Trend output and the output gap in the UK.

Barrell, Ray ; Kirby, Simon

This note looks at estimates of the current scale of the output gap in the UK and at the factors that affect estimates of the trend rate of growth. These issues are central to the debate on macroeconomic policy, both in the short and the long run. The speed at which the economy returns to full capacity, along with the scale of the output gap, will be important factors affecting average growth over the next five years. In the longer term, trend growth at full capacity is not immutable, but rather depends upon the rate of labour augmenting technical progress and the growth of the labour force. In the medium term these factors can be added to by temporary bursts of capital augmenting technical progress and by changes in the user cost of capital that may change the optimal capital-output ratio. Other factors, such as the cost of materials, also affect potential output. Over the past few months there has also been a significant rise in oil prices, and we judge this to have a strong permanent component which will reduce trend growth in the short term and trend output in the longer term. Its implications are more fully discussed in Barrell, Delannoy and Holland in this Review.

We first look at the factors affecting the level of capacity output in the UK, returning to the discussion in Barrell (2009). In particular we assess the evidence on the longer-term impact of the financial crisis on output. Between the first quarter of 2008 and the third quarter of 2009 output fell by 6 1/4 per cent, and some of this fall may be permanent, in that the recession will leave a scar. We then look at the factors affecting the growth of the supply of output, with a particular emphasis on factors affecting the supply of labour. This can be augmented by increasing the participation rate of those under retirement age (the population of working age), by extending working lives as discussed in Barrell, Kirby and Orazgani (2011), or by increasing the level of net immigration, whose effects are discussed in Barrell, FitzGerald and Riley (2010). The third section looks at national saving and net investment. Output and income can be augmented by the accumulation of capital and of income earning assets respectively, but both have been declining as a proportion of National Income, as we discuss. In an open economy with capital mobility an increase in saving will lead to an increase in wealth and income. However, it is unlikely to lead to an increase in capital and output, as the optimal capital stock depends upon the cost of capital. This in turn depends upon factors such as the internationally determined risk-free rate of return. In the final section we look at the factors affecting trend growth in the longer term, building on Barrell, Holland and Liadze (2010) as well as utilising the forecasts and projections undertaken for the Review. We look in particular at technical progress, migration and extending working lives.

The impact of crises on sustainable output

There are a number of ways to estimate the scale of the output gap, and these vary from the use of mechanical filters to detailed analyses of production functions. We rely on the latter, and include the former only to warn against their use in serious policy discussions. The scale of the output gap depends on the fall in output since its peak, which is over 6 per cent, the degree of excess capacity use at that peak, which we estimated at 1 to 2 per cent, the underlying or trend growth of the economy over the period excluding the impact of the financial crisis and also the impact of the crisis on trend output. One could perhaps scale the effects of the crisis on potential output by looking at previous crises and we do this first. The conclusions from our survey support our result that the financial crisis has probably reduced trend output by 3 to 4 per cent and hence we currently have an output gap of at least 4 per cent of GDP.

Interestingly, the majority of financial crises in the OECD in the past 70 years have had no long-run impact on output, and they are not always associated with recessions. In particular, in a cross-country study of

OECD countries for the European Commission, Barrell, Holland and Liadze (2010) find that there is no evidence that any postwar financial crisis prior to 2007 in the UK had a permanent impact on output. There have been a large number of financial crises in the OECD countries since 1970, and as Barrell (2004) discussed, they have been becoming more frequent as markets have been less well regulated. However, not all crises are the same, and any study of crises should test this proposition. Cerra and Saxena (2008) look at crises in OECD and emerging markets using a vector autoregression analysis, and they assume they all have the same effect. As a consequence they find large long-run effects, with crises on average reducing output by 8 per cent in the long run. Cecchetti et al. (2009) do not impose commonality, and look only at OECD and more advanced emerging markets, finding both smaller effects and also that only 9 out of their 40 crises have a significant long-run impact on output. Neither of these studies take account of other factors affecting trend output, such as the disruption caused by sharp increases in oil prices, and this may matter, especially in the current conjuncture much as it did in the 1970s. (1)

There are many other factors affecting trend output, and it is important to abstract from these when looking at the impact of crises on trend output and growth. Barrell, Davis, Karim and Liadze (2010) look at crises in thirteen OECD countries between 1980 and 2008, and take into account other influences on trend output such as Research and Development expenditure and changes in trend labour input (from migration as well as from other sources). They find that only a quarter of all banking crises they study have a significant impact on trend output, and some, such as that in Canada in the early 1980s, are not associated with a subsequent recession. On average, banking crises in OECD countries are associated with a reduction in trend output of around 2 1/2 per cent. However, the crises that do significantly affect trend output have similarities in that they are more system-wide in their nature. Systemic crises, where the majority of the banking system collapses, such as those we saw in Japan in 1991, the US in 1988 and Sweden (2) and Finland in 1991, reduced equilibrium output by over 4 per cent, however. The current crisis is systemic in a large number of countries, and these estimates suggest that it will probably also reduce sustainable output by around 4 per cent, leaving an output gap of 4 per cent or more in the UK.

