Productivity, product quality and workforce skills: food processing in four European countries.

Mason, Geoff ; Ark, Bart van ; Wagner, Karin 等

'borrowing' from other countries. Indeed, all three of the Continental countries examined here are presently undertaking reforms of their own in an effort to improve the workings of their respective vocational education and training systems. Our comparisons suggest that, in order to compete successfully with other advanced industrial nations, Britain needs a coherent set of policies which build on existing institutional structures, for instance, by raising average levels of attainment in core subjects in secondary schools, expanding the provision of full-time and part-time vocational education courses to recognised standards and strengthening the links between vocational colleges and employment-based training schemes.

Appendix A. Estimates of relative 'quality-adjusted' productivity levels 1. Introduction

During most of the 1980s Britain's productivity performance improved sharply in relation to its leading Continental competitors but the differentials remain substantial: recent estimates based on Production Census data show gaps in manufacturing output per person-hour ranging from some 20 to 40 per cent between Britain and Germany, France and the Netherlands (van Ark, 1990a, 1990b; O'Mahony, 1992). It is therefore of continuing interest to examine the reasons for this shortfall in British productivity levels and to assess the implications for British manufacturers' ability to respond quickly and effectively to competitive pressures in international product markets.

In previous studies covering a range of industries, National Institute researchers have compared productivity levels and the quality and utilisation of physical and human capital inputs in matched samples of manufacturing plants in Britain and Germany, and Britain and the Netherlands. In all cases these bilateral comparisons pointed to important links between relative productivity performance and workforce qualification and skill levels (Daly, Hitchens, Wagner, 1985; Steedman and Wagner, 1987, 1989; Mason, Prais, van Ark, 1992).

The present study reports on a detailed comparison of productivity, machinery and skills in matched samples of plants in a single industry--food processing--in four countries: Britain, Germany, the Netherlands and France. (Unless otherwise stated, the term 'Germany' refers throughout to the former Federal Republic.) By extending the range of inter-country variation in this way, it was hoped to deepen our understanding of the effects on relative productivity performance of different national systems of human capital formation: in relation to Britain all three Continental countries are distinguished by higher proportions of vocationally-qualified personnel in the workforce; however, in contrast to the well known 'Dual System' of apprenticeship training in Germany, initial training in the Netherlands and France is largely based on full-time vocational schooling.

Food processing is a prominent example of 'light' manufacturing in which the productivity gap between Britain and most other leading industrialised nations has historically been found to be low relative to other branches of manufacturing (Prais, 1981; Broadberry and Crafts, 1990; Broadberry and Fremdling, 1990). It has therefore been regarded as an industry in which Britain has a comparative advantage or at least in which its comparative disadvantage is relatively small. However, plant-sizes in British food processing are typically larger than in most other industrial nations (even including the United States), and questions have consequently been raised as to why scale-economies have not produced still greater benefits in respect of British productivity performance (Prais, 1981; NEDO, 1982; Maunder, 1988).

In the course of earlier comparative studies, suggestions have accumulated that British manufacturing's 'productivity problem' relative to other advanced industrial nations might consist not just of lower (physical) quantities produced per person employed but also of lower average product quality levels. In Anglo-German comparisons of kitchen furniture and clothing manufacturers, for instance, it was noted that British producers tended to specialise in more standardised and much lower value-added product areas than their German counterparts (Steedman and Wagner, 1987, 1989).

'Quality' of output is notoriously difficult to measure. In consequence, for the present investigation a relatively simple branch of the food industry, namely biscuits, was chosen for examination in order to facilitate productivity comparisons which took explicit account of inter-country differences in the mix of product-qualities produced--and to provide a basis for detailed study of the links between human capital endowments and the pre-dominant product strategies in each country.

Additional reasons for selecting biscuits as a specimen product area included their characteristics of being inter nationally-traded goods produced in competitive market conditions in all four countries. In terms of its basic technology, biscuit-making is representative of 'process' industries which employ continuous-flow methods of production. In such industries workforces are typically divided between technical staff responsible for the smooth running of automated linked machinery and a relatively large proportion of lower-skilled workers engaged in routine loading and packing activities. Comparisons based on this type of industry were expected to provide a valuable contrast with earlier studies based on more craft-intensive industries.

In all we visited some 29 biscuit manufacturing plants--ten in Britain, eight in Germany, six in France and five in the Netherlands. There were several problems to be overcome in drawing up four samples of plants which could be usefully compared with each other and at the same time be regarded as adequately representative of each national industry. In broad terms our sampling strategy was to cover a spread of plants in the inter-quartile employment-size range in each country.(1) As shown by national Censuses of Production, the median plant size in the British industry is substantially larger than in the other three countries; in order to obtain a substantial overlap of plant-sizes in the four samples, additional visits were made to British plants in smaller size groups and to German plants above the upper quartile plant size. However, for purposes of data analysis the sample plants were weighted in such a way that our estimates remained broadly representative of the inter-quartile employment size range in each national population. The effects of this adjustment on sample median plant sizes are shown in Table 1, Note a; Appendix A provides details of the weighting procedure used and compares the productivity estimates derived using this method with alternative estimates based on unweighted data.

In all four countries the plants were initially identified through trade directories. Further information about employment and product lines was sought by telephone before formal requests for visits were made. Response rates in each country were fairly similar with approximately two thirds of plants who were approached about a visit agreeing to participate. All the visits were carried out between October 1989 and December 1991. During the visits semi-structured interviews were held with production and/or personnel managers and, where possible, with shopfloor supervisors as well. Each visit included direct observation of production processes and work organisation on the shopfloor. At least two researchers from different countries were present on 90 per cent of visits. To economise on travelling costs the plants were geographically clustered in all the countries except for the Netherlands (where distances are small).

The order of discussion in this article is as follows: in Section 2 we compare the present structures and recent economic performance of each of the four countries' biscuit industries. Section 3 reports on detailed comparisons of productivity levels and the quality-distribution of output in each national sample of plants. Section 4 describes important inter-country differences in the quality and utilisation of physical capital inputs. In Section 5 we examine the contribution of workforce skills to relative productivity performance and to differences in each national industry's ability to compete successfully in high value added product areas. Section 6 summarises the central findings and public policy implications of the study.

2.Industry structure and performance

The term 'biscuits'--from the French bis-cuit meaning 'twice-cooked' --covers a range of pastry goods which have the common characteristic of being dry, brittle or crisp (in contrast to cakes which retain a relatively high moisture content after being baked). As a result of their relatively long shelf-life, biscuits are far more likely than other baked products to enter international trade.

