Labour Productivity and Foreign Ownership in the UK ... - CiteSeerX

Labour Productivity and Foreign Ownership in the UK

Nicholas OULTON September 1998

NIESR Discussion Paper no. 143

Address for correspondence Bank of England Monetary Analysis Threadneedle Street LONDON EC2R 8AH Email: [email protected]

CONTENTS Page Abstract

1

1. Introduction

2

2. The One Source database of company accounts

3

3. Ownership and productivity: results

5

4. Interpreting the findings

12

5. Concluding remarks

15

References

15

Appendix A Divisions and Classes of SIC80

17

Appendix B Descriptive statistics

19

ABSTRACT Previous studies have found that in manufacturing foreign-owned companies have a substantial productivity lead over domestically-owned ones, but is the same true in the rest of the economy? We investigate this question using a very large database of company accounts. The answer is yes. After controlling for industrial composition and other factors, foreign ownership was found to raise productivity by about a third in non-manufacturing. The foreign productivity lead, which is about the same over UK subsidiaries as over UK independents, can very largely be explained by higher capital per employee and a more skilled labour force.

Keywords

Capital, productivity , foreign ownership

JEL codes

D24, J24, L60

1

1. Introduction1 Foreign-owned firms in manufacturing have substantially higher labour productivity than domestically owned ones. Furthermore, the productivity gap is at least partly explained by the fact that foreign-owned firms have higher capital intensity and use more skilled labour than their domestic counterparts. These findings apply to the UK (Davies and Lyons 1991; Oulton 1998b), the US (Doms and Jensen 1998) and Canada (Globerman et al. 1994). For the UK, this internal productivity gap between foreignand domestically-owned companies is comparable in size to the gap between UK manufacturing and e.g. German or French manufacturing as a whole.

But foreign ownership is not confined to manufacturing which is only around a fifth of GDP anyway. So there is considerable interest in seeing whether the foreign-domestic productivity gap is as large in the rest of the economy. This is the purpose of the present paper. The findings on manufacturing have been derived from studying longitudinal databases of each country’s production census. Outside of manufacturing, no such source exists. However for the UK we can utilise a large electronic database of company accounts, the OneSource database. This database, which has been employed in earlier work on employment growth (Hart and Oulton (1996), (1998) and (1999)) and productivity (Oulton 1998a) is described more fully in the next section.

The database covers both independent companies and subsidiaries. To avoid doublecounting, e.g. including both the parent and the subsidiaries of which it is composed, we divide our companies into four groups: (1) subsidiaries owned by US companies; (2) subsidiaries owned by other foreign companies; (3) subsidiaries owned by UK companies; and (4) independent companies which do not own any subsidiaries. The division between US-owned and other foreign owned subsidiaries was suggested by the results for manufacturing.

1

I owe thanks to the Leverhulme Trust for financing this research as part of a wider project entitled Job generation in the corporate sector [F/59/AD]. This research was carried out at the National Institute of Economic and Social Research prior to my taking up an appointment at the Bank of England. I am grateful to Peter Hart and Martin Weale for helpful comments. The usual disclaimer applies.

2

In what follows, section 2 describes how the sample of companies was derived from the OneSource database. Section 3 presents the main results. Section 4 discusses some alternative interpretations while section 5 concludes.

2. The One Source database of company accounts Our data are derived from the OneSource CD-ROM entitled “UK Companies, Volume 1” for December 1996. This contains the accounts of some 110,000 larger UK companies. The ultimate source is the accounts which companies are legally required to deposit at Companies House. The criteria for inclusion in Volume 1 is stated by OneSource to be: “All public limited companies, all companies with employees greater than 50, and the top companies based on turnover, net worth, total assets, or shareholders funds (whichever is largest) up to a maximum of 110,000 companies”. Only “live” companies are included. Companies which are dormant, dissolved, in liquidation, or in the process of being wound up are excluded.

The database contains the latest available accounts and related information for each included company, including the date of the end of the accounting period. Though the CD-ROM is dated December 1996, the data relate to a somewhat earlier period, which varies between companies according to the date of their accounts. For the great majority of companies, this date falls within 1995 (the average is about two thirds of the way through 1995). Hence for simplicity we refer to the year to which the data relate as 1995.2 Companies are classified under the 1980 SIC.

Output can be measured by either sales or value added (the latter defined as trading profit plus the wage bill). Physical capital is measured by the book value of fixed assets. Clearly, this is likely to be a very noisy measure of the true value, since it is in nominal terms and companies differ both in the time pattern of asset acquisition and in their depreciation practices. Employment is a headcount. There is no breakdown by

2

We excluded 1,104 companies whose accounts predated 1994. A few other companies were also excluded since they claimed to have zero employees, even though supposedly actively trading.