[FIGURE 1 OMITTED]

Mechanical filters are commonly used in economics, and they are often applied to output data to estimate the trend and the scale of the output gap. The Hodrick Prescott (HP) filter is the most commonly used tool, mainly because it is easy to calculate in common packages. As Barrell and Sefton (1995) showed, it is subject to severe endpoint problems and it is inappropriate to use it across an unknown structural break at the end of a data series. It tends to follow the data too closely as it is a symmetric filter, and it effectively double counts the last observations unless it is used with at least five years of forecast as in IMF (2010). (3) In an assessment of filters Christiano and FitzGerald (2003) describe the HP filter as 'worse than useless' for extracting trend output, and we would concur with this conclusion. They advocate the use of the Baxter King style bandpass filter which obviates the need for a forecast. However, this filter is still prone to problems when there is an unknown structural break. Figure 1 plots actual and trend output using this filter. This mechanical tool would suggest that the output gap is at present relatively small and that output was well above capacity in 2007. The lack of serious inflationary pressure in 2007 makes this seem questionable, and the IMF (2010) use inflation as one of their conditioning variables to extract a noticeably larger output gap.

Our preferred approach to calculating the impact of the crisis is similar to that of the OECD and the IMF in that it involves the use of a production function that relates trend output to inputs of labour and capital as well as oil. In particular, we judge that the crisis in the financial sector was in part the consequence of under-pricing risk, and that this will not happen again in the next few years, and hence the effective user cost of capital will rise and the equilibrium capital-output ratio will fall, much as Haugh et al. (2009) suggest happened after previous systemic financial crises. In our global model, NiGEM, we use an estimated and calibrated production function, with an elasticity of substitution between capital and labour of a half, and an elasticity of substitution between other inputs and energy (oil, coal, gas) of one. Hence a rise in the user cost of capital will change the equilibrium level of output, but this will be partly offset by the impact of lower investment on the equilibrium real interest rate as Barrell, Holland and Karim (2010) discuss.

[FIGURE 2 OMITTED]

We judge that the financial market crisis that started in the summer of 2007 and worsened in the autumn of 2008 has led to a sharp short-term decline in output that exceeds its longer-term impact. Using our production function approach we estimate that the permanent scar on output per person hour in the UK might be around 3 per cent. The scar is largely driven by a presumed 2-300 basis point rise in risk premia. The risk premium effect is, as far as we can see, likely to be common across countries, which we can judge from figure 2, which plots the BAA spread on commercial bonds over risk-free government bonds in the UK and the Euro Area over the past decade. There was a major peak during the financial crisis, which may be seen as the shadow price of bank credit, and after that the spread subsided and seems to have settled at or above its 2008 pre-Lehman's shoulder.

Evidence is accumulating that recessions also leave a scar on the number of person hours available to put through the production function. This can either come through its effects on migration, as is discussed in Barrell, Gottschalk, Kirby and Orazgani (2009) or it can come through the reduction in skills and the consequent increase in sustainable unemployment that recessions may produce. As Bell and Blanchflower (2010) discuss, periods of unemployment, especially early in a lifetime, reduce potential earnings and productivity for sustained periods of time. If productivity is reduced by a recession, then workers will find it harder to find jobs, and unemployment spells are likely to be longer, and hence equilibrium unemployment will be higher. We suspect unemployment will settle almost a full percentage point higher after 2015 than observed during the first decade of this century, as we can see from figure 5 below. However, as with all forecasts, we include neither further recessions nor booms, as they should be seen as equally likely, and they may also affect the equilibrium.

As a result of the financial crisis, and given the capital-output ratio in the UK and the level of the user cost of capital, we would judge that output will settle on a trajectory 3 to 4 percentage points lower than it would have been otherwise. (4) We should add to this the 0.5 per cent impact on trend output per person hour of the sustained rise in oil prices we have seen since late 2010. (5) Hence we estimate that sustainable output is currently 4 1/2 per cent lower than we would have expected looking forward from 2007.

The supply of output and factor inputs

The permanent increase in the user cost of capital that follows from a re-evaluation of risk after the crisis is expected to have reduced the equilibrium capital stock. This implies a downward adjustment of investment in order for this to happen. The level of investment in the economy fell by 5 per cent in 2008 and by 15 per cent in 2009, and we expect it to rise by under 4 per cent a year on average between 2010 and 2015. Public sector investment was increased rapidly during 2008 and 2009, growing at an average rate of over 20 per cent a year. Current plans would suggest it will fall by around 10 per cent a year for the next three years, returning it to similar levels as seen in 2007. Housing investment was weak in 2007 and fell rapidly in 2008 and 2009, leaving it at just over half the level seen before the crisis started. Although there remains a potential shortage of housing, we do not expect a rapid recovery without a significant change in planning laws. Business sector investment fell by almost 20 per cent in 2009, and recovered only weakly during 2010, but the third quarter looked more promising. The 20 per cent real devaluation of sterling we have seen since 2007 is likely to strengthen manufacturing relative to the rest of the economy. As manufacturing is more capital intensive than other sectors, this may mean business investment may recover. However, 2011 may see business investment growth of over 5 per cent, and it is unlikely to average much above 5 per cent per annum over the next five years. At these rates of growth, we do not expect the pre-recession peak in business investment to be recovered until 2014, as we can see from figure 3.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Figure 4 plots recent movements in employment as well as in self-employment. The number of self-employed has risen by around 200,000 since the start of the recession in early 2008, whilst total employment has fallen by 260,000. There has been a trend-like growth in self- employment since 1997, and it has not been affected by the downturn and, as we can see from the figure, the trend may have been strengthened. Employment began to recover in early 2009, with a strong increase in part- time employment. Indeed, there has been an increase in employment of 280,000 between the last quarter of 2009 and the third quarter of 2010, whilst dependent employment (employees in employment) has risen by 200,000 over the same period. It would appear, however, that recent improvements in the outlook for labour markets are in part due to firms reassigning employees and the newly-employed to part-time instead of full-time contracts. The rise in part-time employment has been steep since the first part of 2008. Indeed, by the fourth quarter of 2010, 4 per cent of those in employment, or 1.16 million people, were in part-time work because they could not find a full-time job.