Since the initial development of biscuit-making machinery in the early-19th century, the industry has gradually shed its craft origins and engaged in ever more technically complex methods of factory production. Some small craft bakeries remain--and indeed some new ones have developed in recent years in response to new market opportunities for specialist kinds of fresh biscuit--but the present structure of biscuit manufacturing strongly reflects the impact of competitive pressures over many decades which have inspired amalgamations of producers (through mergers and takeovers) and concentration of production in a relatively small number of highly mechanised--in some cases highly automated--plants.

The pre-eminent supply-side factor stimulating this reorganisation process has been the wide scope for technical economies of scale in industrial baking (Pratten 1971) and the opportunities for scale-economies in bulk-buying of materials and in marketing and distribution costs. On the demand side amalgamations of biscuit manufacturers have proceeded in tandem with wide-scale advertising of individual product varieties and brand names which, for some leading firms, has helped create and sustain mass markets for long production runs of standardised biscuit varieties.(2)

Restructuring of biscuit production and its consolidation in fewer, larger plants has taken place in all four of the countries examined here. However, the process began earlier and has developed most fully in the British industry where, as shown in Table 1, average plant sizes are substantially larger than in the other three countries. This outcome can be attributed in part to the faster and more extensive development of nationwide advertising in Britain and the country's earlier development also of a Stock Exchange for industrial securities (facilitating the acquisition of smaller producers by larger firms) (Prais, 1981). All four national industries are now dominated by large enterprises with access to industrial capital markets and indeed some of these leading concerns are multinationals with plants in several European countries, but family-owned firms continue to play a more important role on the Continent (and especially in Germany) than they do in Britain.

Continued rationalisation involving plant closures and reductions in employment has dominated the British biscuit industry's performance in the last decade. Between 1980-90 output rose in Britain by only 16 per cent but, with employment contracting by some 37 per cent over the same period, there was a very sharp increase in labour productivity of 84 per cent, higher than in any of the other three countries. As Table 3 shows, output grew much more rapidly in (West) Germany (up 73 per cent) and in the Netherlands (up 43 per cent) than in Britain and (in spite of fluctuations during the decade) employment levels in 1990 were only about 4 per cent down on 1980 levels in each of those two countries. The German performance in part reflects a remarkable 25 per cent growth of output in a single year (1990) when the domestic market was suddenly expanded by reunification; in the same year German employment rose by some 11 per cent over 1989. In the case of the French industry, output TABULAR DATA OMITTED and productivity growth over the decade was the slowest of the four countries but there were signs of improvement in the late-1980s; the percentage decline in French biscuit employment was about half that in Britain.

Table 1. Distribution of plant sizes in national samples of biscuit
manufacturers

 Britain Netherlands Germany France
 Number of plants

No. of employees
Under 200 2 2
 }4 }4
200-499 2
 }3
500-999 2 2 2
1000-plus 4 0 2 0
Total 10 5 8 6

 Number of employees

Weighted median plant
size in sample(a): 1170 280 350 380

Notes:

(a) The median size is here defined such that half of all employees in each
sample are in plants above that size and half below it. For details of
weighting procedure, see text and Appendix A. The unweighted sample median
plant sizes were: Britain 1040, Netherlands 340, Germany 1000 and France 410.

The faster rates of output growth in the German and Dutch industries during the 1980s partly reflect rapid growth in export volumes in both countries. The relatively high export share of production in the Netherlands (43 per cent of output in volume terms in 1990; see Table 4) is typical of a small open economy. The 38 per cent export share in Germany is markedly higher than in Britain or France which are of similar population size. The British market is characterised by a small (6 per cent) but rising import share of consumption, in particular of higher-priced varieties.(3) The French industry is the only one of the four to run a trade deficit with imports accounting for 27 per cent of consumption while its export ratio is much the same as Britain's.

Table 3. Estimated percentage changes in output (volume), employment and
labour productivity in national biscuit industries, 1980-90

 Percentage change, 1980-90

 Britain Netherlands Germany France

Output (volume) +16 +43 +73 +14
Employment -37 -4 -4 -18
Output per employee +84 +49 +80 +39

Sources: Britain: Output: CSO, Annual Abstract of Statistics and BCCCA Annual
Review, various issues. Employment:CSO, Report on the Census of Production and
Size Analyses of UK Businesses, various issues.

Netherlands: Output: Produktschap voor Granen, Zaden en Peulvruchten.
Employment: CBS, Produktiestatistieken, various issues and Produktschap voor
Granen, Zaden en Peulvruchten.

Germany: Output: SB, Statistisches Jahrbuch, various issues. Employment: SB,
Produzierendes Gewerbe: Kostenstruktur der Unternehmen and Statistisches
Jahrbuch, various issues.

France: Output: INSEE, Annuaire Statistique de La France, various issues and
L'Alliance 7 (Enquete de Branche). Employment: INSEE, Annuaire Statistique and
Ministere de l'Agriculture et de la Foret, Agreste: La Statistique Agricole.
Table 4. Biscuit export and import (volume) shares, 1980 and 1990

Percentages

 Exports/production Imports/consumption

 1980 1990 1980 1990

Britain 12 14 2 6
Netherlands 31 43 12 22
Germany 17 38 29 29
France 10 15 15 27

Source: International Office of Cocoa, Chocolate and Confectionery (IOCCC),
Statistical Bulletin, Brussels, 1991.

3. Productivity and product quality

International differences in labour productivity levels reflect variations in both the average (physical) quantity of output produced per employee (or per employee-hour) and the average quality level of output. As noted in Section 1, the biscuit industry was selected for the present investigation in the hope that the relative simplicity of the product would facilitate comparisons of real productivity levels (taking explicit account of quality--as well as quantity-differences between the four countries).

3.1 Tonnage per employee-hour

In the course of our plant visits we gathered detailed information on biscuit output and associated employment--both direct and indirect--over a recent 12 month period (a sufficient length of time for our calculations not to be distorted by seasonal fluctuations). In all cases we confined our attention to the production of a strictly-defined category of biscuits including, for example, plain, sweet, semi-sweet, savoury and chocolate-coated biscuits but excluding related products such as wafers, waffles, rusks and (in Germany) Lebkuchen as well as any cake or confectionery products. Wherever biscuits were not the sole product of manufacture, guidance was sought from production managers as to how the indirect labour input should be allocated between biscuit and non-biscuit products.