3

type of labour or by skill but we can calculate the average wage which may serve as a proxy for the average level of human capital per worker.

For each company, OneSource gives first, the country of the holding company which owns the company in question and second, the country of the ultimate holding company. Either or both of these may of course be missing; “country” can include the UK. Foreign-owned companies are broken down into two groups, (a) US-owned companies and (b) other foreign-owned companies. A company is classified as USowned if either the country of the ultimate holding company is the US or, if this is missing, the country of the holding company is the US. Other foreign ownership is determined analogously.

Amongst UK-owned companies we distinguish between subsidiaries of UK-owned companies and independent UK companies which do not own any subsidiaries. This is to avoid double-counting: if e.g. a UK-owned company owns five UK subsidiaries, we include the five subsidiaries but not the holding company.3

Avoidance of double

counting leads to the elimination of 21,009 companies. A company is classified as a UK-owned subsidiary if is not an ultimate holding company and either the country of its ultimate holding company is the UK or, if this is missing, the country of its holding company is the UK. A company is classified as a UK-owned independent without subsidiaries if it is not an ultimate holding company and it is not a subsidiary.

These four categories should be mutually exclusive but unfortunately this is not the case in practice. There is an inconsistency in the OneSource database: some companies are classified as subsidiaries by one variable, the subsidiary indicator variable, but as independent by the type of ownership variable; the latter variable is the one used to exclude ultimate holding companies. We cannot resolve this inconsistency so we simply drop the companies which fall into more than one category. This leads to the elimination of 1,447 companies.

4

A further 33,734 companies are lost due to missing data. This is mostly due to missing data for employment or sales. Zero values, mainly for sales or fixed assets, are also encountered, affecting 11,526 companies. These are also dropped, first because zero values are highly implausible for an actively trading company and secondly because we intend to use logs. In summary, we start with 107,829 companies (after eliminating companies with out-of-date accounts), we eliminate a further 67,716 and our eventual sample is 40,113 companies. These 40,113 companies employed collectively 10.020 million people. Excluding self-employment, the private sector, a wider category than the corporate sector, employed just over 17 million in mid-1995 (Economic Trends Annual Supplement 1997, Table 3.8), so the major part of employment in the corporate sector is covered by our analysis.4

3. Ownership and productivity: results Our hypothesis is that productivity may differ between companies because (a) some companies use more inputs per worker and (b) some companies may have access to superior technology or superior business systems or may have superior products, i.e. products which can be sold at a higher price. Input intensity is measured by physical capital per worker and human capital per worker, the latter proxied by the wage. We have no measures of superior technology or products but we can check whether, after controlling for input intensity, higher productivity is associated with ownership, to measure which we employ ownership dummies. Note that different types of ownership may be associated with greater or smaller input intensity. So ownership can have a direct effect on productivity, say if foreign firms have access to superior technology, and an indirect effect, say if foreign firms are more capital intensive.

In addition we employ a number of control variables. We use dummy variables for the SIC80 Class to which each company is assigned (there are 60 Classes in SIC80 3

The accounts for a holding company would normally be consolidated, i.e. they would incorporate the results of its subsidiaries. In some cases, the results of foreign subsidiaries may be included in the accounts of UK-based holding companies. Our procedure ensures that such results are also excluded. 4 There are 49,009 companies for which employment is available. These collectively employed 10.775 million. Value added is available for a smaller number. In the regressions with value added as

5

covering the whole economy: see Appendix A). Since companies’ accounts do not all relate to exactly the same period, the date of each company’s financial year end is included as a control. We include too company age since new companies may have not yet reached their optimal scale. Scale, measured by employment, is also used as a control variable, even though it turns out to vary systematically with ownership type.

We start by considering some descriptive statistics. Appendix B, Table B1, shows the proportion of employment in each type of ownership by SIC80 Class. Overall, UK subsidiaries account for 55% of employment, US-owned companies for 12% and other foreign-owned companies for 22%. The remainder, 11%, is in UK independents. But this latter figure is an underestimate of the population proportion since our sample excludes many smaller companies.

Appendix B, Table B2, shows employment-weighted means of labour productivity (value added per employee, V/L) for each SIC80 Class and for the four types of ownership. It also shows employment-weighted means of the determinants or correlates of productivity: physical capital intensity (K/L), human capital intensity (w), and size (employment, L). All these means are expressed as index numbers with the value for UK independents set equal to 100. This information is summarised in Table 1, which shows percentiles of the distribution across Classes of these employmentweighted means. The rank order for all three measures is foreign-owned first, UK subsidiaries second and last, UK independents. Clearly the distributions for US- and other foreign-owned companies tend to lie above those for UK independents and subsidiaries. Though the largest differences are between UK independents and the rest, the differences between UK subsidiaries and the foreign-owned companies are also substantial. Median productivity is 18% lower in UK subsidiaries than in US-owned companies,

10%

lower

than

in

other

foreign

companies.