The Labour Force Survey based measure of unemployment has started to rise again, after a period of minor falls, as we can see from figure 5. In the three months to November it rose 0.2 percentage points to 7.9 per cent, which represents a total of 2.5 million people unemployed. In contrast, December estimates of the claimant count indicate that the number of people claiming the Jobseeker's Allowance fell, albeit by only 4,100. We do not expect employment to continue to rise and unemployment to continue to fall in the short term, as output growth is likely to weaken in the short term and we expect this to be reflected in more labour shedding. As output growth remains below the potential rate, and given there has also been a rise in part-time employment, we expect unemployment to rise from the end of 2010, peaking at just over 2.8 million in the third quarter of 2011. Public sector job losses have been scaled down as a result of the change in the structure of the plans announced in the 2010 Spending Review, but we would still expect about 400,000 to be shed over the next four years. This will put upward pressure on unemployment, especially of youths and those returning to the labour market, as those released from the public sector should be more skilled than average and should find it relatively easy to flow into jobs, displacing other groups. This change will have lasting effects as longer spells of youth unemployment increase sustainable unemployment. Unless we see another boom, the financial crisis is likely to leave a sustained scar on the sustainable level of unemployment, much as the recession of the early 1980s appeared to have done.

Given the severity of the downturn and the scale of public sector job losses, unemployment is not rising as rapidly as might have been anticipated given that it exceeded three million in the less severe downturn at the start of the 1990s. This may in part be because labour markets appear to have changed and some of the shock to the economy has been absorbed in real wages, as is discussed in Holland et al. (2010). In addition, the nature of the recession appears different from some in the past. The impact of the rise in risk premia is very widespread with all firms having to adjust their capital stock. There has been no particularly noticeable sectoral collapse in demand and manufacturing has been recovering robustly. As a result, insolvencies have been rising much less rapidly than we might have anticipated. Changes in legislation make comparisons of insolvency rates over time problematic, but changes in rates have more meaning. In the recession in the early 1990s insolvencies rose by 140 per cent between the fourth quarter of 1989 and the first quarter of 1993 when they peaked, whilst between the third quarter of 2005 and the last quarter of 2009, also a relative peak, they rose by less than 30 per cent. Although insolvencies have been falling since then as credit conditions have eased, they may start to rise again. However, they are an indication of the different nature of this recession compared with the previous one. The early 1990s saw an overvalued exchange rate and very high real interest rates along with high nominal rates, and these factors caused severe financial stress amongst small and medium sized enterprises. Although new credit is not easily available, real and nominal interest rates are very low at present, and the real exchange rate fell sharply at the start of the recession, and not at the end of it as in 1992.

Accumulation and national wealth

It is important to distinguish between the determinants of income growth and those of output growth. The longer-term evolution of income depends in part upon the accumulation of income earning assets at home and abroad by UK residents and by foreigners in the UK. The income flows for UK residents depend upon their own saving and that of the government. The need for savings depends in part on the risks individuals face, and the more volatile the economy the more likely we are to see precautionary saving. Indeed the past three years may have induced people to save more for this reason alone. The major reason for saving, however, is for retirement, and the longer are working lives the less we might need to save. Retirement ages differ noticeably between countries, as we can see from table 3, with the UK being in the middle of the range. Countries such as the US, where retirement is later than in Europe, but where life expectancy is similar, need to save less than countries such as France. National savings needs, however, also depend on the age structure of the population, and saving should rise in advance of any increase in the proportion of the population above retirement age. This has not been happening in the UK.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Over the past two decades national wealth in the UK has been falling in relation to national output, as we can see from figure 6. The wealth of the nation is made up of the value of the stock of income-earning reproducible assets, and these consist of the physical capital stock (excluding land) and the net stock of foreign income-earning assets. Figure 6 plots the stock of net foreign assets as a share of output, which was around zero in the early 1990s and has been falling since then mainly as a result of continual balance of payments deficits. The value of the capital stock as a proportion of output has also been falling, in part because of the changing nature of production, which has moved away from capital intensive manufacturing towards services. The resulting decline that we see in overall national wealth has not been accompanied by a noticeable increase in rates of return, and hence the share of income available to pay for retirement without transfers from those in work has been falling. As we can see from figure 7, the proportion of the population above working age has not been falling, and is expected to rise.