As a first step in the comparison of labour productivity levels, a simple measure of output (tonnage) per employee-hour was calculated which took no account of quality-differences. By this measure the highest (weighted) average level of productivity was found in our Dutch sample of plants, about 15 per cent ahead of Britain. The average output per unit of labour input in the British plants was only 5 per cent lower than in the French sample and was some 25 per cent higher than in the German plants (even in spite of the above average levels of capacity utilisation prevailing in Germany at the time of our visits).(4)

This relatively good performance of the British biscuit industry primarily reflected the contributions made by its very largest plants which produced in excess of 1000 tons per week and achieved higher levels of labour productivity in terms of (crude) output per person-hour than any of the Continental plants visited. In smaller and medium-sized British plants producing similar tonnages to the great majority of their Continental rivals (between 100 and 500 tons per week), productivity levels in Britain were considerably lower than in the other three countries.

As shown in Table 5, our sample results are consistent only in a very broad sense with recent Production Census-based estimates of productivity differentials in biscuit-making. The Census data cover a wider and more variable range of products in each country--for instance, in the Netherlands biscuits are not distinguished from cake (banket) products--and also include some employees involved with non-biscuit products. Nevertheless, taken together, the two sets of estimates support the conclusion that the British biscuit industry is a comparatively successful industry in international terms. The productivity gaps in biscuits relative to the Netherlands and France are smaller than the estimated differentials for all food processing and for manufacturing as a whole (Van Ark, 1990a, 1990b). In the Anglo-German comparison the higher level of output per employee-hour in British biscuit-making contrasts with a small productivity gap in the German favour in the wider food, drink and tobacco industry, and with an estimated 20 per cent advantage in productivity for German manufacturing as a whole (O'Mahony, 1992).(5)

3.2 'Quality-adjusted' output per employee-hour

In the course of our visits we noted marked differences in the mix of biscuit-qualities produced by each national industry which appeared relevant to comparisons of real productivity levels. For example, the more basic (undecorated) varieties of biscuits accounted for a larger proportion of total output in the British sample than in the other three countries and this was particularly the case in relation to the German sample. As well as requiring less secondary processing and packaging, basic-quality biscuits are typically produced in large batches and are thus more amenable to automation of production than are the higher-quality grades of biscuit.

Table 5. Estimates of productivity levels in biscuit manufacturing(a)

 Output (tons) per employee-hour

 NIESR samples National Production
 (1989-91)(b) Censuses(1990)(c)

Britain 100 100
Netherlands 115 140
Germany 80 75
France 105 125

Notes:

(a) Index numbers: Britain = 100; rounded to nearest five.

(b) Relates to comparable narrowly-defined ranges of biscuits and associated
labour inputs.

(c) Wider coverage of products, varying from country to country, and including
some employees involved with non-biscuit products. The Census product
definitions in each country are as follows:

Britain (Biscuits and crispbread): Rusks, crispbreads and matzos; savoury
biscuits; chocolate covered biscuits; sweetened biscuits; semi-sweetened
biscuits; unsweetened biscuits--plain.

Germany (Dauerbackwaren): Zwieback; Leb- und Honigkuchen, Printen; Hart-und
Weichkeks; Waffeln; Gefullte Riegel; Salz- Kase- und Laugen Geback; sonstige
Dauerbackwaren.

Netherlands: Biscuits, wafels e.d.; Koekjes, banket e.d.

France (Biscuiterie-biscotterie): Biscuits sales; biscuits secs; gaufres et
gaufrettes; biscuits patissiers; autres biscuits divers; patisseries de
conservation; pains d'epices; biscotterie.

Sources for Census-based estimates: Employment and annual total sales value of
biscuits produced: CSO, Report on the Census of Production; SB, Produzierendes
Gewerbe: Kostenstruktur der Unternehmen; CBS, Produktiestatistieken; INSEE,
Annuaire de Statistique Industrielle.

Average ex-factory sales value per ton of biscuits: CSO, Quarterly Sales
Enquiry; SB, Statistiscbes Jahrbuch ber Ernahrung, Landwirtschaft und Forsten;
CBS, Produktiestatistieken; INSEE, Enqute Annuelle d'Entreprise (SCEES).

Annual hours worked per employee: based on estimates for food and drink
industries in each country in B. van Ark, International Comparisons of Output
and Productivity, University of Groningen Growth and Development Centre,
Monograph Series, No. 1.

The detailed information we had gathered enabled us to classify the outputs of all plants visited into three different grades of product defined in terms of technical characteristics such as the number of processes involved in their production and the types of ingredients and packaging materials used.(6) The three grades of biscuit quality were defined as follows:

1. Basic-quality: 'dry' undecorated biscuits made from relatively cheap ingredients (for example, vegetable or animal fats), stacked in cylinder-form and wrapped in a single layer of plastic film.

2. Medium-quality: biscuits requiring at least one secondary production process after baking such as chocolate-coating or sandwich-filling with cream or jam; usually sold with at least two layers of packaging. Undecorated biscuits made with more expensive ingredients such as butter are also included in this category.

3. High-quality: elaborate, multi-textured biscuits requiring two or more secondary (post-baking) production processes, usually stacked in successive small piles, multi-wrapped and boxed; includes assortments of biscuits made from expensive ingredients.

The outcome of this classification of output by quality-grade is shown in Table 6, Part (i). The 35 per cent share of basic-quality biscuits in British output is well above the 15-25 per cent shares found in Germany, France and the Netherlands. Some 35 per cent of German sample output is classified to the high-quality grade, a much higher proportion than in the other three countries. A very large share of production in all four countries (ranging from 50 per cent in Germany to 70 per cent in France) falls into the medium-quality grade.

In respect of ingredients, dimensions, packaging and other physical characteristics, the products classified to the basic-quality grade in each country were remarkably similar. (Indeed, according to a panel of consumers at the Institute, this basic similarity of basic-grade varieties applied also with regard to the rather more subjective aspect of 'taste'). However, products classified to the medium- and high-quality grades were more diverse in nature: for example, many of the German biscuits classed as 'medium' grade would be regarded as 'top of the range' products in the other three countries; similarly, some German high-quality biscuits were more elaborate in their production and packaging than any seen being made elsewhere.

As a first step towards the calculation of a 'quality-adjusted' measure of output per employee-hour in each industry, Institute researchers gathered data on retail prices for the different grades of biscuit in each country, based on products observed in the plants visited; they are shown in Table 6, Part (ii) in the form of price-ratios with the basic grade set at 100 for each country (all calculations were based on retail prices net of VAT). Medium-and high-quality biscuits (as we have defined them) were found to sell in Germany on average for, respectively, about 3.1 and 6.6 times the cost of basic-quality biscuits; in Britain, France and the Netherlands the equivalent price multiples ranged up to 2.5 and 4.2 respectively.