Foreign-owned

the dependent variable which are reported below, there are 36,226 companies employing 9.391 million.

6

Table 1 Productivity and determinants by ownership type: distribution of within Class means across SIC80 Classes (UK independents=100) Variable Ownership V/L

US Other foreign UK-owned US Other foreign UK-owned US Other foreign UK-owned

K/L

w

Percentiles 25th 50th 113.0 139.0 113.1 126.8 99.6 113.8 124.2 149.9 135.6 165.4 91.8 110.1 103.9 119.5 104.4 117.5 92.1 101.9

75th 170.9 150.0 126.0 217.9 226.1 139.6 145.8 128.2 113.8

Source

Appendix B, Table B2.

Note

36 SIC80 Classes (35 for US-owned). Within-Class means are employment-weighted.

companies have much higher capital intensity. Median capital per employee is 36% higher in US companies, 50% higher in other foreign companies. Foreign companies also pay much higher wages, indicating a considerably more skilled labour force. The median wage is 17% higher in US companies, 15% higher in other foreign ones.

In order to see whether these impression stand up to more rigorous analysis, we regress input intensity and other characteristics of companies on the ownership dummies and the controls. We do these regressions separately for manufacturing (SIC80 Divisions 2-4) and non-manufacturing (Divisions 0,1, and 6-9): see Tables 2 and 3. Physical capital intensity is measured by the log of the capital-labour ratio, while the log of the wage acts as a proxy for human capital per worker. We also include the log of company age, dummies for SIC80 Class, and three ownership dummies: US (=1 if US-owned), NON-US (=1 if foreign- but not US-owned), and UKSUB (=1 if owned by

a

UK

company),

with

UK

7

independents

being

the

omitted

Table 2 Determinants of labour productivity and ownership: manufacturing companies in 1995

ln(K/L)

Dependent variable ln(K/L) ln(w)

ln(age)

0.0367** (0.0105)

0.0601** (0.0102)

0.0128** (0.0047)

0.0048 (0.0046)

0.2766** (0.0124)

0.0042 (0.0061)

US

0.2039** (0.0365)

0.3152** (0.0348)

0.2008** (0.0138)

0.1627** (0.0126)

1.3174** (0.0462)

0.2286** (0.0216)

NON-US

0.2663** (0.0325)

0.3538** (0.0310)

0.1333** (0.0125)

0.1033** (0.0124)

1.0364** (0.0360)

0.1267** (0.0164)

-0.1921** -0.1293** (0.0255) (0.0244)

0.0438** (0.0096)

0.0223* (0.0094)

0.7425** (0.0254)

-0.0060 (0.0119)

11,416 0.151

11,416 0.144

11,416 0.188

11,312 0.110

11,416 0.183

11,416 0.175

-

-

ln(V/L)

0.0845** (0.0096)

N R2

-0.0289** (0.0047)

ln(L)

ln(L)

UKSUB

-

ln(w)

-

Source OneSource. Note Companies included are either subsidiaries or independents which do not own any subsidiaries. US=1 if US-owned; NON-US=1 if foreign but not US-owned; UKSUB=1 if company is a subsidiary of a UK company. Omitted category is UK independent companies which do not own any subsidiaries. Constant and dummies for the 60 SIC80 Classes included but not reported. Estimated by OLS. Robust standard errors are in parentheses. * **

Significant at the 5% level Significant at the 1% level

8

Table 3 Determinants of labour productivity and ownership: non-manufacturing companies in 1995

ln(K/L) ln(L)

-0.1490** (0.0080)

Dependent variable ln(K/L) ln(w) ln(w) -

-0.0658** (0.0032)

-

ln(L) -

ln(V/L) -

ln(age)

0.2094** (0.0103)

0.1670** (0.0099)

0.0169** (0.0041)

-0.0018 (0.0040)

US

0.2487** (0.0374)

0.0845* (0.0361)

0.4864** (0.0133)

0.4139** (0.0135)

1.1025** (0.0379)

0.2962** (0.0235)

NON-US

0.1576** (0.0285)

0.0532 (0.0279)

0.3651** (0.0111)

0.3190** (0.0114)

0.7006** (0.0254)

0.2718** (0.0167)

-0.1645** -0.2801** (0.0220) (0.0212)

0.1386** (0.0085)

0.0875** (0.0085)

0.7762** (0.0181)

0.0246* (0.0110)

28,697 0.268

28,697 0.251

28,697 0.239

24,914 0.256

UKSUB

N R2

28,697 0.205

28,697 0.191

0.2845** -0.0280** (0.0090) (0.0057)



9

category. To pick up any possible scale effects, company size (the log of employment) is also included in some regressions.

