Economic performance in France, Germany and the United Kingdom: 1997-2002.
We assess the performance of France, Germany and the United Kingdom over the period 1997-2002. Gross and net output per hour worked are considerably lower in the UK than in France and Germany. GDP in France and the UK have grown at the same rates over the period although real national income
I. Introduction
The past few years are generally regarded as something of a disappointment for Germany, which has experienced periods of economic stagnation while the United Kingdom and France have grown at close to their underlying trend rate. Some estimates suggest that the United Kingdom has closed the economic gap with Germany which opened up in the 1960s. Because rates of change and levels are confused in the popular mind, closure of the gap is often assumed to mean that the UK economy is now performing better than its neigbbours rather than that it is no longer performing worse.
This article reviews the underlying supply and demand factors which lie behind the performance of France, Germany and the United Kingdom. We look first at labour supply and then at capital investment with technical progress (growth of total factor productivity') a residual. We confirm that labour market performance is a key factor holding back the French and German economies.
2. Background: what is the current relative position?
In table 1 we show the economic performance of the three countries as estimated by Barnes (2004) and calculated on the basis of the OECD estimates of purchasing power parity. It should be stressed that the data on GDP/hour worked are regarded as less reliable than those on GDP per worker. In addition to Barnes' figures we show estimates of net product per worker and per hour worked. These are calculated by applying the ratio of net to gross domestic product in the year in question to Barnes' figures and then rebasing so that the UK estimates remain 100. These figures are, for the reasons given in Box A, relevant if we want to assess the capacity of the production process in each country to contribute to economic welfare.
The United Kingdom appears in a better relative position in the net figures than in the gross figures because the share of depreciation in UK GDP is lower than that in the other countries. However, there appears to be an inconsistency between depreciation patterns in UK national accounts and figures produced by OECD on an internationally comparable basis. This is discussed further in section 5. In terms of output per hour worked, both France and Germany remain substantially ahead of the United Kingdom.
There is one further point to make about Germany's performance as shown by these data. Although the performance of the German economy has undoubtedly been affected by re-unification, one might have expected that an initial worsening of the productivity position, as efficient plants in West Germany were averaged with inefficient plants in the East, would have been gradually unwound as plant and working methods in the East were modernised. In fact, over the period shown, these data suggest that the United Kingdom has been steadily reducing the productivity gap with Germany, while the gap with France has not changed much.
There are, nevertheless, substantial concerns about these data. The purchasing power parity data show, over the period 1997-2002, the UK's purchasing power parity rising at 0.2 per cent per annum relative to the United States, which provides the reference point. Purchasing power parity for both France and Germany is declining relative to the United States by just over 0.4 per cent per annum. This means that, averaged over the period, the United Kingdom's GDP and its measures per worker and per hour worked are rising at 0.6 per cent per annum relative to France and Germany even when the volume of output measured by national data is constant. In other words, if the national accounts of all three countries showed, in national prices, productivity being constant, the United Kingdom's output relative to that of the other two countries, measured in purchasing power parity, would have grown at this rate.
The pattern is not, however, monotonic. Between 1998 and 1999 the UK's reported purchasing power parity rose by about 2 percentage points relative to Germany and the movements against France are not much smaller. On the face of it, these movements are surprising since three advanced economies are being compared, none of which suffered from rapid inflation at the time and none of which faced any substantial idiosyncratic shock.
Since the publication of the data by Barnes, the OECD has published its own comparative estimates of GDP per worker and GDP per hour. The figures in table 2 show the United Kingdom in its traditional position behind France and Germany although they confirm that the gap between the United Kingdom and Germany is narrower than that with France. However, even on these data, it is clear that the United Kingdom's performance relative to Germany is considerably better when looked at in net output terms than when seen on the traditional gross output basis.
The reasons for the differences between these two data sets are not clear; the OECD figures are reported as calculated at 2002 purchasing power parities and one might therefore expect them to match Barnes' figures for 2002; the question whether the 2002 purchasing power parity figures are plausible should not affect the issue. If nothing else, they demonstrate the complexity of comparing productivity positions in different countries. But whichever set of data is closer to the truth, the discrepancy between figures for GDP/NDP per worker and per hour worked means that, to enjoy a living standard similar to that on the Continent, people in the United Kingdom have to give up much more leisure time.
We now move on to assess how inputs and outputs have changed between 1997 and 2002 in the three countries of interest.
3. Labour supply
We present here labour supply indicators from Eurostat, the International Labour Organisation and the OECD to provide a picture of labour market conditions in the three countries. The United Kingdom performs better than France and Germany in terms of all indicators. They are shown in table 3.
We can see that female participation rates have improved in all countries, with the biggest improvement from the lowest level in France. France has also shown the strongest improvement in male participation, while in Germany male participation rates have fallen since the late 1990s. Although participation rates of older workers in France are markedly lower than in Germany or the UK, there has been a very substantial improvement in this respect since 1997. The United Kingdom has also improved the labour force participation of workers older than 55, while in Germany participation by men in this age group has declined. Unemployment has declined in all countries, with the decline greatest in France but with the United Kingdom showing the lowest unemployment rate.
The data for hours worked need to be interpreted with some caution. The OECD warns that they are more suited to illustrating changes over time than to making cross-country comparisons. Nevertheless, they suggest that, not only are participation rates higher in the United Kingdom than elsewhere but also that, for people in employment, average working hours are also longer.
Between 1997 and 2002 average hours worked in all three countries display a downward trend, although the trend is less pronounced in the United Kingdom than elsewhere. Looking at hours worked per person of working age, we can see that, in France and Germany, these rose during the period of rapid growth at the end of the century, but have since fallen back, particularly in France. The movements in the UK have been very small. The more striking finding is the evidence that the average person of working age in the UK works over 25 per cent more than in France or Germany.
The sharp decline in France in hours per person employed is probably due to the legislation to reduce the working week from 39 to 35 hours. This came into force on 1 January 2000 for businesses employing 20 or more people and two years later for smaller businesses. Germany has managed a smaller but still substantial decline in hours worked without any such explicit intervention, while in the United Kingdom hours worked have been reasonably stable in comparison.
One of the factors serving to reduce average hours worked is the continuing rise in female relative to male participation. On average women work fewer hours than their male counterparts. Possibly this has a smaller dampening effect on hours worked in France than in the United Kingdom. ILO data suggest that in France, measured on a weekly basis, women work only 5 per cent fewer hours than men. In the United Kingdom women work 10 per cent fewer hours than men.
We now move on to the question of how the labour input has changed in each country between 1997 and 2002. We explore this largely with data drawn from the OECD. Combining data on hours worked and employment with information on changes in the labour force and the population of working age, we can identify the factors behind the change in actual labour supply in each country. We weight together changes in male and female participation because the data do not allow us to identify changes in annual hours worked for each sex. These calculations are shown in table 4.
The figures for France produce a picture consistent with the OECD's own recently published estimates of labour input growth. In the United Kingdom there are differences between vintages of data; those used by the OECD in its calculations have not yet been published, leading to an unidentified component. In Germany the component parts of labour input change are collected from disparate sources and the substantial discrepancy captured by the unidentified component is, unfortunately, not surprising.
The sharp reduction in the rate of unemployment (and equivalent rise in the rate of employment) in France was more than offset by the decline in average hours worked, such that on average between 1997 and 2002 the labour input rose by only a third of the rate of increase in the labour force. In Germany the decline in hours worked between 1997 and 2002 was weaker than in France, but so was the decline in unemployment. Overall, UK labour input rose markedly, French labour input rose somewhat but German labour input was static.
Aside from changes in population, the share of the population in work is one of the key determinants of the labour input and in turn economic growth. Growth in this ratio is given by the sum of the percentage changes in the participation and employment rate in the table above. Between 1997 and 2002 these have improved more quickly in France and Germany than in the UK. France in particular has seen significant improvements in the ratio of employment to the population of working age. But, some of this gain in France has come at the expense of (and possibly as a result of) the decline in hours worked, such that the effect of rising employment on the labour input and economic growth has been offset somewhat.