These inter-country differences in the implied relative valuation placed by consumers on each quality-grade are broadly consistent with the point made above that many biscuits classified to the 'medium-quality' grade in Germany would be regarded as highly elaborate products and sold at 'premium' prices in the other three countries. This was confirmed by our own comparisons of the prices of closely-matched biscuit varieties which had been produced and purchased in each country: for example, British-made biscuits in the upper half of the medium-quality price-range were TABULAR DATA OMITTED typically best matched (in terms of the technical characteristics outlined above) with German biscuits priced in the lower half of that country's medium-price range; similarly, most 'high-quality' British products could only be compared with German biscuits near the bottom of the German high-quality price range.(7)

In order to obtain estimates of the average value-added per ton of output in each national industry, the following three-stage procedure was carried out (see notes to Tables 6 and 7 for details):

1. An index of the average retail sales value per ton produced in each industry was calculated by weighting the retail price-ratios for each country by the proportions of output in each quality-grade (Table 6, Part iii).

2. This index of average retail prices was then converted to an index of average ex-factory prices on the basis of estimated average retail gross margins, exclusive of VAT, in each country.

3. An index of value-added per ton was then derived by recalculating the index of average ex-factory prices on a net of raw materials' basis using Production Census data on inter-country differences in the ratio of materials costs to the total ex-factory sales value of goods produced in biscuit manufacturing.

TABULAR DATA OMITTED

The approximate nature of this measure of 'value-added' will be apparent: amongst other things, it makes no allowance for the costs of industrial and non-industrial services (many of which would apply to the purchase and sale of merchanted goods, not just the biscuits actually produced in our sample plants). In addition, the numerical precision of our estimates could have been improved if even more time and resources had been devoted to data gathering. Our calculations must therefore be regarded as no more than a first estimate of the adjustment needed to take account of inter-country differences in the mix of product-qualities when comparing real productivity levels in this industry.

As shown in the bottom row of Table 7, our estimates suggest that average value added per ton of biscuits in Britain is (in real terms) some 40 per cent lower than in Germany and some 10-15 per cent lower than in France and the Netherlands. When this value-added index is combined with our earlier measures of (crude) output per employee-hour in each national industry, 'quality-adjusted' productivity levels in Britain are found to be some 30 per cent below those in Germany, 20 per cent below the Netherlands and 15 per cent below France.

These results confirm the significant contribution of inter-country variation in the mix of product-qualities to international differences in real productivity levels in biscuit manufacturing. In relation to the British industry, the German productivity advantage derives from the higher average quality of its products which is more than sufficient to offset its lower physical quantity produced per employee-hour.(8) Quality differences are also the predominant element in the productivity gap between Britain and France. The exception in this regard is the Dutch industry which succeeds in out-performing Britain primarily through its more efficient production of medium-quality biscuits.

The estimated differences in 'quality-adjusted' labour productivity levels between Britain and the three Continental countries are broadly mirrored in the pattern of inter-country variation in hourly labour costs: in 1989-91 average labour costs per employee-hour in the food processing and related industries were some 15 to 35 per cent lower in Britain than in the other three countries.(9) The implications of such labour cost differentials for relative price competitiveness are discussed below in Section 6. In the remaining sections of this paper we examine differences in the quality and utilisation of physical and human capital inputs which might help to explain the clear pattern of inter-country difference in real productivity levels.

4. Machinery and production organisation

4.1 The production process

Modern biscuit manufacturing typically employs continuous-flow methods of production with primary ingredients being bulk-fed to mixing machines from where the dough passes through rolling and biscuit-shaping ('forming') machinery before entering conveyor-fed 'travelling' line ovens. In the subsequent post-baking stages of production the biscuits are conveyed by moving belts through secondary processing machines if required (for example, for creaming, chocolate coating or jam-filling) and then to the final stages of wrapping and packing.

The wrapping process is invariably automated but the prior collating and positioning of the biscuits so they can pass smoothly through the wrapping machines is often carried out by hand, as is the subsequent packing of the biscuit packets in cartons ready for transportation to the warehouse. The labour-intensive nature of packing work reaches a peak in those factories which prepare tins of assorted biscuits.

The two main variants to this pattern of production are as follows: firstly, a few plants--those with relatively long production runs of standardised biscuits-- run the entire mixing process on an automated continuous basis; however, the majority of plants visited in all four countries engaged in batch mixing on all their production lines with regular 'changeovers' of dough-forming, wrapping and other equipment further down the line(10); secondly, on production lines where changes of product varieties and packet sizes are relatively infrequent, vibratory TABULAR DATA OMITTED systems or other equipment may be installed to stack the biscuits gently in lanes and feed them into automatic collating machines prior to wrapping; in some cases post-wrapping activities such as cartoning and palletisation may also be automated.

4.2 Age of machinery

Investment in new biscuit-making equipment can enhance productivity performance in several different ways. For instance, each new generation of sophisticated ingredient handling systems permits faster and better controlled distribution of ingredients to mixing machines. New mixing machines tend to be larger and faster than older machines and new ovens are usually both wider and longer than their predecessors, permitting a faster throughput of product. Other advantages of new ovens relate to their controllability with shorter start-up times at the beginning of each day and faster responsiveness to adjustments made by operators; this greater accuracy in control reduces product wastage and permits faster changeovers between product varieties. Technical advances in wrapping machinery continue to improve performance in terms of speed, reliability and ability to undertake complex operations.

In the British, German and French plants visited, the average age of ingredient handling, mixing and baking machinery ranged between 12 and 16 years compared to just under 20 years in the Dutch industry. All four industries had renewed much of their wrapping machinery in the last ten years, reflecting both the heavy wear and tear to which such equipment is subjected and the need to keep up with the rapid pace of technical change.

In preparation and mixing areas the fastest rate of new investment was found in the British plants with about half of all equipment purchased in the last ten years compared to 40 per cent in Germany and 30 per cent in France. However, the German and French samples had both installed slightly more new ovens than Britain over the same time period. In both respects the Dutch plants were well behind with widespread spare capacity leading to only 10 per cent of all process equipment being replaced in the last ten years.