The magnitude of the ownership effects depends very much on whether company size is included separately in the regressions. In general we find that, relative to UK independents, foreign ownership raises input intensity. The effect is particularly strong for physical capital intensity in manufacturing and for human capital intensity in nonmanufacturing. Ownership by a UK company reduces physical capital intensity substantially while raising human capital intensity somewhat, relative to UK independents. Age has a positive effect on physical capital intensity. Also, even after controlling for industrial composition, US companies are significantly larger than all other types of company in both manufacturing and non-manufacturing. Other foreignowned companies are larger than UK independents and UK subsidiaries in manufacturing but in non-manufacturing, while remaining considerably larger than UK independents, they are a little smaller than UK subsidiaries.

Foreign ownership is therefore positively correlated with input intensity. So we shall obtain a maximum estimate of the effect of foreign ownership by regressing productivity on ownership plus controls only. This is done in the last column of Tables 2 and 3. Relative to both UK independents and UK subsidiaries, in manufacturing US ownership raises productivity by [exp(0.2286)-1=] 26%, while other foreign ownership raises it by 14%. In non-manufacturing the US effect is somewhat larger, 34%, while the other foreign effect is not much less, 31%. UK subsidiaries have a statistically significant advantage over independents but it is very small, about 2%.

Next in Table 4 we shows the results of regressing the log of labour productivity on the two input intensities, the ownership dummies and the controls. Our regressions can explain about 48% of the variance in manufacturing and 58% in non-manufacturing. Despite the fact that both physical and human capital are poorly measured in our data, each of these variables is highly significant. The elasticity of value added with respect to

physical

capital

is

remarkably

10

constant

at

about

12%

Table 4 Effect of ownership on labour productivity (value added per employee): manufacturing versus non-manufacturing companies in 1995 (dependent variable is ln(V/L)) Manufacturing ln(K/L)

Non-manufacturing

0.1247** (0.0065)

0.1202** (0.0064)

0.1216** (0.0040)

0.1242** (0.0041)

ln(w)

0.870** (0.0204)

0.8825** (0.0204)

0.8007** (0.0117)

0.8472** (0.0116)

ln(L)

-0.0279** (0.0045)

ln(age)

-0.0051 (0.0050)

US

-

-0.0894** (0.0037)

-

-0.0121* -0.0175** -0.0449** (0.0050) (0.0046) (0.0045)

0.0899** (0.0193)

0.0543** (0.0183)

0.0946** (0.0198)

-0.0166 (0.0192)

NON-US

0.0165 (0.0137)

-0.0102 (0.0128)

0.1050** (0.0135)

0.0366** (0.0131)

UKSUB

0.0266** (0.0089)

0.0073 (0.0085)

0.0894** (0.0086)

0.0265** (0.0083)

11,312 0.485

11,312 0.482

24,914 0.581

24,914 0.566

N R2



11

while the elasticity with respect to human capital is 80-88%. The effect of age is negative (log age squared was also tried but was insignificant).

The effect of size is also negative. This is somewhat surprising since in this type of regression a positive coefficient is usually found, as indeed was the case with results based on the ARD (Oulton 1998b). Including size does not affect the coefficients on the input intensities very much. But it does affect the size of the ownership effects. Without size, these latter are small or insignificant. The largest effect is for US ownership in manufacturing, 5%. With size included, they become larger. In nonmanufacturing all three effects are of the order of 9-10%; in manufacturing only the US effect is of this size.5

In the regressions of Table 4, the ownership dummies measure the direct effects of ownership; the indirect effects must also be taken into account. We have already seen from Table 3 that ownership has highly significant effects on capital intensity. We therefore conclude that most or even all of the effects of ownership on productivity are indirect, i.e. foreign ownership leads to higher human and physical capital intensity and this accounts for the productivity gap.

4. Interpreting the findings: three alternatives The substantial productivity lead of foreign-owned companies shown by our results is in line with a large literature stressing the productive effects of foreign investment (e.g. Dunning 1981; Barrell and Pain 1997). But now the obvious question to ask is, if foreign-owned establishments, located in Britain and employing British workers, use high human and physical capital intensity to achieve high productivity, why don’t British-owned establishments do the same? There are a number of possible explanations.

5

Similar results were obtained with sales per employee as the dependent variable. But the foreign ownership effects are larger. This suggests that foreign ownership affects the extent to which companies use intermediate input. In other words, foreign-owned companies tend to be more reliant on outsourcing.