Given the age structure of the population and current retirement ages, along with increases in expected lives that are anticipated over the next decade, it is reasonably clear that a further improvement in the net saving balance of the nation is needed to ensure we can pay our pensioners at the rate at which they have been promised. Table 4 sets out the components of the accumulation of wealth in the UK both over the past decade and over our forecast horizon. Net national saving is expected to rise, but even by 2015 it will be below the level seen in the 1990s. However, as alternatives to raising saving, it is possible either to renege on these promises or to raise the age at which retirement takes place. (6)

Output growth, migration and retirement

Capital services inputs and the available supply of labour help determine output along with the skills of the labour force and technical progress or knowledge that is not directly embodied within labour. Longer-term projections for output growth after such a severe recession are inevitably clouded by projections for the speed with which the economy returns to capacity, wherever that may be. The discussion at the start of this note would lead us to think that we have a long-run structural output gap of around 4 per cent at present. As we expect this to close over the next decade it would not be surprising if output growth were to average around 0.4 percentage points more than the underlying trend over that period. This calculation affects the interpretation of the figures in table 5 below. In the decade between 1997 and 2007 hourly productivity growth in the UK was robust, and it was bettered only by the US, Finland and Sweden in the group of countries (which includes Japan, Italy, Canada, Germany, France, and Spain) studied by Barrell, Holland and Liadze (2010). They attribute this relative improvement in performance to skill improvements and increases in competitive pressures. Improvements in the skills of the workforce were probably the most important, especially the increase in the number of graduates which allowed the UK to take advantage of the product innovations associated with developments in biotechnology and computing. Total factor productivity growth is driven by labour-augmenting and capital-augmenting technical progress, and much of the former is embedded in the skills of the workforce. Going forward, productivity growth is likely to be slower than over this period, in part because we may be at the end of what has to be a temporary period of capital-augmenting technical progress, and also because the effects of skills accumulation have already started to decline with the compression of the graduate mark-up. Given the increase in risk premia we discuss above, capital deepening in excess of that required by labour-augmenting technical progress, and hence growth going forwards, will depend upon the (slowing) rate of technical progress and increases in labour supply. (7)

Between 1997 and 2007 around 60 per cent of the almost three million increase in employment came from migration, as defined by country of birth. By the fourth quarter of 2008 around 3.8 million workers in the UK were born abroad, whilst total employment was 29.4 million. It was widely anticipated that the severity of the recession would reduce the stock of migrants in the UK, as major source countries such as Poland and the Indian sub-continent were not suffering as badly as the UK. (8) As we can see from table 6, the numbers of those born abroad who were in employment fell more rapidly than average between the end of 2008 and late 2009. However, since the end of 2009 the numbers of those born abroad who are in employment has risen, and in the third quarter of 2010 it was higher than in late 2008. The increase is more than accounted for by the increase in the stock of migrants from the eight new member states of the European Union, including Poland. We would expect the introduction of more stringent controls on immigration to restrict the growth of the supply of labour to below that seen over the earlier period, but it cannot affect the flow from EU members.

Given these factors, it is hard to see that over the next decade trend growth in the UK will be much above 2 per cent a year, unless an effort is made to increase the supply of labour by raising retirement ages. As Barrell, Kirby and Orazgani (2011) show, current legislation on state pension ages (which excludes the proposed increase at the end of the current decade) will raise the workforce by almost 3 million by 2030, adding about 0.25 per cent per annum to trend growth over this period. The study assumed average retirement ages for men and women of 65 in 2020, and 66 in 2026. If we wish to raise trend growth and average incomes over the next decade, the most effective way to do this would be to raise the age at which people leave the labour force. The abolition of the default retirement age is a step in this direction, as it is possible that up to 10 per cent of the workforce exit employment into retirement earlier than they might like. Retirements are bunched, as figure 8 would suggest, and unbunching them might extend the labour force, but only by about 0.2 percentage points. As with all extensions to working lives, the effects on unemployment will exist, but they will be temporary, and will be experienced in particular by groups with high turnover such as the young and the unskilled. Recently announced extensions to work experience allowances would help ameliorate these problems.

[FIGURE 8 OMITTED]

Probably the most effective way to raise trend growth over the next decade would be to raise the retirement age by three years relatively quickly. Labour markets are able to absorb increases in the supply of labour, as we saw with the increase in the stock of migrants after 1997, but they do need some assistance. Growing demand and policies to reduce youth unemployment are essential complements to raising the retirement age. If conditions are right, as they should be over the next decade, three years on working lives would raise trend growth by up to a quarter of a per cent a year. If the bundle of policies that are required to raise the retirement age were put in place, and policy is clearly announced, then it is possible that forward-looking consumers will realise they need to save less. This will boost demand, as would the response of forward-looking firms who would realise they needed to scrap less capital.

Conclusion

It is difficult to gauge the scale of the output gap after such a severe recession, but we would judge that it is 4 cent or more. The nature of the recession and the impacts of the preceding financial crisis on risk premia on investment will combine to produce a lower equilibrium capital-output ratio and a higher level of sustainable unemployment. Policy might be able to reduce the latter through labour market programmes, but the former is immutable. The impacts of the rise in risk premia will be felt across the whole economy, and the widespread nature of the shock along with low interest rates and a weak exchange rate have meant there has not been a sharp rise in insolvencies. The increase in oil prices we have seen in the past three months, if sustained, will also reduce the sustainable level of output, with the $25 a barrel increase anticipated for next year reducing trend growth in the UK by 0.1 per cent a year for five years. Coming out of the crisis we cannot expect trend growth to be much greater than 2 per cent a year unless we see a significant increase in the labour force either from later retirement or increased immigration. Given the low level of national wealth we have available to pay for retirement an increase in the length of working lives would be wise. Without this we project trend labour input will rise by around 0.6 to 0.7 per cent a year and trend technical progress will be around 1.4 to 1.6 per cent again over the next decade. It is easier to revise the former than the latter.

REFERENCES

Barrell, R. (2004), 'The design of macro economic policy in Europe', NIESR, mimeo, ESRC Seminar, Bloomberg's, 24 June.