Differences in the age and sophistication of capital equipment in the four industries therefore showed little relation to the inter-country differences in productivity performance outlined in Section 3. In particular there appeared to be no great disadvantage for the Dutch sample in operating with older mixing machines and ovens so long as their control systems had been updated. In all four countries, and particularly in the Netherlands, we observed process equipment aged in excess of 30 years which was running smoothly; conversely, and particularly in Britain (as discussed further below), new and sophisticated machinery was sometimes visibly malfunctioning. In wrapping departments, as mentioned, heavy investment in new machinery has occurred in all four industries but even here we noted the importance--most evident in Germany--of 'low-tech' modifications to old equipment designed to maintain a smooth flow of production; detailed examples of such incremental improvements are cited in Section 5.5 below.

4.3 National origin of machinery

In the British and German firms visited, ingredient handling and mixing equipment and tunnel ovens were largely of domestic origin; in France about half of process equipment was French-made and half sourced from other European countries including Germany and Britain; in the Dutch industry German suppliers accounted for most (about four fifths) of process equipment and British firms for the remainder. The heavy reliance on domestic suppliers in three of the four countries--and on nearby German suppliers in the Dutch case--reflects a preference for ease of access to after-sales, maintenance and repair services; however, this has become less of a priority in recent years as improvements in transport and communications have enabled the leading European manufacturers of food processing equipment to improve services to foreign customers.

In contrast to the continued availability of high-quality British-made baking machinery, there are hardly any TABULAR DATA OMITTED remaining British producers of complex, high-speed wrapping equipment and the great majority of wrapping machines used in British biscuit plants were imported from Switzerland or Italy. In the German, Dutch and French industries there were still significant proportions of locally-made wrapping equipment but the leading Swiss and Italian suppliers accounted for about three quarters of the machines used in France, just over half in the Netherlands and about a third in Germany.

4.4 Machinery utilisation

Capital equipment was operated most intensively in the British biscuit industry with nearly all plants visited engaging in full or partial 24-hour working, albeit usually for five rather than seven days per week. In France half the plants had 24-hour operations and half had two shifts per day. In Germany most plants operated only two shifts per day, in part because of legal restrictions on the employment of women (packing workers) during night shifts. In the Netherlands spare capacity limited production to a single shift per day in the majority of plants visited (with some evening work in the high season).

These inter-country differences in machine utilisation affected labour productivity performance in several ways. Where production lines were largely dedicated to single products--with few changes in packet sizes--there were clear benefits in continuous multi-shift working; examples of this type of production were observed in all four countries but were most common in Britain. However, most producers were under pressure to respond quickly to diverse and rapidly-changing customer requirements, with even ostensibly similar products having different specifications in terms of ingredients, dimensions, packaging and so on. In this context plants making less intensive use of equipment were better placed to schedule major changeovers of machinery for times when production lines were not fully manned (for example overnight); conversely, plants working a 24-hour day were less able to avoid inefficient use of labour while changeovers were being made.

4.5 Maintenance and repairs

Productivity levels in each industry were also affected by different rates of machine breakdown and malfunctioning which, in a continuous-flow production process, have serious consequences in terms of product wastage and the additional labour required to sort out problems. During our plant visits we observed some examples in all four countries of production lines which had been halted in mid-production or were not working smoothly; however, the incidence of equipment failure was highest in the British sample with the most common problems arising from hard-worked conveyor belts and wrapping and carton-sealing machines and, in some cases, from longstanding difficulties in getting complex new ingredient handling and mixing systems to perform to requirements.

These observations were borne out by records of 'emergency downtime' rates in the four samples: British plants reported an average 10 per cent of planned machine-working time lost due to breakdowns and other unexpected stoppages compared to only about 3-4 per cent downtime in the other three countries. This disparity was associated with different priorities in respect of preventative maintenance: only a fifth of the British plants visited reported serious efforts to carry out routine maintenance; by contrast, four-fifths of the Dutch and German plants had implemented full planned maintenance programmes as had two-thirds of the French plants.

As in other international comparisons carried out by the National Institute, the majority of British biscuit plants appeared to be trapped in a vicious circle with high levels of emergency maintenance militating against the introduction of preventative maintenance procedures which might help reduce the incidence of breakdowns. This dilemma partly reflected the pressure on British managers to work machinery harder and longer than on the Continent in order to ensure a rapid 'payback' on new investments. At the same time, as described below, high levels of emergency downtime in Britain were also associated with relative deficiencies in production management, supervisory and other workforce skills. In the next section we assess inter-country differences in human capital endowments in detail and consider their effects on productivity performance and the predominant choice of product strategy in each industry.

5 Workforce skills, productivity and product strategy

5.1 Occupational structure

Direct labour accounted for between 70-80 per cent of total employment in all four national samples with the great majority of shopfloor workers employed in post-baking (secondary processing, wrapping and packing) areas of production. The employees involved in the ingredient handling, dough-mixing, forming and baking stages of biscuit production are defined here as 'process workers': they include mixermen/women (and/or computer operators in a mixing control room), machine operators and ovensmen/women. On average there were approximately three wrapping and packing workers for every process worker in the British, French and Dutch samples and a 4:1 ratio in the German plants (reflecting the greater incidence of assortment packing in Germany).

Among indirect workers the largest single groups of staff in most plants were in engineering and maintenance departments (with responsibility for setting and re-setting of complex wrapping and packing equipment as well as dealing with repairs and servicing of machinery) and in administration and sales. Other key indirect functions were in technical support areas such as production planning (day-to-day scheduling), production or 'project' engineering (involved both with the planning of work methods and sequences and with the commissioning and installation of new equipment), new product development and the design and management of quality control and hygiene systems. In large multi-plant firms, product development is usually undertaken in separate research centres; however, several of the branch plants visited in each country had staff members involved from time to time in pre-production testing of new biscuit varieties or packaging materials.

5.2 Vocational qualifications and training

The highest levels of formal vocational qualifications were found in the German sample and the lowest in Britain, with the Dutch and French industries occupying an intermediate position in this respect. In all four countries wrapping and packing work was carried out by a largely unqualified workforce. The main differences arose in the more technically demanding occupational areas--process, engineering and technical support departments and senior management--where the proportions of staff qualified to craft level or above were some 85 per cent in Germany, 65 per cent in the Netherlands, 45 per cent in France and 25 per cent in Britain.