12

First, UK-owned companies may face a higher cost of capital than foreign-owned ones. Financial constraints are now widely believed to be an important influence on investment (Caballero 1997; Chirinko 1993; Hubbard 1998). Foreign companies are not presumably constrained to acquire funds for investment from the UK financial system, or at least not to the same extent as UK ones, so deficiencies in the UK system may be hindering investment by UK companies. Foreign companies may also have a lower cost of internal funds (Miles 1993). An obvious objection to this is that large UK companies are themselves multinationals and face the same global capital market as foreign multinationals. However, the argument may have some force for smaller companies. And it is still possible that when making investment decisions out of retained profits even large UK companies are constrained by the short-termist views of the UK stock exchange.6

Second, UK firms may face a less favourable risk-return trade-off than foreign ones and consequently may prefer less capital-intensive technologies. UK companies, even the large multinational ones, almost certainly make a higher proportion of their sales in the UK than do foreign companies. They may be heavily influenced in their investment decisions by the memory of bad experiences in the three long recessions of the last 25 years (the working lifetime of the people now running UK manufacturing companies). If the UK is perceived as having greater macro instability than other countries, then even if UK firms are no more risk averse than foreign ones, they will perceive their overall risk level as higher. By contrast, the large foreign multinationals which operate in UK manufacturing may be better able to balance the risk of poor outcomes in the UK against the chance of good ones elsewhere. Consequently, their preferred capitallabour ratio may be higher. This argument assumes that capital intensive technologies are riskier. This in turn may be justified if investment in physical capital is at least partially irreversible while labour and other inputs may be adjusted at relatively low cost.

Finally, as a third hypothesis, foreign companies may be using superior technology and business methods which happen to be more intensive in both capital and skilled labour.

6

I owe this point to Steve Bond.

13

UK companies may just be slow to learn from and apply the best foreign practice, for several possible reasons. First, the relevant knowledge may be commercially confidential or located in the heads of foreign managers. Second, there may be work force resistance to change. In the latter case, it might not pay for an established firm to adopt the superior technology because of the upfront cost of strikes, etc. This will be all the more likely if the firm is a satisficer rather than a maximiser.7

Objections can be raised against this last explanation too. It would seem rather odd if superior technology is in general more intensive in both capital and skilled labour. Some superior business methods, e.g. just in time, require less (inventory) capital not more. Also, the larger UK companies at least must be well aware of their foreign rivals’ technology and could hire foreign managers if they so desired. And how potent is work force resistance after the trade union reforms of the 1980s? While there are few areas of manufacturing which are not exposed to foreign competition, the same is not true of services. So the fact that the productivity gap seems to be about the same in manufacturing as in non-manufacturing argues against competition or the lack of it being the explanation.

Doms and Jensen (1998) in their study of US manufacturing were able to break down their domestically-owned (i.e. US) firms into those which are multinationals and those which operate only in the home market. They find that the real difference is between multinationals and non-multinationals, not foreign and domestically owned firms. This suggests a fourth explanation based on the theory of foreign investment. At any moment there is a range of capabilities amongst a country’s firms. The better companies develop specific advantages. These allow them to compete successfully in foreign markets and consequently to go multinational (Dunning 1981). The foreigndomestic productivity gap which we observe simply reflects this process. Indeed, the observed gap is on this view rather misleading since the performance of the more successful, domestically owned multinationals is being obscured by their less successful colleagues who operate only in the home market.

7

Baily and Gersbach (1995) argue that the crucial factor in inducing firms to adopt best practice technology is exposure to global, not just local or regional, competition.

14

Whether this explanation works for the UK as well as for the US is unclear. It would require much more work beyond the scope of this paper to identify the UK multinationals in the OneSource database. But even if some British multinationals have high productivity, they must still represent a comparatively small proportion of UK employment, otherwise we would not find that the employment-weighted mean of productivity is generally lower in UK subsidiaries (Tables 1 and B2). In any case, it is not clear that this fourth explanation is different from the other three since the specific advantages of multinationals have to show up in some measurable way.

5. Conclusions The Introduction asked, do foreign-owned companies have as big a lead in labour productivity in the rest of the economy as they do in manufacturing? The answer is yes. In fact the lead is larger. After controlling for industrial composition and other factors, US ownership was found to raise productivity by 26% in manufacturing; other foreign ownership raises it 14%. These estimates are comparable to those found using a different data source, the ARD (Oulton 1998b). In the rest of the economy, US and other foreign ownership raise productivity by even more, 34% and 31% respectively. For both manufacturing and non-manufacturing, the foreign lead is about the same over UK subsidiaries as over UK independents. The foreign productivity lead can very largely be explained, or at least accounted for, by higher capital per employee and a more skilled labour force.

References Baily, M.N. and Gersbach, H. (1995). “Efficiency in Manufacturing and the Need for Global Competition”. Brookings Papers on Economic Activity: Microeconomics 1995, pp. 307-347. Barrell, R. and Pain, N. (1997). “Foreign Direct Investment, Technological Change, and Economic Growth within Europe”. Economic Journal, 107 (November), pp. 1770-1786.