--(2009), 'Long-term scarring from the financial crisis', National Institute Economic Review, 210, October, pp. 36-8.

Barrell, R., Davis, E.P., Karim, D. and Liadze, I. (2010), 'The effects of banking crises on output in OECD countries', National Institute discussion Paper 358.

Barrell, R., FitzGerald, J. and Riley, R. (2010), 'EU enlargement and migration: assessing the macroeconomic impacts', Journal of Common Market Studies, 48, 2, pp. 373-95.

Barrell, R., Gottschalk, S., Kirby, S. and Orazgani, A. (2009), 'Projections of migration inflows under alternative scenarios for the world economy', Department of Communities and Local Government economics paper no. 3.

Barrell, R., Holland, D. and Karim, D. (2010), 'Tighter financial regulation and its impact on global growth', National Institute Economic Review, 213, July, pp. F39-44.

Barrell, R., Holland, D. and Kirby, S. (2010), 'Retirement and economic recovery', National Institute Economic Review, 213, July, pp. F4-8.

Barrell, R., Holland, D. and Liadze, I. (2010), 'Accounting for UK economic performance 1973-2009', National Institute Discussion Paper no. 359.

Barrell, R. and Kirby, S. (2008), 'The budgetary implications of global shocks to cycles and trends in output', ESRI Budget Outlook, October, Dublin.

Barrell, R., Kirby, S. and Orazgani, A. (2011), 'The macroeconomic impact from extending working lives', Department for Work and Pensions Economics Paper no. 95.

Barrell, R. and Sefton, J. (1995), 'Output gaps: some evidence for the UK, France and Germany, National Institute Economic Review, 151, February.

Bell, D.N.F. and Blanchflower, D.G. (2010), 'UK unemployment in the Great Recession', National Institute Economic Review, 214, October, pp. R3-25.

Cecchetti, S.G., Kohler, M. and Upper, C. (2009), 'Financial crises and economic activity', NBER Working Paper no. 15379.

Cerra, V. and Saxena, S.C. (2008), 'Growth dynamics: the myth of economic recovery', American Economic Review, 98:1, pp. 439-57.

Christiano, LJ. and FitzGerald, T.J. (2003), 'The band pass filter', International Economic Review, 44, 2, pp. 435-65.

Dicks, M. (2010), 'The effect of the financial crisis on the UK's productive capacity: surveying the damage', presentation to 2010 Green Budget launch, Institute for Fiscal Studies.

Haskell, J., Clayton, T., Goodridge, P., Pesole, A., Barnett, D., Chamberlin, G., Jones, R., Khan, K. and Turvey, A. (2009)w, 'Innovation, knowledge spending and productivity growth in the UK', Interim report for NESTA Innovation Index Programme, November.

Haugh, D., Ollivaud, P. and Turner, D. (2009), 'The macroeconomic consequences of banking crises in OECD countries', OECD Economics Department Working Paper no. 683, Paris, OECD.

Holland, D., Kirby, S. and Whitworth, R. (2010), 'An international comparison of employment in recovery', National Institute Economic Review, 214, October, pp. F35-40.

IMF (2010), United Kingdom: Selected Issues Paper, IMF country report 10/337.

OECD (2010), 'Prospects for growth and imbalances beyond the short term', Economic Outlook, 2010/I, Paris, OECD.

Office for National Statistics (2009), 'The labour market and retirement', Pension Trends, London, ONS, http:// www.statistics.gov.u k/downloads/theme_compendia/ pensiontrends/Pension_Trends_ch04.pdf.

Oulton, N. (2010), 'Long-term implications of the ICT revolution: applying the lessons of growth theory and growth accounting', CEP discussion paper no. 1027, November.

NOTES

(1) Barrell, Holland and Liadze (2010) show that the apparent effects on trend output of the secondary banking crisis in the UK in the first half of the 1970s, discussed by Dicks (2010), are explained by the common effects of the coincident oil price rise.

(2) It is difficult to discern a significant long-run impact of the 1991 crisis on trend output in Sweden as it was associated with reforms to the structure of the economy that raised productivity. This appears to be uncommon. In only one case was there (insignificant) evidence that a crisis raised trend output, and even here in Denmark it would be unwise to suggest that the forces of creative destruction were at work.

(3) The IMF report does not advocate the use of such filters, only uses it to illustrate the range of output gap estimates.

(4) Our estimate of the scar is marginally higher than that of the

OECD (2010), in part because we judge the impact of the observed (and the implicit) risk premium to be higher.

(5) The rise in oil prices will reduce trend growth by 0.1 per cent a year for the next five years. Barrell and Kirby (2008) discussed the effects of oil and risk premia on trend output, and came up with similar conclusions. More details can be found in Barrell, Delannoy and Holland in this Review.

(6) Barrell, Holland and Kirby (2010) discussed the impacts of raising the retirement age.

(7) Haskell et al. (2009) and Oulton (2010) both argue that recent technical change has been driven by intangible investment. It is not clear that this will rise in future. Oulton also shows that the use of high tech products is more important than producing them.

(8) Barrell, Gottschalk, Kirby and Orazgani (2009) took this view.