Both Germany and France have high unemployment in comparison to the UK. The 3 percentage point reduction in the French unemployment rate between 1997 and 2002 narrowed the gap between unemployment in France and the UK from almost 5 percentage points in 1997 to fewer than 4 percentage points in 2002. However, there is some reason for believing that the fall in unemployment in France is largely cyclical; over the period output moved from below trend to above trend. It is not obvious that France has begun to remedy the underlying structural problems typically put forth as the cause of its relatively high rate of unemployment. In comparison to the United Kingdom, the French and German economies are characterised by high labour taxes, replacement ratios and union coverage. They are all factors which serve to raise unemployment (see e.g. Nickell, 1998; Nickell and van Ours, 2000); we explore the first of these in section 4. In the United Kingdom these institutions underwent significant reform during the 1980s, helping to reduce equilibrium unemployment throughout the 1990s. The current government's ambitious Welfare-to-Work programme, gradually implemented since 1997, is likely to have helped in reducing equilibrium unemployment further.
An older population further exacerbates Germany's weak performance in comparison to the UK and France. In comparison to the population as a whole, older people display low participation rates. Splitting the population into age bands 20-24, 25-34, 35-49 and 50+ we calculate that labour market participation would be almost 2 percentage points higher in Germany if it had the same demographic structure as either France or the United Kingdom. The age compositions of the population in France and the United Kingdom are very similar, although the distribution of participation rates across age groups is very different. Young and old people in France display comparatively low rates of labour market participation. The reverse is the case for people age 25-49.
Between 1997 and 2000, the demographic composition has changed in all three countries in a way that is unfavourable to the participation rate. In all three countries the number of people aged over 50 increased between 1997 and 2000, measured as a proportion of the population aged 20+. In the UK and Germany the share of the population aged over 50 has risen by approximately 1 percentage point over this three-year period. The equivalent figure for France is 1.4 percentage points. Assuming age-specific participation rates are unchanged at 1997 levels, we can calculate the effect of the ageing of the population on participation. We split the population into age bands 20-24, 25-34, 35-49 and 50+ as above. This simple calculation suggests that the change in the age distribution of the population has reduced participation for those age 20 and above by 3/4 of a per cent in France, by 1/2 a per cent in Germany and by a little less than 1/2 a per cent in the United Kingdom between 1997 and 2000.
While the problems of Germany are those of a low level of employment and static labour input, France also suffers from the first of these. It has managed to grow at the same rate as the United Kingdom despite slow growth of total labour input. Plainly, if it were able to increase its hours worked, it would be in a substantially better position than the United Kingdom.
We now focus on the role of employment taxes as an obstacle to employment because the variation in these is large and the data are both up to date and for a period long enough to make a panel analysis possible. Since there are undoubtedly also other variables involved, our results should be interpreted more as a proxy for the benefits of labour market reform than as an indication of the effects of employment taxes on their own.
4. Tax and benefits as influences on employment in the European Union
There are good theoretical grounds for expecting tax structures to have an influence on labour participation and unemployment rates. Unemployment benefit schemes often imply very high marginal tax rates on people who move from unemployment to employment. If the move from unemployment to employment results in relatively little extra income, then unemployment is likely to be high. The effect is, of course, augmented if people drawing unemployment benefits are in fact working in the black economy. In such circumstances they may find themselves financially worse off by moving back to work. Thus the effective supply of labour may be influenced by benefit arrangements.
On the demand side similar effects may apply. Tax on labour income drives a wedge between the cost of labour to the employer and the benefit to the employee. Tax on labour income may be charged either to the employee (as income tax and national insurance contributions) or to the employer as national insurance and social security contributions. Since the tax structure is the same for all employers, it cannot logically make any difference to the overall outcome whether the tax is collected by the employer or the employee; the effect on labour costs gross of taxes must be the same, however the tax burden appears to be split between employer and employee.
Data for the effective labour income tax rate on low earners in EU countries are available for the period 1996-2002. The data for 1997 and 2002 are shown in table 5. Those on the marginal tax rate on moving from unemployment to employment cover only 2001 and 2002 and we show the figures for 2002. (4)
>From these data we can see that employment rates rose in all three countries between 1997 and 2002. The tax wedges were reduced over the same period and both changes were particularly large in France. We therefore wish to investigate how far these tax reductions account for the growth in the employment ratios over the period and also, as part of the same exercise, identify the extent to which France and Germany could raise their employment ratios by cutting their tax wedges to those in the UK. There is a risk in this analysis that a part of the role attributed to the tax wedge may in fact be due to other differences for which there are no comparable data. The marginal tax rate may also be important, but with data for this only for 2001 and 2002, it is impossible to say whether this is the case.
We explore how far it is possible to explain first unemployment rates and then employment ratios for the period 1996-2002 in terms of the tax wedges and also in terms of the state of demand as represented by the output gap in each of the European Union countries excluding Luxemburg. (5) We carry out a pooled regression looking at the coefficients in three types of models. The first is simply a pooling of the data. The second introduces country dummies to deal with country-specific fixed effects and the third also introduces time dummies. The results of the regressions are shown in table 7.
As is so often the case, the results are sensitive to the method of analysis. When country dummies are present, so that only variation in unemployment rates within the sample period is to be explained, then movements in the tax wedge play a powerful and significant effect. When the year dummies are also present, the tax wedge effect remains significant but its size is reduced.
Employment rates are calculated as the proportion of the working age population employed or self-employed. Thus, unlike unemployment rates, they consolidate all people in the age group who do not work, whether or not they are looking for work.
The effects on employment rates are, after controlling for country and year effects, very similar to those on the unemployment rate (with signs reversed because we are now looking at employment rather than unemployment). The coefficients for men and women are, once the fixed effects are removed, very similar and we also show the results with the coefficients for men and women restricted to be equal; this restriction is easily accepted. It is wrong to pretend that these equations offer evidence independent of the earlier regression on unemployment about the power of the tax wedge to discourage employment. But they plainly serve to strengthen the earlier conclusions.
These estimates suggest that a 1 percentage point reduction in the effective rate of taxation reduces the unemployment rate by 0.33 percentage points. We find results similar in magnitude for the effect of the tax wedge on the employment rate, but with opposite signs. Pooling the male and female sub-samples we estimate that a 1 percentage point reduction in the effective rate of taxation raises the employment rate by 0.34 percentage points. These numbers imply that, between 1997 and 2002, the fall in the tax wedge reduced the unemployment rate by just over 1/2 a percentage point in Germany and by approximately 1 1/4 percentage points in France and the UK. Using the coefficient on the tax wedge we show, in table 8, the estimated effect of the change in the tax wedge on the employment rate and also the change in employment which might be expected in France and Germany if they moved to the UK tax wedge.
These estimates suggest that, while tax wedge reductions, possibly as proxies for wider labour market reforms, were valuable contributors to increased employment, and particularly so in France and the United Kingdom, nevertheless they account for a relatively small proportion of the overall change except in the United Kingdom. However, by moving to the UK tax wedge on low paid workers, France would halve its employment gap relative to the UK and Germany would achieve a slightly higher employment rate than the UK. This does not, of course, mean that such a change is costless. Tax revenues are needed for public spending and judgement is always needed when trading off the benefits of lower taxes against losing the benefits of some public spending.
Others have found a similar impact of the tax wedge on unemployment. For example, Daveri and Tabellini (2000) estimate that, between 1965 and 1995, the 14 percentage point rise in labour tax rates in the EU may have accounted for a 4 percentage point rise in the EU unemployntent rate. In other words, to facilitate comparison to our numbers, a 1 percentage point reduction in the tax wedge reduces unemployment by around 0.29 percentage points. Nickell (1998) finds a similar relationship between the tax wedge and unemployment. Based on a sample of twenty OECD countries observed between 1983 and 1994, his estimates suggest that a 1 percentage point reduction in the tax wedge results in a 0.27 percentage point reduction in the rate of unemployment, when unemployment is around 10 per cent. The effect is smaller when unemployment is smaller. His measure of the wedge includes excise taxes, as well as payroll and income taxes.