15

Caballero, R.J. (1997). “Aggregate Investment”. NBER Working Paper No. 6264. Cambridge, MA. Davies, S.W. and Lyons, B.R. (1991). “Characterising relative performance: the productivity advantage of foreign owned firms in the UK.” Oxford Economic Papers, 43 (October), 584-595. Dunning, J.H. (1981). International Production and the Multinational Enterprise. London: George Allen and Unwin. Chirinko, R.S. (1993). “Business Fixed Investment Spending: A Critical Survey of Modelling Strategies, Empirical Results, and Policy Implications”. Journal of Economic Literature, 31 (December), 1875-1911. Doms, M.E. and Jensen, J.B. (1998). “Comparing Wages, Skills, and Productivity Between Domestic and Foreign Owned Manufacturing Establishments in the United States”. Mimeo. Globerman, S., Ries, J.C. and Vertinsky, I. (1994). “The Economic Performance of Foreign Affiliates in Canada”. Canadian Journal of Economics, XXVII, No. 1 (February), pp. 143-156. Hart, P.E. and Oulton, N. (1996). Growth and size of firm. The Economic Journal, 106, 1242-1252. Hart, P.E. and Oulton, N. (1998). Job creation and destruction in the corporate sector: the relative importance of births, deaths and survivors.

National Institute of

Economic and Social Research Discussion Paper No. 134. Hart, P.E. and Oulton, N. (1999). “Gibrat, Galton and job generation”. Forthcoming in International Journal of the Economics of Business. Hubbard, G. (1998). “Capital-Market Imperfections and Investment”. Journal of Economic Literature, XXXVI (March), pp. 193-225. Miles, D. (1993). “Testing for short termism in the UK stock market” Economic Journal, 103, pp. 1379-1396. Oulton, N. (1998a). “Competition and the Dispersion of Labour Productivity Amongst UK Companies”. Oxford Economic Papers, 50, 23-38. Oulton, N. (1998b). “Investment, Capital and Foreign Ownership in UK Manufacturing”. National Institute of Economic and Social Research Discussion Paper No. 141 (August).

16

APPENDIX A Divisions and Classes of SIC80 Division 0 Agriculture, forestry and fishing 01 Agriculture and horticulture 02 Forestry 03 Fishing Division 1 Energy and water 11 Coal extraction and manufacture of solid fuel 12 Coke ovens 13 Extraction of mineral oil and natural gas 14 Mineral oil processing 15 Nuclear fuel production 16 Production and distribution of electricity, gas and other forms of energy 17 Water supply industry Division 2 Metals, mineral products and chemicals 21 Extraction and preparation of metalliferous ores 22 Metal manufacturing 23 Extraction of minerals not elsewhere specified 24 Manufacturing of non-metallic mineral products 25 Chemical industry 26 Production of manufacturing-made fibres Division 3 Metal goods, engineering and chemicals 31 Manufacture of metal goods not elsewhere specified 32 Mechanical engineering 33 Manufacturing of office machinery and data processing equipment 34 Electrical and electronic engineering 35 Manufacture of motor vehicles and parts 36 Manufacture of other transport equipment 37 Instrument engineering Division 4 Other manufacturing industries 41/42 Food, drink and tobacco manufacturing industries 43 Textile industry 44 Manufacturing of leather and leather goods 45 Footwear and clothing industries 46 Timber and wooden furniture industries 47 Manufacturing of paper and paper products; printing and publishing 48 Processing of rubber and plastics 49 Other manufacturing industries Division 5 Construction 50 Construction

17

Divisions and Classes of SIC80 (continued) Division 6 Distribution, hotels and catering, and repair 61 Wholesale distribution (except dealing in scrap and waste materials) 62 Dealing in scrap and waste materials 63 Commission agents 64/65 Retail distribution 66 Hotels and catering 67 Repair of consumer goods and vehicles Division 7 Transport and communications 71 Railways 72 Other inland transport 74 Sea transport 75 Air transport 76 Supporting services to transport 77 Miscellaneous transport services and storage not elsewhere specified 79 Postal services and telecommunications Division 8 Banking, insurance and real estate 81 Banking and finance 82 Insurance, except for compulsory social security 83 Business services 84 Renting of movables 85 Owning and dealing in real estate Division 9 Social and personal services 91 Public administration, national defence and compulsory social security 92 Sanitary services 93 Education 94 Research and development 95 Medical and other health services: veterinary services 96 Other services provided to the general public 97 Recreational services and other cultural services 98 Personal services 99 Domestic services

18

APPENDIX B Descriptive Statistics Table B1 Employment by ownership type and SIC80 Class