Inter-country differences in the qualifications and training of each occupational category are examined in detail in Appendix C. Broadly speaking, the pattern of qualifications in each sample reflected the institutional structure of vocational education and training provision in each country. Thus the majority of vocationally-qualified German employees had passed through the 'Dual System' which combines employment-based training with part-time attendance at vocational schools. By contrast, in the Dutch and French plants most vocational qualifications had been acquired in the course of full-time vocational schooling which for many employees had begun in the latter stages of compulsory education. In addition, the Dutch sample had relatively high proportions of employees in key occupations with technician-level qualifications, reflecting the widespread provision of vocational education at age 16-plus in the Netherlands. In the British case there was a mix of qualifications gained through employment-based apprenticeships and full-time vocational courses but, in common with most other branches of manufacturing, the proportions of employees with either type of qualification were relatively low by Continental standards.(11)

5.3 Process skills and product quality

In German plants the vast majority of production process workers were craft-trained bakers who had been recruited on the open market. In addition three quarters of German production supervisors had undertaken further training to Meister standard in a range of technical and management areas which had equipped them well to liaise with technical support services such as maintenance and to advise on incremental improvements to production processes.

By contrast, in the British plants visited there were no process workers and very few supervisors with vocational qualifications and some managers went so far as to say that they would not give priority to craft-skilled bakers or other qualified persons for production jobs even if such employees were readily available to them. Such differences in perceived skill requirements were closely related to the predominant product strategies in each of the two countries.

Process work in biscuit manufacturing largely centres around the achievement and maintenance of detailed product specifications in the face of considerable variability in raw material qualities (texture, moisture and so on) and differences in weather conditions (affecting ambient temperatures). In the course of production, variables such as the temperature and consistency of the dough mix and the diameter, thickness, weight, shape, moisture content and colour of the biscuits need to be repeatedly monitored. As conditions change from day to day (or batch to batch) it may be necessary to make a series of alterations to the dough-mixes initially specified by standard recipes and to adjust forming machine-settings, conveyor belt speeds and oven temperatures (Manley, 1991).

The skills and knowledge required for this work rise sharply as the number and complexity of products made is increased. At the same time small batch sizes and the use of expensive ingredients reduce the scope for protracted 'trial runs' with individual varieties before full production begins. In consequence the several German plants specialising in small- and medium-batch production of high-quality multi-textured biscuits greatly depended on the 'know-how' and experience of the craft-trained bakers in their process departments.

By contrast, in the British industry the designation of process work as 'semi-skilled' reflected its primary orientation towards longer runs of relatively simple types of biscuit. In this type of production the need for judgement and skill on the part of process workers is much reduced. This is particularly the case in factories where manual monitoring and adjustment activity has been supplemented or replaced by computerised control systems. However, even in these circumstances there is a recurrent need for some manual intervention to maintain a consistent product and wastage may occur if operators do not fully understand the consequences of their actions or inaction (Buchanan and Boddy, 1983, Ch. 11): in highly automated British factories specialising in bulk production of basic-quality biscuits, it was still considered necessary for process workers to 'stick to their own jobs' on particular production lines in order to cope with possible problems that might arise even with ostensibly uncomplicated products.

In relation to these German and British extremes of product strategy and process skill requirements, the Dutch and French samples both occupied intermediate positions. As outlined in Section 3 above, the average value added per unit of output in the Netherlands and France was higher than in Britain but was still substantially below the average unit-value of German output. Both the Dutch and French samples had significant proportions of vocationally-qualified process workers and supervisors but few of them were apprentice-trained bakers as in Germany; this deficiency had been partially offset by longer average periods of on-the-job training than occurred in the British industry.

5.4 Shopfloor manning levels

In the British plants visited the mixing and baking stages on each individual oven-line were usually covered by a three-person team consisting of a mixing operator, forming-machine operator and ovensman/woman. By contrast in all three Continental industries similar teams of process workers typically had responsibility for at least two oven-lines at a time and had been trained to switch flexibly between different task areas as required. In some larger British plants the deployment of process workers was restrained by long-established manning agreements with unions, on which slow negotiations were in progress. However, doubts were expressed by several British employers and union representatives as to the ability of narrowly-trained process workers to deal with problems on more than one oven-line at a time.

In post-baking areas, manning levels were naturally highest in plants engaged in small- and medium-batch production of biscuits requiring multiple secondary processes (chocolate-coating, cream-filling, and so on) and subsequent packing in assortments. This type of product strategy was most commonly observed in Germany and (to a lesser extent) France; the associated high labour inputs were reflected in the higher average value added per unit of output in those two national samples.

In British plants geared to basic- or medium-quality product strategies, manning levels in wrapping and packing areas were frequently higher than in comparable Continental plants but did not contribute positively to value added: additional personnel in these British plants were typically required to sort out problems caused by higher rates of machine breakdown and malfunctioning, for example, to stack biscuits temporarily while machines were being repaired, to unwrap packets which had not passed properly through wrapping machines or had been rejected by check-weighing machines and to sweep up biscuits which had spilled onto the floor.

5.5 Engineering skills and production flow

In all four countries maintenance staff were generally required to hold at least a craft-level engineering qualification but this minimum standard had only been fully met in the German and the Dutch plants (indeed in the latter country roughly half the maintenance personnel were qualified to technician level). In both Britain and France about a fifth of engineering workers were vocationally unqualified; in France this deficiency was partially offset by recent recruitment of engineering staff with technician-level qualifications and by the relatively large size of French maintenance departments.(12)

Some of the reasons for the relatively high rates of emergency downtime and product wastage in the British sample were conveyed by managers' assessments of their engineering departments: in one plant we heard of maintenance staff being 'slow to respond to problems' and failing to deal with problems in such a way as to prevent their recurring; in another the introduction of a preventative maintenance policy was ruled out on the grounds that 'jobs might take too long' or not be done properly; in a third case where a large proportion of maintenance workers had been recruited from the shopfloor, some of them were even described as 'failed bakers'. Some British plants also reported residual problems of demarcations between maintenance and process workers and, within maintenance departments, between mechanical fitters and electricians which had no parallels in the German, French and Dutch samples.

As noted earlier (Section 4.5), in most Continental plants there was a clear link between relatively low rates of emergency downtime and the implementation of planned maintenance programmes. Concerted efforts to maintain a smooth production flow were most evident in German plants, many of which had an essentially 'pro-active' engineering policy: apart from undertaking preventative maintenance, engineering staff worked in close collaboration with production supervisors on machinery development and adaptation with a view to speeding up production and eliminating bottlenecks and delays.

Typical examples of this methodical approach to problem-solving in German plants included: the installation of turn-tables for the temporary absorption of biscuits if a machine was halted further down the line; modifications to standard wrapping and case-packing equipment designed to reduce both re-setting and maintenance times; the use of overhead cranes to allow flexible reconfiguration of conveyor belts and wrapping machinery; and the in-house development of 'pick up and place' equipment to help biscuits pass through multiple secondary processing operations. According to one German manager, it could take 'up to ten years' of incremental improvements to achieve the desired throughput on capital equipment.