The magnitude of our estimates is nevertheless at the larger end of the range provided in the literature. Surveying the literature, Nickell and van Ours (2000) suggest that a 1 percentage point fall in the tax wedge is associated with a 0.12 percentage point fall in the rate of unemployment. This estimate is obtained by averaging across the results published from a number of crosscountry studies.
It is perhaps not very surprising that our coefficient is at the upper end of the range. While we have removed country and year-specific effects by means of dummy variables, it is clear from table 5 that high tax rates on low-paid workers are associated with high marginal tax rates when unemployed workers take up low-paid work. They may also be associated with institutional obstacles to labour market flexibility. Thus our coefficients could be interpreted as indicating what might be achieved by a general package of labour market reforms rather than suggesting what reduced taxes on low-paid workers could achieve on their own.
However, there is some reason to believe that our figures may represent quite reasonable estimates of the effect of recent declines in taxation on unemployment for France, if perhaps less so for Germany and the UK. Belot and van Ours (2000) suggest that the relationship between taxation and unemployment depends on the other institutional factors in place. They find that the effect of labour taxes on unemployment is greater the larger the benefit replacement rate, union coverage and the degree of coordination between the actors involved in the collective wage bargain, and the lower the degree of union coverage. These characteristics are all features of the French labour market.
5. Capital input
Official statistics provide estimates of the net capital stock in the three countries we are considering, although the figures exclude non-produced capital such as land. We begin with a description of the conditions in which capital is used in the three countries, looking at depreciation rates, net rates of return, capital-output ratios and the growth rate of the net capital stock.
Rates of return and depreciation
The national accounts for the three countries and their capital stock data provide information needed to work out rates of return and depreciation.
We calculate depreciation rates by dividing depreciation (of each of the two types of capital) at current prices by the market value of the net capital stock at the start of the year. The net rate of return is calculated by dividing operating surplus net of depreciation by the same figure. We have to take account of mixed incomes, the incomes of the self-employed, which include wages, as well as business profits. There are two obvious ways of allocating mixed incomes between wages and profit. It might seem sensible to assume that the wage per self-employed person is the same as that of an employed person. This, however, neglects the fact that some people may be both self-employed and employed. Instead we assume that two thirds of mixed income is wage income, with the balance being profits gross of depreciation.
We also show capital-output ratios, measured as the value of the net capital stock at the start of the year divided by GDP at market prices. All ratios are measured as proportions of capital stocks at the start of the year. As table 9 shows, returns have been considerably higher in the UK and France than in Germany. The figures for both rates of return and capital-output ratios are calculated neglecting the value of land in the national capital stock. The resulting upward bias in the rate of return is offset by a downward bias arising from the fact that no return is imputed to publicly-owned capital used for consumption purposes. (6) We present averages for the period 1998-2002 because these are relevant to economic growth between 1997 and 2002.
The rate of return on capital in the UK has declined over the period (from 8.2 per cent per annum to 6.7 per cent per annum) while the other returns are very stable. The depreciation rates on housing are very similar in all three economies while those in France and Britain on other capital are higher than those in Germany. There is an obvious question, whether or how far this accounts for the differences in the capital-output ratios in the three different countries. Comparing the United Kingdom and Germany, we see that the whole economy depreciation rates are very similar; Germany's reported high capitalisation must be due to its historically high investment rate rather than slow writing down of capital. France, however, has a higher depreciation rate than Germany, and there is a question how far this makes France appear less well capitalised than Germany. (7)
A high assumed depreciation rate has the effect of reducing profits after depreciation, but it also reduces the estimated capital stock. The former depresses the estimated rate of return while the latter raises it. On average the gross share of investment in GDP is equal to the capital-output ratio multiplied by the growth rate plus the depreciation rate
I/Y = (g + [delta])K/Y
>From this we can work out the effect of variations in the depreciation rate, [delta], on the derived capital-output ratio (K/Y) and on the share of depreciation in gross output ([delta] K/Y) for a given ratio of gross investment to GDP.
K/Y = 1/g + [delta] I/Y and [delta]K/Y = [delta]/g + [delta] I/Y
To use these formulae we need to look at the investment ratio over a reasonably long period; we use the average for 1987-2002 shown in the table. Applying these formulae, assuming a trend growth rate of 2 1/4 per cent per annum and using the average investment ratios for 1987-2002 shown in table 13 and average whole-economy depreciation rates for 1998-2002, we derive capital-output ratios of 3.11 for Germany, 2.65 for France and 2.48 for the United Kingdom.
Thus the model gives a reasonable guide to the observed capital-output ratios. If France used the mean of German and UK depreciation rates, then its capital-output ratio would rise by 0.19 points. This would reduce the estimate of the whole economy rate of return for the period from 6.64 per cent per annum to 6.19 per cent per annum, leaving it still well above the German rate. Thus the low rate of return on capital in Germany cannot be explained by different depreciation assumptions.
Low returns in Germany: cyclical or structural
It is widely recognised that profits are pro-cyclical (Hultgren, 1965, Rotemberg and Woodford, 1999) and the question therefore arises whether the low profit rates in Germany in particular could be ascribed to cyclical factors. There are in fact a number of reasons for thinking that this is not the case. The rates of return summarised in table 9 for Germany have been very stable, changing little between 2000, which was a year of rapid economic growth, and 2002 which was a year of weak economic growth. A cyclical explanation of low returns would be expected to show a considerable variation between the two. Secondly, if we look at the ratios of net capital to GDP in current prices, cyclically weak output growth should be expected to lead to rising capital-GDP ratios, with the capital stock designed to reflect cyclically normal levels of output. There is no evidence of a cyclical rise in the capital stock.
Further light on this can be shed by exploring the differences in rates of return which would be attributable to the differences in capital-output ratios if the marginal product is measured by the rate of return. There is evidence that the production structure can be represented reasonably well by a CES production function of the form
Y = A(t)[{[mu][K.sup.-[rho]] +(1 - [mu])[L.sup.-[rho]]}.sup.-1/[rho]]
where Y is output gross of depreciation, K is net capital input and L is labour input. A(t) represents total factor productivity as a function of time. If capital is paid its marginal cost, then the rate of return gross of depreciation is given by the expression
r + [delta] = [mu]A[(t).sup.-[rho]][(K/Y).sup.-[rho]-1]
We can use this to explore the link between the gross rates of return and the capital-output ratios in the different countries. We find, comparing countries i and j that
[r.sub.i] + [[delta].sub.i]/[r.sub.j] + [[delta].sub.j] = [{[K.sub.i]/[Y.sub.i] [Y.sub.j]/[K.sub.j]}.sup.-[rho]-1],
assuming that the production functions are the same in both countries.
With this formula, taking France as the reference country, we calculate the rates of return (fitted rates) which would be found in the UK and Germany, on the assumption that the values of [mu], [rho] and A(t) are the same in all three countries. We assume a value of 2/3 consistent with median elasticity of substitution between labour and capital of 0.6 provided by Rowthorn (1999, p. 415).
Thus we can see that, taking France as a reference point, the lower gross return in Germany is largely commensurate with the higher capital-output ratio. However, the return on capital in the United Kingdom is well below what would be expected given the low capital-output ratio.
This analysis does not, of course, indicate what causes the differences in capital-output ratios in the different countries. Tax rates on profits seem to be higher in the UK than in France and higher there than in Germany. If rates of return net of tax are equalised internationally, then high taxes will lead to high gross rates of return and correspondingly low capital stocks. (8) Compared with Germany, a part of the high return in France may be explained in this way. But it does not seem possible to account for the low return on capital in the UK by this mechanism.