SIC80 Class 1 2 3 11 12 13 14 15 16 17 21 22 23 24 25 26 31 32 33 34 35 36 37 41 42 43 44 45 46 47 48 49 50 61 62 63

Number of Total companies employment 522 13 37 41 1 141 47 1 55 47 6 313 103 397 853 16 898 2,235 244 1,413 449 265 398 605 474 479 47 511 648 1,935 902 268 3,066 9,087 130 503

156,947 2,549 1,958 15,049 155 51,078 37,380 545 37,077 49,438 6,270 88,475 29,448 131,342 290,101 6,967 155,604 416,808 107,137 430,780 263,191 157,785 61,244 266,887 560,509 189,246 4,367 169,671 100,137 429,218 183,089 36,912 420,583 774,596 6,573 25,849

Proportions of employment (%) UK UK US-owned Other foreign - subsidiary independent owned 0.8 28.9 60.4 10.0 0.0 73.8 11.8 14.4 0.0 15.6 64.0 20.4 0.0 0.1 94.8 5.2 0.0 0.0 100.0 0.0 45.7 33.7 19.6 1.0 38.4 47.0 13.7 0.9 0.0 0.0 100.0 0.0 12.2 0.4 61.6 25.8 0.0 22.2 77.0 0.7 0.0 45.5 50.4 4.1 5.7 37.3 51.9 5.1 0.0 13.7 81.8 4.5 8.0 25.0 58.6 8.5 37.0 22.6 37.2 3.3 67.8 3.4 28.2 0.6 16.9 21.4 49.2 12.4 25.2 20.8 44.5 9.5 37.0 45.4 14.4 3.3 13.5 31.3 46.3 9.0 40.9 32.0 23.4 3.7 2.6 19.0 74.9 3.5 29.5 23.4 38.4 8.6 6.9 22.7 56.2 14.2 11.4 12.1 74.1 2.4 5.2 6.1 80.4 8.3 5.3 5.8 63.9 25.0 6.7 8.7 72.1 12.4 5.2 10.7 63.1 21.0 9.7 45.9 33.7 10.8 9.8 28.3 51.1 10.9 21.4 12.8 48.1 17.8 4.7 15.8 66.4 13.1 14.3 32.6 39.5 13.6 10.9 22.0 40.8 26.4 53.9 18.7 21.8 5.6 19

Table B1, continued 64 65 66 67 71 72 74 75 76 77 79 81 82 83 84 85 91 92 93 94 95 96 97 98 Total

1,083 750,775 2,559 563,832 1,170 594,147 155 22,485 18 16,357 870 317,857 103 42,947 69 101,432 166 40,637 1,252 155,823 129 48,159 761 487,014 250 205,033 6,578 866,083 464 68,627 2,318 32,109 47 5,544 313 311,644 639 63,371 172 26,995 486 119,462 719 61,338 1,235 171,403 303 37,305 49,009 10,775,344

0.8 2.7 16.2 2.7 0.0 2.9 1.1 0.0 4.8 8.1 8.0 4.6 3.9 21.2 21.0 0.1 0.0 12.9 1.5 11.1 7.1 0.2 3.1 18.0 12.0

20

13.3 11.2 26.7 27.4 5.4 11.4 39.9 17.2 20.3 35.7 14.4 24.4 27.3 16.3 13.3 19.9 21.8 29.2 2.8 19.5 15.0 0.5 7.9 3.6 21.7

78.6 74.9 47.1 54.4 92.9 77.9 55.8 23.0 64.1 46.0 72.7 62.2 67.7 50.8 57.2 59.2 9.1 45.8 23.0 42.3 47.0 11.9 72.2 63.8 55.2

7.3 11.2 10.0 15.5 1.7 7.9 3.2 59.8 10.7 10.2 4.8 8.8 1.1 11.7 8.5 20.8 69.2 12.2 72.7 27.1 30.9 87.5 16.8 14.6 11.1

Table B2 Employment-weighted means of productivity, human and physical capital intensity, wages and size (employment), by ownership type and SIC80 Class: ownership type 4 (UK independents) = 100

Class 22 24 25 31 32 33 34 35 36 37 41 42 43 45 46 47 48 49 61 62

N 269 320 712 735 1883 206 1189 390 200 333 527 405 417 439 556 1566 800 227 7443 113

1 133.1 169.2 183.5 133.6 129.7 226.7 169.9 139.1 118.7 147.6 171.8 216.1 102.0 149.8 126.2 101.7 136.9 148.3 172.3 129.5

V/L 2 135.8 114.0 161.7 162.2 125.8 119.9 137.5 141.5 82.0 130.6 191.8 122.4 108.5 116.3 148.7 73.0 119.6 149.5 123.0 162.3

3 118.9 110.8 343.9 121.2 99.8 111.0 107.4 75.1 101.5 122.1 133.1 124.8 75.7 96.0 104.7 117.3 118.3 112.6 93.0 99.1