Although some examples of efficiency-enhancing innovations were also observed in the other three countries, they were much less common than in Germany. The German sample appeared to gain a substantial advantage from the combination of technical competence and detailed knowledge of production conditions embodied in its Meister-trained supervisors. The Dutch and French plants were able to match their German rivals in minimising emergency downtime through preventative maintenance; the majority of Dutch plants benefited from the diagnostic skills of a significant proportion of technician-qualified engineering staff and several French plants had tackled production flow problems by deploying groups of skilled engineering workers to work full-time in shop-floor wrapping and packing departments. However, the majority of Dutch and French supervisors were much less well-qualified than German Meister to collaborate systematically with engineering specialists on process innovations.

In the British industry, as already noted, the great majority of production supervisors lacked any form of technical or other vocational training and engineering departments were largely absorbed in dealing with emergency repairs. In some cases these problems required the regular attention of highly-qualified managers and project engineers who were thus diverted from their primary tasks of making long-term improvements in processes and products.

5.6 Flexibility in production planning

A key element in the efficient maintenance of production flow is the sequencing of different product batches to minimise the number of changeovers required and the time needed to carry them out. Roughly half the German plants had allocated responsibility for weekly and daily production scheduling to vocationally-qualified staff, either office-based personnel with technical or business administration qualifications, or Meister-trained supervisors who combined it with their other duties on the shopfloor. In the other German plants, as in most British, Dutch and French plants, production planning was carried out by experienced but unqualified employees or (in small plants) by the production manager himself.

In spite of this broad similarity in the qualifications of planning personnel in Britain and the majority of Continental plants visited, it was much more common in British plants to hear complaints about the 'pressures' involved in adjusting production schedules in response to unexpected changes in customer orders. An ability to respond quickly to the needs of large supermarket chains was a priority in all four countries. As much as anything the problems reported in British plants appeared to result from wider skill deficiencies, for instance, the limited transferability of process workers between different products and tasks. In addition, several British managers mentioned restrictions on flexibility caused by the need for wrapping machines to 'settle down' after every product or packet-size change-over. Such difficulties were rarely mentioned in the other three countries; their prevalence in Britain was consistent with higher reported rates of machine malfunctioning in general.

5.7 Upgrading workforce skills

As a consequence of competitive pressures in domestic and export markets and the need to keep up with new technological developments, the majority of plants visited in all four countries had recently increased their expenditure on continuing and updating training for certain groups of employees. In Britain and Germany this mainly consisted of short external courses for maintenance and other technical support staff in areas such as electronics. Several plants in both countries had also taken advantage of technical courses or training materials for supervisors and process workers which were provided by their respective industry trade associations or by further education colleges. In the Netherlands the incidence of continuing training appeared to be smaller than in Britain or Germany; most of the Dutch plants visited expressed greater interest in upgrading their skills base by raising their recruitment criteria for supervisory and process positions and improving their initial training programmes.

In France since 1971 all companies have been legally obliged to spend a minimum proportion of their total wage and salary bill on continuing training. By 1988 the national average 'participation rate' was some 2.7 per cent--well above the legal minimum of 1.2 per cent--but with marked variations between enterprises in different employment size-groups and industrial sectors (Berton and Podevin, 1991). Recent declarations on training expenditure by the plants in our French biscuit sample showed a weighted annual average of 3.5 per cent of total wages spent on continuing training, above average for both the wider food processing industry and the size-groups represented in the sample. As in many other industries, there was a considerable variation in the training received by different occupational groups with about three quarters of 'technicians/supervisors' and 'engineers/managers' receiving some formal training each year compared to some 30 per cent of manual workers (including engineering as well as shopfloor employees). The reported 'expectation of training' (total training hours divided by total number of employees) was about 25 hours per annum for the two highly-qualified groups and 12 hours for manual workers.

No detailed comparison of French spending on continuing training with the other three samples can be attempted as employers in the other countries have no incentive to record and declare training activity as fully as do French employers. Nonetheless our observations and discussions during plant visits pointed to a broad conclusion that formal training for existing employees in the French sample was both more extensive and more innovative than in Britain, Germany or the Netherlands. Notable examples of continuing training initiatives more commonly, or only, found in French plants included:

* 'multi-skilling' training programmes in engineering departments designed to turn both mechanical fitters and electricians into electronics specialists (automaticiens)

* full adult apprenticeships leading to a formal craft qualification designed to turn female packing workers into skilled machine setters (regleuses)

* systematic technical and other training over two-three years for key process workers with a 'hands-on' supervisory role (responsables produits, responsables techniques)

This last example reflected a considered attempt by some French plants to develop supervisors akin to German Meister. However, at present it seems unlikely that French supervisory skills in aggregate will attain German levels; indeed, many of the supervisors in the German sample had received some further training over and above their earlier craft and Meister training.

Our overall assessment of training levels in the four countries can perhaps be summarised as follows: the relatively high volume of continuing training in the French industry--promoted by the legislative framework within which French employers operate--has helped raise French productivity levels above those in Britain and near to those of the Netherlands (partially offsetting the Netherlands' more extensive provision of initial vocational education and training). Although the French and Dutch industries are ahead of Britain, they still lag behind German producers in respect of both human capital endowments and productivity performance. The mix of initial and continuing training received by German employees continues to support a highly successful strategy of rapid growth in small- and medium-batch production of elaborate, high value added biscuits which would be hard for the other three countries' industries-and particularly Britain--to emulate.

6 Summary and conclusions

This study has sought to shed light on the determinants of international differences in productivity levels by comparing productivity, machinery and workforce skills in matched samples of food processing plants in Britain, Germany, France and the Netherlands. A relatively simple food product--biscuits--was deliberately chosen for investigation in order to facilitate productivity comparisons which would take explicit account of inter-country differences in the mix of product-qualities as well as differences in (physical) quantities produced. Nonetheless, the results reported here on quality-differences should be regarded as no more than a first attempt to treat in a precise way the issues associated with these complex aspects of economic performance.

6.1 Main findings of the study

Detailed information on biscuit outputs and associated labour inputs was collected from a total of 29 plants in the four countries. The main conclusions can be broadly summarised as follows:

1. When differences in 'quality' were neglected, and productivity was measured simply in terms of average tons produced per employee-hour, the British sample of plants as a whole compared favourably with Continental plants. In more detail: biscuit output per unit of labour input in Britain was much the same as in France and was some 25 per cent higher than in Germany; the overall leader by this measure of productivity was the Dutch sample with a 10-15 per cent advantage in tons per employee-hour over Britain and France.