The calculation does, however, make it difficult to argue that returns in Germany are low because of the cyclical state of the economy. If that were the case we would expect only a part of the gap between returns in France and Germany to be accounted for by the variation in the capital-output ratio. It may instead be that risk premia are lower there than elsewhere but that is outside the scope of this paper. This illustration of the UK's productivity problem can also be viewed from a more 'conventional' perspective. Tables 1 and 2 pointed to output per hour worked being around 20 per cent higher in Germany than in the UK. Germany's capital-output ratio is 30 per cent higher than in the UK. With the share of income gross of depreciation going to capital being about 30 per cent of GDP at market prices and 33 per cent at basic prices, only about half of the gap can be explained by Germany's capital intensity. The balance is the UK's productivity problem.
The contribution of capital to output growth
With this background we can calculate the contribution to growth in output arising from fixed investment. As we explained in Box A, we present results on two assumptions. First we assume that the contribution of investment is measured by growth in the net capital stock, and secondly that it is measured by growth in capital services. In Box 1 it was explained how this allows us to identify different contributions made by depreciation in the two cases. Obviously, as we show later, the results depend on whether GDP is measured at basic or market prices. Here in table 12 we provide details of the calculations only for GDP at basic prices. This is consistent with most other supply analysis of economic growth. The figures for growth in capital services are taken directly from OECD data. We note that, although the United Kingdom has a share of net investment in GDP only slightly higher than in France and Germany, the rate of growth of capital services, both ICT and other, is substantially higher. This is not impossible but the magnitude of the gap is surprising.
Box A. GDPM, GDPB and Real Net National Income (1) GDP at basic prices and at market prices GDP in real terms is, for most purposes, now presented at market prices (GDPM) and is calculated using a chain-linked index approximation to an optimal Divisa index. Headline economic growth is presented in this way and it provides a natural framework to look at the contribution of various components of demand to overall growth. Growth accounting, however, explains the growth in real GDP at basic prices (GDPB) which correspond approximately to those prices that would be observed before adding on indirect taxes. In this paper we seek to show how demand and supply factors have contributed to GDP growth. Given obvious benefits in having the same measure of GDP on the demand and supply sides, we therefore need to modify conventional growth accounting by showing, as an additional contribution to growth, the effects of changes in the adjustment to basic prices. We calculate this by taking the percentage increase in the adjustment to basic prices measured in constant prices and multiply it by the share of the basic price adjustment in GDP at current market prices. In a smoothly evolving economy, if this adjustment grows in line with GDPM, so that both GDPM and GDPB grow at the same rates, it then follows that the contributions of the factors of production, labour and capital and of residual total factor productivity will be depressed by the ratio basic price adjustment/GDPM. While this allows us to balance demand and supply and may also shed some light on why one country has out-performed another, it is not the standard means of presenting growth accounting results. We therefore also present our growth accounting results as contributions to the growth of GDPB so as to make them immediately comparable with other findings. The contribution of capital accumulation to net output and depreciation When assessing growth in either measure of GDP it is of considerable interest to know how far growth is accounted for by an addition to the depreciation charge, since this cannot be expected to play any role in increasing living standards. The importance of the issue is further augmented because it is argued that the contribution made by capital to GDP growth should be assessed in terms of the additional flow of capital services rather than the increase in the capital stock (Groth, Gutierrez-Domenech and Srinivasan, 2004). The two differ if the depreciation rate of new capital is different from that of the existing capital stock. Such a situation is obviously likely to arise if there is substantial investment in ICT (information and communication technology) capital, since the latter almost certainly depreciates much faster than, say, housing. We proceed as follows in our analysis. We present estimates of the contribution of capital on two bases--first using the growth of the capital stock given in national sources as the measure of change in capital input and secondly using OECD (2) estimates for the increase in the supply of capital services. In both cases we use the percentage increase in the net capital stock, (3) multiplied by the share of profits net of depreciation in GDPB or GDPM as appropriate to estimate the contribution of capital accumulation to net output. To calculate the contribution of depreciation on the first basis we take the share of depreciation in GDPM or GDPB and multiply it by the increase in the net capital stock to obtain the figure we require. For the second calculation we take the percentage increase in the flow of capital services, multiplied by the share of gross profits in GDPB or GDPM to measure the contribution of capital accumulation to gross output. The difference between this and the net contribution of capital represents the contribution to depreciation of the growth in capital services calculated on the second basis. Underlying this is the assumption that the net rate of return is the same on all types of capital, i.e. that the capital market is efficient. Both methods of calculation are of some importance because, as we demonstrate, the OECD data for growth in capital services have properties which raise questions about their coherence with other economic data. GDP and Real Net National Disposable Income Analysis of neither GDPM nor GDPB is, however, adequate if one wishes to understand what is happening to living standards over time, or to compare living standards in different countries. For these purposes it is more sensible to focus oncomparisons of real net national income at market prices. There are a number of important differences between gross domestic product and real net national disposable income at market prices. First of all, as is obvious from the two titles, the first is gross of depreciation while the second is net of depreciation. The logic of working with a gross measure is obvious; the measurement of depreciation is never likely to appear very precise. Output is a gross observed measure and depreciation measures the extent to which capital is used up in the production process. However it does not make much sense to compare two countries or periods on a gross basis if they have very different capital intensities of production and thus very different depreciation charges against gross output. Secondly, it is necessary to take account of income from abroad, whether accruing as factor income or paid as transfers. A country which borrows heavily from abroad in order to finance its capital stock will face payments of factor income abroad which a country reliant on its own saving in order to provide its capital stock does not have to make. Thirdly, when studying how changes take place over time, it is necessary to take account of variations in the terms of trade. A country whose import prices fall may not experience any change to real GDP but it does enjoy an increase in its real income. The System of National Accounts (UN, 1993) shows an awareness of this terms of trade effect but it is not clear that an economic meaning can be given to the methods proposed to identify it. Here, instead, we adopt a means of measuring the terms of trade effect which has a clear meaning in terms of future consumption possibilities. Finally, there is the question of changes in the prices of investment goods relative to consumption goods. The money price of investment goods measures their worth in terms of consumption goods. The (expected) real interest rate measured in terms of a consumption bundle indicates the trade-off between consumption foregone and future consumption possibilities. It follows that, if we want a measure of economic activity which reflects current and future consumption opportunities, then we need to measure net investment in terms of the consumption foregone to pay for it, and not with reference to some notional price index of investment goods. To put the point the other way round, if investment goods are steadily becoming cheaper relative to consumption goods, then a part of the return to capital measured in the national accounts is best seen as compensation for the falling price of capital goods. It is, in effect, like depreciation and not available for net investment or consumption. These considerations point to the fact that in order to produce an income aggregate related to current and future consumption possibilities, net national disposable income should be deflated by the price of consumption goods. Oulton (2002) presents an analysis of US economic growth on this basis and Sefton and Weale (2000) present a full account of this measure of income, showing the relationship between it and current and future consumption. Accordingly, we use the term Real Net Disposable Income, RNDI to indicate real disposable income deflated by the price index which deflates private and government consumption expenditures taken together. The following relationship holds between GDPM and real national disposable income (RNDI); differenced and measured as proportions of initial GDPM, it facilitates understanding of the link between movements in real GDPM and movements in real income (1) GDPM/[P.sub.GDP] = RNDI/[P.sub.C] + Depreciation/[P.sub.C] - Net factor and transfer income/[P.sub.C] + GDPM (I/[P.sub.GDP] - I/[P.sub.C]) The last term in this equation is the terms of trade effect, since it measures the adjustment needed to move from output valued in terms of consumption to output valued in terms of production. It is obvious that it has no meaning except in relative terms. It allows us to measure the impact of terms of trade effects relative to the period in which prices are one. The expression allows us to decompose the relationship between growth in real national income growth and real GDPM growth either in the form shown, identifying the contribution of each right-hand side component to growth in GDPM or, by making real national disposable income the subject and analysing changes to the change in this. We use the first approach so as to align our figures with other decompositions of the growth in GDPM.
