the impact of Brexit-related factors.

Chart 4.10 The manufacturing and finance sectors have spent the most time planning for Brexit

CFO time spent on Brexit planning by sector ^(a)

Sources: DMP Survey and Bank calculations.

(a) Question: ‘On average, how many hours a week is the CFO of your business spending on preparing for Brexit at the moment?’. Respondents are asked to choose between ‘None’, ‘Up to 1 hour’, ‘1 to 5 hours’, ‘6 to 10 hours’, ‘More than 10 hours’ and ‘Don’t know’. Responses collected between August and October 2019. Point estimates are constructed by using values of 0, 0.5, 3, 8 and 15 for the respective categories.

Chart 4.11 Brexit planning has increasingly weighed on productivity growth since the referendum

Impact of Brexit planning on labour productivity growth ^(a)

Sources: Bureau van Dijk, DMP Survey and Bank calculations.

(a) Bank staff estimates based on DMP Survey responses and company accounts data. Brexit planning comprises time (CFO and CEO hours) and resources spent planning for Brexit. Unweighted regression based estimates are scaled to be consistent with aggregate effects estimated by Bloom et al (2019).

4.2 The outlook for potential supply growth

The supply capacity of the economy depends on the amount of available labour — potential labour supply — and how productively that can be put to use.

Underlying productivity growth is expected to remain weak…

Productivity growth is estimated to have averaged around ½% per year since the financial crisis, compared with 2¼% in the previous decade. The MPC judges that it is unlikely to recover substantially over the forecast period, given how persistently weak it has been in the past.

Brexit will continue to weigh on productivity growth over the forecast period. The MPC’s forecasts assume that there is an immediate but orderly move to new trading arrangements with the EU on 1 January 2021. That is expected to increase trade barriers which weigh on productivity growth, given its well-established relationship with openness (see Box 1, November Report). There are risks around the extent of the drag that higher trade barriers exert on productivity growth, however.

Nonetheless, productivity growth is expected to recover somewhat. The impact of Brexit planning on growth fades, as companies are expected to spend a similar amount of time and money as they have recently. Business investment growth is expected to pick up as Brexit-related uncertainty falls back. That supports a slight pickup in capital deepening — the amount of capital available per hour worked — which has weighed on productivity growth over the past year (Chart 4.12).

The recovery is supported by a pickup in the growth of total factor productivity — the efficiency with which capital and labour are used to produce output. Investment in intangibles — assets which are not physical in nature such as software — has been found to be a key driver of innovation and productivity growth. This is particularly the case for research and development (R&D) expenditure, which has been steadily rising for many years (Chart 4.13), with evidence suggesting it provides a boost to productivity growth after two to six years, on average.^[3] More generally, improved capital stock estimates from the ONS implemented in Blue Book 2019 show that the share of intangibles in total capital is higher than previously thought, and has been rising in recent years.

Chart 4.12 Capital deepening has weighed on productivity growth since the crisis, on average

Contributions to four-quarter growth in hourly labour productivity

Sources: ONS and Bank calculations.

(a) Market sector output per hour.
(b) Calculated as a residual.

Chart 4.13 Research and development expenditure has increased

Expenditure on tangible and intangible investment and research and development ^(a)

Sources: ONS and Bank calculations.

(a) Annual averages. Data for R&D expenditure are to 2018. Data for tangible and intangible investment are to 2019 Q3. 2019 Q4 is assumed to be equal to an average of Q1–Q3.

…while potential labour supply growth is modest.

Potential labour supply growth is modest over the forecast period, and is driven by population growth (Table 4.A). The MPC’s forecast is conditioned on the ONS’s principal population projection, published in October 2019. That implies that the working-age population will continue to grow at a similar pace to recent years.

Potential labour supply grows at a somewhat slower pace compared to recent years. That is in part because the equilibrium unemployment rate is not projected to decline as it is estimated to have done in the 2015–19 period. It also reflects the trend participation rate flattening off at around 64%, having been rising in recent years. That is the net result of two offsetting factors: the ageing of the population, which will tend to pull down on participation rates; and increases in participation within older age groups. Desired average hours continue to fall, as they have done recently.

Taken together, potential supply growth is projected to remain subdued, and weaker than expected a year ago.

The MPC judges that potential supply growth will remain subdued over the forecast period, at around 1% on average. It initially falls a little from its current rate of around 1%, before rising to around 1½% in 2023 Q1.

The projection for potential supply growth is weaker than the MPC expected one year ago. In part, that reflects the weakness of productivity growth over the past year, which extends the pattern seen since the financial crisis. Brexit‑related factors are also expected to have a larger effect over the forecast period (Section 1).

Weaker potential supply growth reduces the pace of GDP growth that is consistent with the MPC meeting its 2% inflation target — it acts as a ‘speed limit’ on the economy.