2. The range of product qualities in each country was considerable, varying from simple undecorated biscuits, stacked in cylinder-form and wrapped in a single layer of plastic film, to elaborate multi-textured chocolate-coated varieties, stacked in successive small piles, multi-wrapped and boxed. The more basic types of biscuits accounted for a larger proportion of total output in the British sample than in the other three countries, and the more complex varieties constituted a larger share of output in Germany than elsewhere. Estimates of average quality levels based on ex-factory price differentials (net of raw materials costs) suggest the average value added per ton of biscuits in Germany was some 75 per cent higher than in Britain; in France and the Netherlands average quality levels were approximately 10-15 per cent above Britain.

3. International differences in average quality per ton are thus at least as important as differences in crude productivity measured by tons produced per employee-hour. Taking quantity- and quality-differences together to obtain a 'quality-adjusted' measure of labour productivity, it was found that real output per employee-hour was highest in the German sample, about 40 per cent higher than in Britain; in the Netherlands and France real productivity levels were some 20-25 per cent higher than in Britain.

4. In terms of age and sophistication, there was very little difference in the physical capital employed in Germany, France and Britain; the Dutch industry tended to rely more on older process equipment but had undertaken almost as much investment in new wrapping machinery as the other three industries. By contrast, in respect of human capital endowments, there were important differences between the four countries which were linked to both relative productivity performance and the predominant choice of product strategy in each case.

5. German plants specialising in small- and medium-batch production of elaborate, high value-added grades of biscuit greatly depended on the skills and knowledge of apprentice-trained bakers in process departments coupled with the technical and management expertise of Meister-trained supervisors. In addition productivity in most German plants was visibly enhanced by a 'pro-active' engineering policy involving not just preventative maintenance but also systematic collaboration between production supervisors and engineering departments on incremental process improvements and innovations.

6. In the Dutch and French industries, the average value added per unit of output was higher than in Britain but in both cases was well below German levels. The Dutch and French samples also tended to occupy an intermediate position in respect of shopfloor qualification levels: both had larger proportions of vocationally-qualified process workers and supervisors than Britain but few of these employees had been specifically trained in baking as in Germany and, in common with Britain, the Dutch and French plants had no equivalent of the German Meister available to them as production supervisors. In both countries these deficiencies had been partially offset by relatively long periods of initial on-the-job training equipping process workers to carry out tasks in all the main areas of operation.

7. Through the implementation of preventative maintenance, the Dutch and French samples also succeeded in matching average German levels of emergency downtime (under 5 per cent compared with an average 10 per cent in Britain). In the Netherlands several plants benefited from the diagnostic skills of technician-qualified engineering staff; in the French industry considerable resources had been devoted to continuing training and updating for machine-setting and maintenance personnel.

8. In Britain the relative lack of skilled process and supervisory staff hampered the performance of small- and medium-sized plants producing a wide range of biscuits. More success was achieved by larger highly-automated plants engaged in the bulk production of relatively uncomplicated varieties of biscuit but some of the benefits of scale-economies were lost due to excessive rates of emergency down-time and product wastage and the limitations of narrowly-trained employees. Manning levels in British process departments were comparatively high because the majority of workers had not been trained to switch flexibly between different tasks and were not expected to cover more than one oven-line at a time, and additional personnel were also frequently required in wrapping and packing areas to sort out the problems caused by machine breakdown and malfunctioning.

9. Senior managers in British biscuit plants were well-qualified academically compared to their Continental counterparts but, as in other industries investigated by the National Institute, they were frequently pre-occupied by day-to-day production problems which in other countries were typically dealt with by supervisors and technical support staff. Because this study was concerned primarily with the production capabilities of the main part of the workforce, very little attention was paid to the size and qualifications of the product development and sales and marketing departments in each country; these aspects deserve attention in future research.

6.2 Policy implications

10. Given the present structure of workforce skills in Britain--heavily polarised between a small minority of highly-qualified personnel and a large majority of low-skilled workers--it is understandable that many branches of British manufacturing have tended to specialise in highly automated mass production of relatively low value added goods, as is characteristic of biscuit manufacturing. One concern about this pattern of specialisation must be the limitations it places on future growth in real incomes in Britain. However, more immediate problems derive from the limited ability of many British producers of standardised goods to compete effectively against foreign imports from lower-wage newly-industrialising countries.

11. In the particular case of biscuits, there is little cause for concern about the future competitiveness of the industry given the highly specific nature of domestic tastes and preferences--sustained in part by large-scale advertising--which tend to limit the threat of import penetration. In addition, the price competitiveness of the British industry relative to Continental rivals is enhanced by its comparatively low labour costs which help offset the labour productivity shortcomings identified in this study. Several of the larger British-based firms have recently made substantial investments in new equipment and new product development and, in some cases, have also made strategic purchases of Continental production facilities.

12. However, these relatively favourable prospects for British biscuit manufacturers contrast sharply with those confronting other industries investigated by National Institute researchers. In the furniture and clothing industries, for instance, British firms which have pursued a relatively low value added product strategy are poorly-placed to withstand import competition from low-cost standardised goods or to respond quickly to rapid changes in consumer demand in favour of higher quality products (Steedman and Wagner, 1987, 1989).

13. The results of this four-country study--allied with those of earlier comparisons--therefore have important implications for British policy-makers. Future growth in output and employment in many industries is likely to increasingly depend on the ability to compete effectively in small-batch, high value added manufacturing. As our detailed examination of biscuit manufacturing suggests, the successful implementation of such a product strategy calls not just for the employment of highly-qualified staff in management and technical support areas but also for shopfloor and engineering personnel to possess the necessary skills to maintain production flow and quality standards in the context of rapid changeovers of machinery and raw materials. However, only very small proportions of each age cohort in Britain are trained to recognised craft- or technician-level standards. International comparisons serve to highlight the institutional deficiencies which contribute to Britain's limited provision of intermediate skills, for example:

* the traditional system of apprenticeship training lacks the legal foundations of its German counterpart;

* in contrast to countries such as the Netherlands and France, there is as yet no provision for a significant proportion of British secondary school pupils to combine general education with substantive vocational preparation, thus laying a foundation for subsequent employment-based training;

* there are limited incentives for employers to devote more resources to adult training or for individuals to invest their own time and money in skills acquisition.

In arguing the need for expanded development of intermediate skills in Britain, it is not intended to recommend direct policy or institutional @ @