It can be seen that the use of the capital services figures shows a larger contribution to depreciation than is obtained from using net capital stock data in all three countries. The effect is particularly marked in the United Kingdom where ICT investment has been a large contributor to growth in the capital stock. An implication of this is that the quality of output growth in the United Kingdom is not as good as in the other countries. The impact of this can be seen by considering the long run that these figures imply. Over the period 1997-2002 basic price output in the United Kingdom grew by 2.6 per cent per annum (table 13). If the contribution from depreciation were maintained at that shown in table 12, then the share of depreciation in GDP at basic prices would be expected to rise from its average share of 12.6 per cent in GDP asymptotically to 1.08/2.6 = 42 per cent. In France the increase projected is much more modest, from 16.0 per cent to 21.4 per cent, and in Germany from 16.6 per cent to 27.8 per cent. The contributions associated with depreciation are lower if national prices rather than harmonised ICT prices are used, but the disparity, particularly for the United Kingdom, remains. Even using national prices, one would expect to see the share of depreciation in UK GDP rising markedly during the period; this is not apparent in the data used to calculate the averages presented in table 11. Either the estimates of the contribution of capital are wrong or the UK depreciation figures provided by the Office for National Statistics are too low. (9) Thus there are good reasons for assessing growth performance in the light of capital contributions calculated from both the capital services and the capital stock data. We also note that the omission of land from capital stock data makes it likely that estimates of growth in net assets or capital services overstate the true growth. This in turn makes it possible that estimates of the contribution of capital are biased upwards.
6. The state of demand
The problems of the large Euro Area economies are often regarded as cyclical, to be alleviated by means of expansionary monetary and perhaps fiscal policies. It is impossible to comment on their recent growth performance without forming some view oil whether this is in fact the case or not. We have noted above the low levels of labour input and take these as our starting point. In France and the United Kingdom hours worked per person of working age are close to trend while in Germany they are 1 per cent below trend. (10)
The shortfall in hours worked in Germany could be attributable to decisions made by individuals in the light of expectations of future wage movements, the intertemporal substitution hypothesis. Until recently it had been thought there was little evidence to support this but Lee (2001) has argued that careful analysis of micro-economic data does support the hypothesis. For the shortfall in hours worked in Germany to be explained in this way, it would be necessary for there to be an expectation of an unusually rapid increase in hourly wages as would be associated if actual wages were below their trend value. In fact wages are above trend in both France and Germany. The hypothesis of intertemporal substitution cannot account for the low labour input in either country. We explored in section 4 an alternative structural view that the employment taxation plays an important role.
>From the demand side, there is the question of whether low levels of consumption are depressing the French and German economies. In 2002 consumption was above trend in both France and Germany, while long-term real interest rates were more or less on trend. By 2003 consumption in Germany was 1/2 per cent below trend; in France it was above trend by a similar amount. UK consumption was almost on trend. Long-term real interest rates were below trend in all countries (by 0.6 percentage points in France, 0.4 percentage points in Germany and 0.3 percentage points in the United Kingdom). The standard intertemporal optimisation model of consumer behaviour suggests (Svensson, 2000) that the percentage deviation of consumption from trend should equal minus the integral of the deviation of the long-term real interest rate from trend from now to the indefinite future multiplied by the intertemporal elasticity of substitution believed to be between a half and one fifth; see Hall (1988). Given that long-term interest rates are below trend and not expected to rise above it, one would expect consumption to be above rather than below trend. So it is difficult to conclude that consumption is depressed as a result of an excessively restrictive monetary policy. Nevertheless, it is almost certainly the case that consumption demand has grown more strongly in the United Kingdom than in France and Germany because of greater house price buoyancy there and the interaction of this with consumer demand (Barrell, Metz and Riley, 2004). This linkage is not reflected in the standard model of intertemporal optimisation. Rapid expansion of public sector consumption has also been a factor behind the overall growth in consumption in the United Kingdom.
The general impression which results from these observations is that the poor economic performance in Germany is the result of medium-term factors rather than cyclical disturbances. Faster growth in demand might have raised labour input by bringing hours worked per person of working age back to trend. However, over five years this would have increased the rate of growth of labour input by only 0.2 percentage points per annum. It is not clear that the economy has the supply capacity to meet a rapid expansion of demand because there does not seem to be the extra labour needed to facilitate a sharp reduction in the capital-output ratio. Thus it is likely that substantially faster demand growth would have been met by an increase in net imports rather than by matching extra output.
7. Analysis of economic growth
We bring these findings together in table 13, which shows growth in France, Germany and the United Kingdom over the period 1997-2002 from a number of different perspectives. The first group of figures presents an analysis of the growth in components of demand over the period. While overall growth in France and the UK were very similar, in France consumption growth accounted for only two thirds of the total overall growth, while in the United Kingdom it absorbed the whole of the growth in demand. Investment growth was reconciled with the overall expansion of output only by a sharp negative contribution from net trade. Germany suffered the combined effects of slow consumption growth and falling investment demand. The latter can be seen in the context of the high capital-output ratio and low rate of profit on which we have already commented (section 5). Finally, it is striking that the share of consumption in GDP is considerably higher in the United Kingdom than in the two continental countries.
The supply analysis is carried out using the standard techniques of growth accounting discussed above. The changes in labour input are the annual figures underlying those in table 3 multiplied by the share of labour income in GDP at market and basic prices respectively, depending on which definition of GDP is being explained. The contributions of capital to GDP growth in basic prices are those shown in table 12, while the corresponding figures for market prices are derived by using shares of gross and net profit in GDP at market rather than basic prices when evaluating the contributions.
Total factor productivity growth is that part of economic growth which is not accounted for by changes in factor inputs or changes in the basic price adjustment. There is an arbitrary element to the split between residual growth and growth of labour input. Our measures of labour input are not adjusted for changes to the quality of labour, arising from poorly educated older people retiring from the labour force and being replaced by better educated young people. Had we made that adjustment we would have shown a larger labour input and a smaller residual. Mason and O'Mahony (2004) suggest that, over the period 1995-2000, this contributed to output per worker of 0.32 per cent points per annum in the United Kingdom, 0.47 per cent points in France, and 0.01 per cent per annum in Germany. This goes part of the way to explain the differences in total factor productivity growth as calculated using capital stock data. On the other hand, if one accepts the total factor productivity figures calculated from the capital services figures, then adjustment for labour quality changes appears to widen the gap between Germany's and the United Kingdom's total factor productivity growth.
In France and Germany the main source of growth is total factor productivity growth. This is true in the United Kingdom if the capital input is measured using capital stock data. However, if one uses the OECD capital services data, it is the extra depreciation arising from ICT intensive investment. Using these estimates, the good performance of France is due to high residual total factor productivity growth, while Germany's performance relative to the United Kingdom is depressed by low labour input and a low depreciation contribution.
As table 3 showed, the stagnant labour input is associated with both a falling employment ratio and declining hours worked per person employed; these will have to be reversed if Germany is to succeed in growing in the same way as France and the United Kingdom. Germany has, however, also experienced a decline in the population aged 15-64 and, compared to the United Kingdom where the population in the same age range has grown at 0.7 per cent per annum, this is a substantial handicap.
Finally, we look at the relationship between real national disposable income and GDP. This is defined using the identity of equation (1). The decomposition is additive so that positive entries show why the increase in real national disposable income is larger than that of GDP and negative entries the opposite. Thus the positive entry for the terms of trade for Germany contributes to national income expanding less than GDP, while the negative net foreign income for the United Kingdom contributes to an expansion of real national disposable income larger than that of GDP. When looking at contributions to growth on the supply side, the depreciation figures showed the impact of extra depreciation associated with new investment. Here, by contrast, they indicate the overall change in depreciation.
This table brings out the striking fact that while the United Kingdom's GDP grew by 0.8 per cent per annum more than Germany's, its real national disposable income grew by 2 per cent per annum more, at least if official depreciation figures are believed. The similar performance of France and Germany in terms of GDP was turned into a gap of 0.7 per cent per annum in terms of real national disposable income. The table also sheds some light on the observation often made that consumption in Germany has been depressed. It shows that, although the growth rate of consumption was below that of GDP, it was faster than that of real national disposable income. Weak consumption performance cannot be seen as a cause of Germany's poor outturn, instead we must look to sluggish output growth worsened by an adverse movement in the terms of trade (rise in the price of consumption relative to that of output).