1 164.5 192.2 226.8 112.1 129.2 295.4 272.1 267.4 137.8 204.1 207.9 131.6 119.1 133.9 137.8 149.9 154.7 130.7 180.8 83.0

K/L 2 184.0 145.0 262.1 236.4 157.6 117.1 224.5 322.7 69.8 160.6 237.7 84.6 142.1 165.7 298.1 131.9 164.3 136.8 157.3 188.5

21

3 138.4 124.5 256.2 98.3 94.2 86.6 108.7 61.5 81.9 105.5 137.4 143.3 62.5 104.8 110.3 113.1 120.7 114.7 94.6 66.1

1 110.6 139.9 137.9 96.2 119.5 149.7 142.0 133.5 106.4 120.6 131.0 160.6 131.3 118.5 111.4 99.3 130.1 116.1 165.4 156.5

w 2 120.4 111.8 131.6 123.8 112.1 132.2 127.0 124.4 103.0 113.9 140.2 81.9 119.2 117.8 115.6 61.2 113.7 121.5 110.1 139.5

3 120.0 107.6 139.4 108.4 102.4 100.8 108.4 71.2 96.4 102.9 130.1 84.9 80.1 92.8 100.8 97.2 104.8 103.5 82.4 117.9

L 1 2 3 131.1 1080.1 572.1 698.8 1460.8 1301.2 1946.6 986.9 3034.4 2442.2 608.5 458.5 1658.6 1669.5 1692.3 7037.1 11941.0 577.9 1060.4 1346.8 4307.5 6063.3 8528.1 2484.6 43.6 345.1 1380.4 459.9 709.1 287.2 1077.4 2278.4 1081.7 2230.6 4231.7 15114.0 937.7 348.6 10512.4 392.3 550.8 3316.6 1186.5 362.8 607.0 62.4 2636.4 57.4 1163.5 1449.4 924.7 335.0 297.7 704.8 6855.7 6695.5 5601.7 327.7 521.7 770.8

Table B2, continued Class 63 64 65 66 72 74 76 77 79 81 82 83 84 85 92 97 Key

N 307 895 2245 880 753 73 107 936 89 339 115 4037 376 570 244 812

1 81.9 91.3 139.0 72.2 142.1 284.8 80.7 95.0 120.7 77.0 167.4 198.9 107.3 . 291.3 213.3

V/L 2 151.5 116.2 128.9 110.4 101.4 127.8 105.7 98.4 88.7 86.1 117.6 165.0 205.7 309.4 228.3 133.3

3 84.1 115.0 80.9 107.1 79.0 149.8 129.4 110.9 370.0 60.9 135.1 122.6 164.4 177.0 120.1 149.9

1 142.8 117.2 142.3 29.5 240.1 176.0 53.1 80.6 1863.4 79.0 399.5 209.0 106.6 . 798.2 234.9

K/L 2 239.2 167.2 188.8 154.4 56.1 53.7 100.2 99.4 3301.5 37.4 182.4 257.9 170.4 165.0 184.8 230.7

3 76.4 199.0 88.7 109.8 50.8 92.8 202.5 102.6 2485.1 31.3 131.8 179.4 190.7 277.7 156.2 135.6

1 82.4 92.3 81.6 77.7 117.7 186.8 118.1 101.3 158.6 86.4 231.5 149.5 78.0 . 252.8 106.8

w 2 140.8 84.5 117.2 86.1 104.3 102.0 108.3 97.5 123.2 104.4 190.3 155.8 155.6 122.6 188.6 87.7

3 1 89.3 30651.9 96.7 136.6 76.0 795.9 98.8 5079.9 90.0 1296.5 100.5 13.9 102.6 85.0 101.4 889.9 129.9 167.0 80.2 2642.6 187.1 445.6 119.1 2277.6 116.3 1670.1 122.0 . 113.0 1023.0 84.2 396.6

L 2 3 1795.7 601.1 3371.4 3341.3 1039.6 10671.0 6064.3 2934.8 3168.8 6550.2 383.5 261.3 65.2 342.6 2169.6 1046.1 223.0 1044.2 11129.8 77177.9 2004.3 6778.4 702.4 648.6 1332.3 1368.9 829.8 435.0 1409.0 1073.3 1120.9 1747.5

Ownership type 1: US-owned subsidiary Ownership type 2: Other foreign-owned subsidiary Ownership type 3: UK-owned subsidiary Ownership type 4: UK independent without subsidiaries N: number of companies.

Note Divisions 0, 1 and 5 and Classes 21, 23, 26, 44, 67, 71, 75, 91, 93, 94, 95, 96 and 98 excluded since the number of foreign companies was low. See Appendix A for the names of the Classes.

22