Table 1. Fixed investment and capital

billion [pounds sterling], 2006 prices

 Gross fixed investment (a)

 Business Private General Total
 investment housing (b) government

2005 122.1 67.2 23.7 213.6
2006 127.9 73.9 25.4 227.2
2007 144.0 74.1 27.0 245.1
2008 142.4 56.7 33.6 232.8
2009 115.5 41.5 40.0 197.0
2010 118.7 43.4 40.7 202.8
2011 125.0 47.2 34.3 206.5
2012 131.4 49.8 31.0 212.2
2013 138.5 54.0 29.2 221.7
2014 146.0 59.2 28.7 233.9
2015 153.3 64.7 29.5 247.5

Percentage changes
2000/1999 4.4 -0.9 -4.6 2.7
2001/2000 1.5 0.3 23.2 2.6
2002/2001 1.2 7.6 13.7 3.6
2003/2002 -1.0 0.0 17.4 1.1
2004/2003 1.2 14.0 7.7 5.1
2005/2004 4.5 -4.8 10.9 2.4
2006/2005 4.8 10.0 7.3 6.4
2007/2006 12.5 0.2 6.4 7.8
2008/2007 -1.1 -23.4 24.5 -5.0
2009/2008 -18.9 -26.9 18.9 -15.4
2010/2009 2.7 4.8 1.8 3.0
2011/2010 5.3 8.6 -15.6 1.8
2012/2011 5.1 5.5 -9.7 2.8
2013/2012 5.4 8.4 -5.9 4.4
201412013 5.4 9.6 -1.5 5.5
2015/2014 5.0 9.4 2.5 5.8

 User Corporate Capital stock
 cost profit
 of share of Private Public (c)
 capital (%) GDP (%)

2005 17.1 24.2 2163.9 554.1
2006 16.6 25.0 2219.2 564.1
2007 16.5 25.2 2301.2 577.1
2008 16.1 26.2 2370.7 593.2
2009 14.7 25.0 2401.3 612.2
2010 14.5 23.8 2414.2 632.9
2011 16.5 24.9 2433.5 646.6
2012 17.1 25.5 2458.5 656.7
2013 17.0 26.0 2491.6 664.6
2014 16.9 26.4 2533.9 671.9
2015 16.9 26.8 2585.3 679.6

Percentage changes
2000/1999 3.0 -0.2
2001/2000 2.6 1.3
2002/2001 1.4 6.2
2003/2002 3.3 -1.5
2004/2003 2.3 1.8
2005/2004 2.4 1.9
2006/2005 2.6 1.8
2007/2006 3.7 2.3
2008/2007 3.0 2.8
2009/2008 1.3 3.2
2010/2009 0.5 3.4
2011/2010 0.8 2.2
2012/2011 1.0 1.6
2013/2012 1.3 1.2
201412013 1.7 1.1
2015/2014 2.0 1.1

Notes: (a) Fixed investment figures exclude the effect of the transfer
of BFNL nuclear reactors to central government in 2005Q2. (b) Includes
private sector transfer costs of non-produced assets. (c) Including
public sector non-financial corporations.

Table 2. Productivity and the labour market

Thousands

 Employment ILO Population
 unemploy- Labour of
 Employees Total(a) ment force(b) working
 age

2000 23977 27484 1587 29071 36138
2001 24185 27711 1489 29200 36406
2002 24387 27922 1528 29450 36620
2003 24427 28187 1489 29676 36824
2004 24646 28485 1424 29909 37083
2005 24929 28775 1466 30241 37419
2006 25096 29027 1672 30699 37708
2007 25209 29225 1653 30878 37916
2008 25408 29441 1781 31221 38090
2009 24939 28978 2394 3117 38236
2010 24844 19029 2481 31511 38395
2011 24683 28901 2741 31644 38711
2012 24985 29227 2568 31794 38935
2013 25327 29602 2357 31959 39169
2014 25641 29950 2175 32125 39406
2015 25906 30249 2041 32291 39638

Percentage changes

2000/1999 1.6 1.2 -8.1 0.6 0.6
2001/2000 0.9 0.8 -6.2 0.4 0.7
2002/2001 0.8 0.8 2.6 0.9 0.6
2003/2002 0.2 0.9 -2.6 0.8 0.6
2004/2003 0.9 1.1 -4.3 0.8 0.7
2005/2004 1.2 1.0 2.9 1.1 0.9
2006/2005 0.7 0.9 14.1 1.5 0.8
2007/2006 0.5 0.7 -1.2 0.6 0.6
2008/2007 0.8 0.7 7.7 1.1 0.5
2009/2008 -1.8 -1.6 34.5 0.5 0.4
2010/2009 -0.4 0.2 3.6 0.4 0.4
2011/2010 -0.6 -0.4 10.5 0.4 0.8
2012/2011 1.2 1.1 -6.4 0.5 0.6
2013/2012 1.4 1.3 -8.2 0.5 0.6
2014/2013 1.2 1.2 -7.7 0.5 0.6
2015/2014 1.0 1.0 -6.2 0.5 0.6

 Productivity Unemployment, %
 (2006=100)
 Claimant ILO unem-
 Per hour Manufact- rate ployment
 uring rate

2000 89.4 77.1 3.6 5.5
2001 90.6 79.3 3.2 5.1
2002 92.5 81.5 3.1 5.2
2003 94.7 85.3 3.0 5.0
2004 96.6 91.2 2.7 4.8
2005 97.7 95.4 2.7 4.8
2006 100.0 100.0 3.0 5.4
2007 102.0 102.6 2.7 5.4
2008 101.4 102.4 2.8 5.7
2009 99.5 98.4 4.7 7.6
2010 100.6 104.7 4.6 7.9
2011 102.3 110.7 5.2 8.7
2012 102.8 116.0 4.6 8.1
2013 103.6 120.7 4.0 7.4
2014 104.6 124.8 3.5 6.8
2015 106.0 128.6 3.1 6.3