The table also shows that UK real national disposable income is a larger share of GDP than is income in France or Germany. This arises from a combination of lower depreciation charges (because the economy is less capital intensive) and a favourable balance of net income from abroad. Relative to real national income consumption in the UK is still high but the excess over France and Germany is lower than the GDP figures suggest.
8. Conclusions
Study of the growth performance of France, Germany and the United Kingdom illustrates that in GDP terms the performance of both France and the United Kingdom was very similar in the period 1997-2002. In 2003, for which we do not have the full data needed for our analysis, the growth performance of the United Kingdom was of course considerably better than that of France. But with only one year of markedly different performance there is at present no reason to attribute this to anything other than the fact that growth rates of countries with similar performances do not always coincide. In income terms the United Kingdom has, nevertheless, performed considerably better than France. This cannot have anything to do with supply side issues as they are conventionally studied because the difference does not arise in the production process.
Germany, by contrast, is the sick member of the group. It has suffered from slow growth in GDP (although its performance when GDP is measured at basic prices is better than when it is measured at market prices) and been weakened further by adverse terms of trade movements. On the demand side output has been particularly depressed by falling demand. We attribute this to a high capital-output ratio and a low net rate of return on capital. The share of labour income in GDP is higher than that of the United Kingdom even though the German economy is more capital intensive.
However, on the supply side, Germany's problems seem to be more to do with stagnant labour input than with either declining investment or a low rate of technical progress. There may of course be a number of factors explaining worsening labour supply conditions in Germany. However a panel analysis of the European Union does suggest that there is a link between the overall employment tax burden and labour supply. There may also be a link between marginal tax rates faced by the unemployed on returning to employment and their willingness to take up work.
In these respects both Germany and France perform badly. If both countries were able to expand their labour inputs to the levels seen in the United Kingdom, then they would be able to enjoy a sustained period of relatively rapid economic growth and it is likely that they would resume the economic lead over the United Kingdom which they enjoyed for much of the late 20th century. On the other hand, it can also be remarked that, were the United Kingdom to find a solution to its productivity problem--that output per hour worked is lower than in France and Germany--it too would have the opportunity for faster growth.
Table 1. ONS estimates of comparative productivity
GDP per worker NPD per worker
France Germany UK France Germany UK
1997 119.6 105.1 100 115.2 100.4 100
1998 119.1 103.3 100 115.0 98.9 100
1999 118.8 104.2 100 114.5 99.6 100
2000 116.5 103 100 112.0 98.3 100
2001 115.1 100.5 100 110.6 96.0 100
2002 112.9 98.8 100 108.7 94.4 100
GDP per hour worked NDP per hour worked
1997 132.4 121.7 100 127.6 116.2 100
1998 132.4 119.7 100 127.9 114.6 100
1999 131.6 120.8 100 126.6 115.5 100
2000 132.3 119.8 100 127.2 114.3 100
2001 132.9 118.1 100 127.7 112.8 100
2002 131.7 116.4 100 126.8 111.2 100
Source: Barnes (2004) and OECD National Accounts. Net figures are
calculated on the assumption that gross data relate to GDP at market
prices.
Table 2. OECD estimates of comparative
productivity, 2002
France Germany United Kingdom
GDP/worker 118.3 107.8 100
GDP/hour 130.7 127.4 100
NDP/worker 114.3 102.9 100
NDP/hour 126.3 121.7 100
Source: See footnote 2. National estimates of depreciation shares.
Table 3. Labour supply
1997 1998 1999
Female participation (%)
Germany 55.3 55.8 57.4
France 52.4 53.1 54.0
United Kingdom 63.1 63.6 64.2
Male participation (%)
Germany 71.9 71.9 72.8
France 66.9 67.4 68.0
United Kingdom 76.6 77.3 77.7
Female participation aged 55-64(%)
Germany 28.7 28.3 28.8
France 25.0 24.4 25.4
United Kingdom 38.5 39.2 39.9
Male participation aged 55-64 (%)
Germany 47.5 47.2 46.8
France 33.2 32.5 32.3
United Kingdom 58.4 59.1 59.7
Unemployment rate (%)
Germany 9.7 9.1 8.4
France 11.8 11.4 10.7
United Kingdom 6.9 6.2 5.9
Average annual hours worked per person
in employment
France 1565 1553 1547
Germany 1496 1489 1479
United Kingdom 1737 1731 1719
Average hours worked per person aged
15-64
France 954 957 963
Germany 937 944 956
United Kingdom 1214 1217 1215
2000 2001 2002
Female participation (%)
Germany 58.1 58.7 58.8
France 55.2 56.0 56.7
United Kingdom 64.8 65.0 65.3
Male participation (%)
Germany 72.9 72.8 71.7
France 69.2 69.7 69.5
United Kingdom 78.1 78.3 78.0
Female participation aged 55-64(%)
Germany 29.0 29.4 30.1
France 26.3 27.8 30.6
United Kingdom 41.7 43.1 44.7
Male participation aged 55-64 (%)
Germany 46.4 46.5 47.1
France 33.6 36.2 39.3
United Kingdom 60.1 61.7 62.6
Unemployment rate (%)
Germany 7.8 7.8 8.6
France 9.3 8.5 8.8
United Kingdom 5.4 5.0 5.1
Average annual hours worked per person
in employment
France 1500 1477 1459
Germany 1463 1451 1444
United Kingdom 1708 1711 1707
Average hours worked per person aged
15-64
France 958 956 946
Germany 947 945 934
United Kingdom 1214 1217 1221
Source: Eurostat Structural Indicators and OECD Labour Market
Statistics.
Table 4. Contributions to changes in the labour
input
Annual France Germany United Kingdom
% change 1997-2002 1997-2002 1997-2002
Population age 15-64 0.37 -0.15 0.72
+ Participation ratio 0.60 0.29 0.02
= Labour force 0.98 0.14 0.74
+ Employment rate 0.71 0.24 0.42
+ Average hours
worked -1.39 -0.70 -0.35
+ Unidentified 0.03 0.38 -0.24
= Labour input (OECD) 0.32 0.07 0.58
Source: OECD Labour Market Statistics, ILO Labour Statistics Database,
and NIESR calculations.
Table 5. Unemployment, employment and disincentives to work
Unemployment Employment Tax Marginal
rate (%) (a) rate (%) wedge tax
rate (%)
1997 2002 1997 2002 1997 2002 2002
France 11.8 8.8 59.7 63.1 41.6 37.8 84.0
Germany 9.7 8.6 63.6 65.3 47.7 45.9 88.4
UK 6.9 5.1 69.9 71.7 28.4 24.7 70.3
EU average 10.0 7.7 60.8 64.2 40.8 37.8 79.1
Source: Eurostat Structural Indicators. Note: (a) Average of men and
women.
Table 6. Unemployment and the tax wedge
Dependent No dummies Country Country
variable is present dummies and year
unemployment present dummies
rate present
Output gap 1.32 -0.715 -0.276
S.E/p value 0.40 [0.002] 0.127 [0.000] 0.168 [0.105]
Tax Wedge 0.054 0.552 0.327
S.E/p value 0.027 [0.050] 0.072 [0.000] 0.112 [0.005]
[R.sup.2] 0.1255 0.909 0.930
Note: Robust standard errors are displayed.