Percentage changes

2000/1999 3.7 5.5
2001/2000 1.3 2.9
2002/2001 2.1 2.8
2003/2002 2.4 4.6
2004/2003 2.0 6.9
2005/2004 1.1 4.6
2006/2005 2.4 4.9
2007/2006 1.9 2.6
2008/2007 -0.5 -0.3
2009/2008 -1.9 -3.8
2010/2009 1.1 6.3
2011/2010 1.7 5.8
2012/2011 0.4 4.8
2013/2012 0.8 4.0
2014/2013 1.0 3.4
2015/2014 1.3 3.0

Notes: (a) Includes self-employed, government-supported trainees and
unpaid family members. (b) Employment plus ILO unemployment

Table 3. Average effective age of retirement versus the
official age, 2002-7

 Men Women

 Effective Official Effective Official

Canada 63.3 65 61.9 65
Germany 62.1 65 61.0 65
France 58.7 60 59.5 60
Italy 60.8 57 60.8 57
Japan 69.5 63 66.5 61
United Kingdom 63.2 65 61.9 60
United States 64.6 65.83 63.9 65.83

Source: OECD.

Table 4. Accumulation

As a percentage of GDP

 Households Companies General government

 Saving invest- Saving invest- Saving invest-
 ment ment ment

2000 3.2 4.0 9.0 12.5 2.8 1.2
2001 4.2 4.3 8.7 11.9 2.5 1.2
2002 3.3 4.7 12.2 11.1 -0.1 1.3
2003 3.5 4.9 12.9 10.2 -1.3 1.7
2004 2.5 5.4 13.6 9.8 -1.1 1.8
2005 2.7 5.3 12.8 10.1 -1.1 1.7
2006 2.3 5.6 12.0 10.2 -0.2 1.7
2007 1.7 5.8 14.3 10.8 -0.4 1.6
2008 1.3 4.6 15.0 9.8 -1.3 2.2
2009 4.2 3.4 13.9 7.4 -6.3 2.6
2010 3.4 3.4 15.2 8.5 -6.5 2.5
2011 3.3 3.7 13.9 8.6 -4.9 1.9
2012 4.3 3.9 13.2 8.5 -3.9 1.6
2013 4.8 4.1 12.4 8.5 -2.7 1.4
2014 5.2 4.5 11.7 8.5 -1.6 1.3
2015 5.3 4.9 11.0 8.6 -0.8 1.2

 Whole economy Finance Net
 from national
 Saving invest- abroad saving (a)
 ment

2000 15.0 17.7 2.6 3.6
2001 15.4 17.5 2.1 4.1
2002 15.4 17.1 1.7 4.0
2003 15.1 16.7 1.6 4.1
2004 15.0 17.1 2.1 3.8
2005 14.5 17.1 2.6 3.4
2006 14.1 17.5 3.4 3.0
2007 15.6 18.2 2.6 4.6
2008 15.0 16.7 1.6 4.6
2009 11.8 13.5 1.7 1.4
2010 12.2 14.4 2.2 1.0
2011 12.2 14.2 2.0 1.1
2012 13.6 14.0 0.4 2.8
2013 14.5 14.0 -0.5 4.0
2014 15.3 14.3 -1.0 5.0
2015 15.5 14.7 -0.9 5.4

Note: (a) Whole economy investment less finance from abroad and
depreciation.

Table 5. Long-term projections

All figures percentage change unless otherwise stated

 1997 1998 1999 2000 2001

GDP (market prices) 3.3 3.6 3.5 3.9 2.5
Average earnings 3.9 6.3 4.5 5.7 5.1
GDP deflator (market prices) 2.8 2.2 2.1 1.2 2.1
Consumer Prices Index 1.8 1.6 1.3 0.8 1.2
Per capita GDP 3.0 3.3 3.1 3.6 2.1
Whole economy productivity(a) 1.5 3.0 2.7 3.7 1.3
Labour input(b) 1.7 0.8 0.9 0.2 1.0
ILO unemployment rate (%) 7.0 6.3 6.0 5.5 5.1
Current account (% of GDP) -0.1 -0.4 -2.4 -2.6 -2.1
Total managed expenditure
 (% of GDP) 38.8 37.7 36.6 36.6 37.2
Public sector net borrowing
 (% of GDP) 2.0 -0.1 -1.2 -1.8 -1.0
Public sector net debt
 (% of GDP) 42.7 40.4 38.1 33.7 30.9
Effective exchange rate
 (2005=100) 81.9 88.4 91.9 95.4 95.4
Bank Rate (%) 6.5 7.2 5.3 6.0 5.1
3 month interest rates (%) 6.9 7.3 5.4 6.1 5.0
10 year interest rates (%) 7.0 5.5 5.0 5.3 4.9

 2002 2003 2004 2005 2006

GDP (market prices) 2.1 2.8 3.0 2.2 2.8
Average earnings 3.2 4.9 3.8 3.6 4.5
GDP deflator (market prices) 3.1 3.1 2.5 2.0 3.0
Consumer Prices Index 1.3 1.4 1.3 2.1 2.3
Per capita GDP 1.7 2.4 2.5 1.5 2.2
Whole economy productivity(a) 2.1 2.4 2.0 1.1 2.4
Labour input(b) -0.3 0.4 0.9 1.2 0.5
ILO unemployment rate (%) 5.2 5.0 4.8 4.8 5.4
Current account (% of GDP) -1.7 -1.6 -2.1 -2.6 -3.4
Total managed expenditure
 (% of GDP) 38.3 39.6 40.1 41.1 41.0
Public sector net borrowing
 (% of GDP) 1.7 3.2 3.2 3.2 2.4
Public sector net debt
 (% of GDP) 30.4 31.6 33.3 35.3 36.4
Effective exchange rate
 (2005=100) 98.8 96.5 101.6 100.0 100.7
Bank Rate (%) 4.0 3.7 4.4 4.6 4.6
3 month interest rates (%) 4.0 3.7 4.6 4.7 4.8
10 year interest rates (%) 4.9 4.5 4.9 4.4 4.5

 2007 2008 2009 2010 2011

GDP (market prices) 2.7 -0.1 -4.9 1.4 1.5
Average earnings 5.0 1.5 2.5 3.4 3.0
GDP deflator (market prices) 3.0 3.0 1.4 2.9 3.3
Consumer Prices Index 2.3 3.6 2.2 3.3 3.8
Per capita GDP 2.0 -0.7 -5.5 0.9 0.9
Whole economy productivity(a) 1.9 -0.5 -1.9 1.1 1.7
Labour input(b) 0.8 0.4 -2.9 0.4 -0.2
ILO unemployment rate (%) 5.4 5.7 7.6 7.9 8.7
Current account (% of GDP) -2.6 -1.6 -1.7 -2.2 -2.0
Total managed expenditure
 (% of GDP) 40.8 42.6 47.7 47.8 45.7
Public sector net borrowing
 (% of GDP) 2.5 4.6 10.8 10.2 7.9
Public sector net debt
 (% of GDP) 36.7 38.8 48.6 57.0 63.6
Effective exchange rate
 (2005=100) 102.9 90.7 81.2 81.0 82.1
Bank Rate (%) 5.5 4.7 0.6 0.5 0.8
3 month interest rates (%) 6.0 5.5 1.2 0.7 1.0
10 year interest rates (%) 5.0 4.5 3.7 3.6 3.8

 2012 2013 2014 2015 2016-20

GDP (market prices) 1.8 2.4 2.5 2.4 2.5
Average earnings 2.8 3.2 3.3 3.1 3.5
GDP deflator (market prices) 2.2 2.2 2.1 1.9 1.9
Consumer Prices Index 1.8 1.8 2.0 2.0 2.0
Per capita GDP 1.2 1.8 1.9 1.8 1.8
Whole economy productivity(a) 0.4 0.8 1.0 1.3 2.0
Labour input(b) 1.5 1.7 1.5 1.0 0.4
ILO unemployment rate (%) 8.1 7.4 6.8 6.3 6.1
Current account (% of GDP) -0.4 0.5 1.0 0.9 0.1
Total managed expenditure
 (% of GDP) 44.5 43.2 42.1 41.5 41.1
Public sector net borrowing
 (% of GDP) 6.6 5.2 3.9 3.1 2.2
Public sector net debt
 (% of GDP) 68.7 71.8 73.5 74.1 72.3
Effective exchange rate
 (2005=100) 82.5 82.7 82.8 83.1 84.3
Bank Rate (%) 1.8 2.7 3.3 3.6 4.5
3 month interest rates (%) 1.9 2.8 3.3 3.7 4.6
10 year interest rates (%) 4.1 4.4 4.5 4.7 4.9

Notes: (a) Per hour. (b) Total hours worked.

Table 6. Employment levels by country of birth

 of which:
 Total Non-UK
 employment born Old EU 14 New EU 8 US and
 Australasia

1997 26695 2013 571 34 151
1998 27017 2140 626 33 165
1999 27385 2120 606 34 172
2000 27611 2270 639 35 196
2001 27855 2359 618 43 189
2002 28136 2533 639 52 201
2003 28321 2561 621 61 200
2004 28675 2811 628 118 230
2005 28850 3008 634 221 218
2006 29151 3317 630 356 239
2007 29473 3603 670 486 231
2008 29412 3818 60 485 204
2009 28986 3720 660 481 239
2008 Q1 29414 3676 697 509 236
2008 Q2 29444 3707 696 506 233
2008 Q3 29499 3722 661 516 209
2008 Q4 29412 3818 686 485 204
2009 Q1 29095 3804 676 520 204
2009 Q2 28832 3732 670 522 216
2009 Q3 29000 3682 667 502 226
2009 Q4 28986 3720 660 481 239
2010 Q1 28745 3701 655 508 239
2010 Q2 28933 3846 680 560 236
2010 Q3 29297 3886 653 593 219

 of which:

 Africa Indian
 sub continent

1997 364 340
1998 418 347
1999 409 361
2000 467 366
2001 466 379
2002 523 407
2003 554 414
2004 599 435
2005 637 482
2006 683 535
2007 703 547
2008 756 620
2009 686 612
2008 Q1 707 561
2008 Q2 733 568
2008 Q3 741 607
2008 Q4 756 620
2009 Q1 738 608
2009 Q2 702 599
2009 Q3 686 578
2009 Q4 686 612
2010 Q1 698 590
2010 Q2 708 631
2010 Q3 748 629

Source: Labour force statistics.