Table 7. Employment and the tax wedge
Dependent variable is No dummies present Country dummies
employment rate present
Men
Output Gap 1.300 0.746
S.E/p value 0.65[0.049] 0.187 [0.000]
Tax wedge -0.149 -0.529
S.E/p value 0.04[0.000] 0.091 [0.000]
[R.sup.2] 0.131 0.936
Women
Output gap 2.517 1.108
S.E/p value 0.996[0.013] 0.271 [0.000]
Tax wedge 0.139 -0.911
S.E/p value 0.114[0.226] 0.105 [0.000]
[R.sup.2] 0.038 0.978
Dependent variable is Country and year Restricted with
employment rate dummies present separate country
and year dummies
for men and women
Men
Output Gap 0.342 0.347
S.E/p value 0.241[0.159] 0.189[0.068]
Tax wedge -0.387 -0.340
S.E/p value 0.150[0.011] 0.081[0.000]
[R.sup.2] 0.942
Women
Output gap 0.352
S.E/p value 0.253[0.168] F-test of restriction
Tax wedge -0.293 F(2,152)=0.17
S.E/p value 0.165[0.081]
[R.sup.2] 0.989 0.991
Note: Robust standard errors are displayed.
Table 8. Tax wedge changes as a source of rising employment
Change to Change to Effect of tax
tax wedge employment wedge change
rate on employment
rate
France -3.8 3.4 1.3
Germany -1.8 1.7 0.6
United Kingdom -3.7 1.8 1.2
Deviation from Impact on Employment rate
UK tax wedge employment of with UK tax
(2002) adopting UK wedge
tax wedge
France 13.1 4.0 67.1
Germany 21.1 7.0 72.3
United Kingdom 0.0 0.1 71.7
Table 9. Capital, depreciation and rates of return,
1998-2002 average
Whole economy
Depreciation Rate of Capital/
return output ratio
(% p.a.) (% p.a.) (years)
Germany 4.83 4.80 3.08
France 5.35 6.64 2.67
United Kingdom 4.85 7.17 2.26
Source: Authors' calculations from national accounts and capital stock
data. All average rates of return are calculated by compounding.
Table 10. Investment ratios, per cent
1998 1999 2000 2001 2002
Gross fixed investment as proportion of GDP at market prices
France 18.4 19.2 20.2 20.1 19.5
Germany 21.4 21.6 21.7 20.3 18.6
United Kingdom 17.5 17.1 16.9 16.8 16.3
Depreciation as proportion of GDP at market prices
France 14.0 14.0 14.2 14.4 14.6
Germany 14.8 14.7 14.9 15.0 15.1
United Kingdom 10.8 11.0 11.0 11.0 11.1
Depreciation as proportion of GDP at basic prices
France 15.8 15.7 15.9 16.1 16.3
Germany 16.4 16.5 16.7 16.8 16.8
United Kingdom 12.4 12.7 12.7 12.7 12.8
Average Average
1987-2002 1998-2002
Gross fixed investment as proportion of GDP at market prices
France 20.1 19.5
Germany 22.0 20.7
United Kingdom 17.6 16.9
Depreciation as proportion of GDP at market prices
France 14.3
Germany 14.9
United Kingdom 11.0
Depreciation as proportion of GDP at basic prices
France 16.0
Germany 16.6
United Kingdom 12.6
Source: Authors' calculations from national accounts.
Table 11. Rates of return
Gross rate Capital/output Fitted rate
of return ratio
(% p.a.) (years) (% p.a.)
France 12.34 2.67 12.34
Germany 9.87 3.08 9.73
UK 12.37 2.26 16.30
Note: Gross rates of return are calculated compounding the net rate and
the depreciation rate.
Table 12. The contribution of capital to growth of
real GDP at basic prices, per cent per annum,
1997-2002
France Germany United
Kingdom
A. Growth of net capital stock 2.12 1.82 2.61
B. Growth of ICT services 1.11 1.16 2.81
C. Growth of other services 1.65 1.33 2.45
D = B + C Growth of all
capital services 2.76 2.49 5.26
E. Contribution to net output
growth as percentage of
GDPB calculated from net
capital stock 0.42 0.30 0.49
F. Contribution to gross
output growth as percentage
of GDPB calculated from
net capital stock 0.76 0.60 0.82
G = F - E. Contribution to
depreciation as percentage
of GDPB calculated from
net capital stock 0.34 0.30 0.53
H. Contribution to gross
output growth calculated
from capital services 0.99 0.83 1.57
I = H - E. Contribution to
depreciation calculated from
capital services 0.57 0.53 1.08
Source: OECD and authors' calculations.
Table 13. Contributions to growth 1997-2002
Contributions to growth
(% points p.a.)
France Germany UK
Components of demand to growth in GDP at market prices
Consumption 2.0 1.0 2.8
Investment 0.9 -0.2 0.8
Net trade -0.1 0.7 -0.8
Total 2.7 1.6 2.7
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital stock data
Labour Hours 0.2 0.0 0.4
Net capital Net output 0.4 0.3 0.4
Depreciation 0.3 0.3 0.3
Factor cost adjustment 0.2 0.0 0.4
Total factor productivity 1.7 1.0 1.3
Total 2.7 1.6 2.7
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital services data
Labour 0.2 0.0 0.4
Capital Net output 0.4 0.3 0.4
services Depreciation 0.5 0.5 0.9
Factor cost adjustment 0.2 0.0 0.4
Total factor productivity 1.5 0.8 0.6
Total 2.7 1.6 2.7
Components of income
RNDI 2.2 1.1 2.6
Depreciation 0.4 0.2 0.3
Less net foreign income 0.0 0.0 -0.3
Terms of trade 0.1 0.3 0.1
Total 2.7 1.6 2.7
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock
Labour Hours 0.2 0.0 0.4
Net capital Net output 0.4 0.3 0.5
Depreciation 0.3 0.3 0.3
Total factor productivity 1.9 1.1 1.4
Total 2.8 1.8 2.6
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock from capital services
Labour Hours 0.2 0.0 0.4
Net capital Net output 0.4 0.3 0.5
Depreciation 0.6 0.5 1.1
Total factor productivity 1.6 0.9 0.7
Total 2.8 1.8 2.6
As contributions to real net national disposable income
Consumption 2.0 1.2 3.2
Savings 0.5 0.0 0.0
Total 2.5 1.2 3.2
Average annual growth
rates (% points p.a.)
France Germany UK
Components of demand to growth in GDP at market prices
Consumption 2.7 1.5 3.5
Investment 4.9 -1.1 4.3
Net trade n/a n/a n/a
Total 2.7 1.6 2.7
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital stock data
Labour Hours 0.3 0.1 0.6
Net capital Net output 2.1 1.8 2.6
Depreciation n/a n/a n/a
Factor cost adjustment 2.5 -0.2 3.3
Total factor productivity n/a n/a n/a
Total 2.7 1.6 2.7
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital services data
Labour 0.3 0.1 0.6
Capital Net output 2.8 2.5 5.3
services Depreciation n/a n/a n/a
Factor cost adjustment 2.5 -0.2 3.3
Total factor productivity n/a n/a n/a
Total 2.7 1.6 2.7
Components of income
RNDI 2.6 1.4 2.9
Depreciation 2.9 1.6 2.9
Less net foreign income n/a n/a n/a
Terms of trade n/a n/a n/a
Total 2.7 1.6 2.7
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock
Labour Hours 0.3 0.1 0.6
Net capital Net output 2.1 1.8 2.6
Depreciation n/a n/a n/a
Total factor productivity n/a n/a n/a
Total 2.8 1.8 2.6
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock from capital services
Labour Hours 0.3 0.1 0.6
Net capital Net output 2.8 2.5 5.3
Depreciation n/a n/a n/a
Total factor productivity n/a n/a n/a
Total 2.8 1.8 2.6
As contributions to real net national disposable income
Consumption 2.2 1.3 3.4
Savings 3.5 -0.2 -6.9
Total 2.5 1.2 3.2
Average share
(current prices, %)
France Germany UK
Components of demand to growth in GDP at market prices
Consumption 77.6 77.6 84.5
Investment 19.6 20.6 17.3
Net trade 2.7 1.8 -1.8
Total 100.0 100.0 100.0
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital stock data
Labour Hours 57.3 59.9 59.3
Net capital Net output 17.7 14.7 16.6
Depreciation 14.3 14.9 11.0
Factor cost adjustment 10.7 10.4 13.2
Total factor productivity n/a n/a n/a
Total 100.0 100.0 100.0
Supply contribution to growth in GDP at market prices: capital
contribution calculated from capital services data
Labour 57.3 59.9 59.3
Capital Net output 17.7 14.7 16.6
services Depreciation 14.3 14.9 11.0
Factor cost adjustment 10.7 10.4 13.2
Total factor productivity n/a n/a n/a
Total 100.0 100.0 100.0
Components of income
RNDI 86.0 83.6 89.4
Depreciation 13.9 14.9 11.0
Less net foreign income 0.1 1.5 -0.3
Terms of trade n/a n/a n/a
Total 100.0 100.0 100.0
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock
Labour Hours 64.2 66.9 68.3
Net capital Net output 19.8 16.5 19.1
Depreciation 16.0 16.6 12.6
Total factor productivity n/a n/a n/a
Total 100.0 100.0 100.0
Supply contribution to GDP at basic prices: contribution of capital
calculated from net capital stock from capital services
Labour Hours 64.2 66.9 68.3
Net capital Net output 19.8 16.5 19.1
Depreciation 16.0 16.6 12.6
Total factor productivity n/a n/a n/a
Total 100.0 100.0 100.0
As contributions to real net national disposable income
Consumption 91.0 92.8 94.4
Savings 9.0 7.2 5.6
Total 100.0 100.0 100.0
Sources: Authors' calculations.
NOTES
(1) The lack of awareness of these issues is demonstrated by the fact that a recent discussion paper issued by the UK Government (HM Treasury, 2004) makes no reference to these issues which are central to assessing economic performance.
(2) http://www.oecd.org/topicstatsportal /0,2647,en_2825_30453906_1_1_1_1_1,00.html#30453948
(3) This, of course, takes account of the return to structures but ignores the return to land.
(4) The data are available from http://europa.eu.int/comm/eurostat/newcronos/queen /display.do?screen=welcome&close=/&language=en&product=YES&root =YES_copy_539019591709#/
(5) The output gap data are calculated using a band-pass filter (Massmann, Mitchell and Weale, 2003), applied to series whose forecast values are taken from NiGEM.
(6) To the extent that it facilitates private sector production the surplus presumably accrues to the private sector. Private profits are higher than they would be because no charge is made to the production sector for use of the roads.
(7) O'Mahony (1999) has argued that for capital stock figures to be internationally comparable, common (and in her case US) depreciation rates should be used. This has the effect of changing estimates of net national income from the published data. Here we wish to comment on the published data, so instead we carry out an algebraic analysis of the effects of changes to depreciation assumptions on the share of depreciation and the rate of return.
(8) The same is true in a closed economy if the rate of return net of taxes on capital is equated to the subjective discount rate (Chamley, 1986). In either case tax on income from capital is shifted onto labour.
(9) The third possibility, that the return on capital at the margin in the UK is much higher than the return on average, seems most unlikely in a generally competitive economy.
(10) All the trend variables we discuss are calculated using the Baxter-King band pass (Baxter and King, 1999; Massmann, Mitchell and Weale, 2003). This is a mechanical means of smoothing data by removing high-frequency fluctuations. The trend estimates close to the end of the period depend on the assumptions made about future values of the variables. We use values from our macroeconomic forecasts of the countries concerned.
REFERENCES
Barnes, M. (2004), 'International comparisons of productivity: better data improve UK productivity position', Office for National Statistics http://www.statistics.gov.uk/articles/nojournal/Feb_ICP.pdf
Barrell, R, Metz, R. and Riley, R. (2004), 'Recent UK growth: a comparison with France, Germany and the US', National Institute Economic Review, 187, pp. 53-7.
Baxter, M. and King, R.G. (1999), 'Measuring business cycles: approximate band-pass filters for economic time series', Review of Economics and Statistics, 81, pp. 575-593.
Belot, M. and van Ours, J. (2000), 'Does the recent success of some OECD countries in lowering their unemployment rates lie in the clever design of their labour market reforms?', CEPR Discussion Paper No. 2492.
Chamley, C. (1986), 'Optimal taxation of capital income in general equilibrium with infinite lives', Econometrica, 54, pp. 607-22.
Daveri, F. and Tabellini, G. (2000), 'Unemployment, growth and taxation in industrial countries', Economic Policy, 30, pp. 47-104.
Dutta, J., Sefton, J. and Weale, M.R. (2000), 'Real national income', National Institute Discussion Paper No 163.
Groth, C., Gutierrez-Domenech, M. and Srinivasan, S. (2004), 'Measuring total factor productivity growth in the United Kingdom', Bank of England Quarterly Bulletin, pp. 63-73.
Hall, R.E. (1988), 'Intertemporal substitution in consumption', Journal of Political Economy, 96, pp. 339-57.
HM Treasury (2004), 'Productivity in the UK 5: benchmarking UK productivity performance', London, HM Treasury.
Hultgren, T. (1965), Costs, Prices and Profits: Their Cyclical Relations, New York, NBER.
Lee, C. (2001), 'Finite sample bias in IV estimation of inter-temporal labour supply models: is the inter-temporal elasticity of substitution really small?', Review of Economics and Statistics LXXXIII, pp. 638-46.
Massmann, M, Mitchell, J. and Weale, M.R. (2003), 'Business cycles and turning points: a survey of statistical techniques', National Institute Economic Review, 183, pp. 90-106.
Mason, G. and O'Mahony, M. (2004), 'ICT investments, workforce skills and industrial performance in the UK', in Raising UK Productivity--Developing the Evidence Base for Policy, DTI Economics Paper No 8.
Nickell, S. (1998), 'Unemployment: questions and some answers', Economic Journal, 108, pp. 802-16.
Nickell, S. and van Ours, J. (2000), 'The Netherlands and the United Kingdom: a European unemployment miracle?', Economic Policy, 30, pp. 135-80.
O'Mahony, M. (1999), Britain's Productivity Performance: 1950-1996, London, National institute of Economic and Social Research.
Oulton, N. (2002), 'Productivity versus welfare: or, GDP versus Weitzman's NDP', Bank of England Working Paper No. 163, forthcoming in Review of Income and Wealth.
Rowthorn, R. (1999), 'Unemployment, wage bargaining and capital-labour substitution', Cambridge Journal of Economics, 23, 4, pp. 413-26.
Rotemberg, J. and Woodford, M. (1999), 'The cyclical behaviour of prices and costs', in Taylor, J. and Woodford, M. (eds), Handbook of Macroeconomics, Amsterdam, North Holland, pp. 1051-135.
Svensson, L.E.O. (2000), 'Open-economy inflation targeting', Journal of International Economics, 50, pp. 155-83.
Robert Metz, Rebecca Riley and Martin Weale *
* National Institute of Economic and Social Research. We are grateful to Ray Barrell and Mary O'Mahony for their comments. This paper was presented at the Institut de la Gestion Publique et du Development Economique, Ministry of Finance, Paris on 2 April 2004. We thank the ESRC for financial support.
In addition, make sure to read these articles:
- Better European Train Travel
- How Small Business Owners Feel About Doing Business
- Serving Up Some Seductive Fallacies
- Study: Germany, France, U.k. Lead In Europe B'band
- Companies Lulled Into False Sense Of Security By URL Blocking Technology
- Hammerson Chief Exec Sees REITs Changing United Kingdom's Retail Picture
- More Than 80% of EMEA Speech Deployment Revenues Are Generated In France, Germany, and the...
- Sony Connects With U.K., France, Germany
- Sony Connects with U.K., France, Germany
- France Telecom Announces the Launch of Its VoiLaPortal in the United Kingdom.
- Car Plugs in the United Kingdom Industry Profile Is an Essential Resource for Top-Level...
- TIW Wins UMTS License in the United Kingdom TIW andHutchison Announce Cooperation on U.K.
- Hardware: How are the markets faring?
Email
Digg It
del.icio.us
