Vocational education and productivity in the Netherlands and Britain.

Mason, Geoff ; Prais, S.J. ; van Ark, Bart 等

The contribution of differences in the Dutch and British education and training systems to the significant Dutch advantage in manufacturing productivity levels is examined in this article. The Dutch schooling system is characterised by high standards in mathematics, the provision of vocational education at ages 14-16 for a third of all pupils, and widespread vocational education at 16 +. The proportion of the Dutch workforce attaining vocational qualification approaches that of Germany and is well ahead of Britain. Comparisons of productivity. machinery and skills in matched samples of British and Dutch manufacturing plants were carried out in selected branches of two industries--engineering and food-processing. Higher average levels of workforce skills and knowledge in the Dutch samples were found to contribute to bigher productivity through better maintenance of machinery, greater consistency of product-quality and lower manning-levels (greater workforce flexibility, less learning-time on new jobs). The Dutch productivity advantage was greatest in product areas where small- or medium-sized batches are demanded by the market. '

They are eagerly debating on the reform of schools in the whole kingdom... COMENIUS on his first impressions of England, 1641(1)

We want to have a school of wisdom,...training them in everything necessary to life, in a manner so complete and so sure that nobody could ever be reproached that he knew something inexactly...or could not apply it in the true and right way. COMENIUS on teaching methods(2)

1. Introduction

The improvement of the vocational skills of the workforce continues to be an important policy issue in Britain, with growing interest in the detailed experience of successful neighbouring countries for lessons from which Britain might benefit. A contrast is now often drawn between the German approach to vocational qualification, based on virtually obligatory part-time attendance at vocational schools, combined with a traineeship at a place of employment, for all 15-18 year-olds not in full-time education; and the French approach--to some extent similar to the Japanese--which relies more on full-time vocational schools, beginning for some pupils at the age of 14-15 as an option during compulsory schooling. The consequences of these countries' training and education systems for the structure of workforce qualifications and for productivity have been examined in previous studies by the National Institute.(3) The present study extends these comparisons to the Netherlands. This country has: - -- a less centralised and less prescriptive approach to education than France or Germany;

-- a distinct and successful schooling system, including a substantial practical curriculum to train pupils 'in everything necessary to life' and to apply their knowledge 'in the true and right way', as Comenius put it in the above quotation;(4) and

-- attained a level of manufacturing productivity which, as measured from the Censuses of Production, is above Germany's and clearly above that of the UK.(5)

The first task of the present paper is to outline those features of the present Dutch education and training system which distinguish it from Britain and which seem likely to bear on industrial productivity; this is attempted in the second section. But we need to go further and examine, at least by way of illustrative example, how those differences operate in economic actuality; for that purpose we visited matched samples of manufacturing plants in two industries in both countries, namely engineering and a branch of food manufacturing.

The wide interest in engineering comparisons does not need elaboration, and follows previous comparisons between Britain and Germany for that industry carried out by the Institute in 1983-4?1 In choosing a second industry for comparison with the Netherlands, we wished to select an industry with a different mix of skill and organisational characteristics. The Institute's previous investigations were based on industries in which craft skills at the operative level were supremely important-metal working, wood working, clothing manufacture-and production was organised in batches in response to orders received. Modern food-manufacturing, chosen for the present study, is akin to what is often termed a 'process industry'--where teams of process-workers and maintenance staff supervise the continuous and precise running of automatic linked processes and machinery; production of main lines is often for stock, and customers' orders are delivered from stock. Aside from technical staff, there is considerable employment of those without (or with only low levels of) vocational qualification to deal with routine activities such as packing and loading. It was necessary for our purposes that the precise branch of the industry selected for our comparisons should be viable in both countries, that is, not dependent on local natural resources or patents; and, because of our limited research budget, that it should not be too complex. The manufacture of biscuits, as a particular example of food manufacturing, seemed to satisfy these requirements.

We visited a total of 36 plants in the two countries with a total of 12,500 employees: 21 engineering plants (12 in Britain, 9 in the Netherlands) and 15 biscuit-manufacturing plants (10 in Britain and 5 in the Netherlands). To ease comparisons of productivity, the engineering plants were chosen in each country from those manufacturing three particular production-ranges: centrifugal pumps, industrial hydraulic valves and small coil-springs (this is similar to the procedure in our previous engineering comparisons between Britain and Germany).(7) We also had discussions with machinery suppliers, vocational schools and colleges; and we compared vocational examinations in the two countries with the help of college teachers. Approximately two-thirds of the plants in each country originally approached by telephone for interview agreed to participate; the visits were carried out in 1989-91.(8)

The plants included in our samples were chosen to cover, as far as possible, the middle range of sizes in each country (as shown in official statistics, supplemented by informal trade estimates where necessary). The median size of plant in both industries was considerably larger in Britain than in the Netherlands, particularly so in biscuit production (see table 1; for details of the size distributions see Appendix A). In the interests of comparability a greater spread of biscuit plant-sizes was therefore chosen, so as to secure a degree of overlap between the countries. In engineering half or more of the plants in our samples in both countries had 100-200 employees(9); in biscuit production, half or more of the plants in our samples in both countries were in the rather wider range of 50-400 employees. The types of product could be matched fairly closely in engineering; but in biscuits there were some differences in the mix of qualities produced in the two countries (as discussed in Section 3 below).

On our visits to manufacturers we examined, as in our previous matched-plant studies: output per employee, type of machinery, and workforce qualifications; the results are presented in sections 3-5 below which deal, respectively, with topics related to those three headings. Section 6 summarises our findings.(10)

2. Vocational education

To understand how the Dutch workforce attains its skills it is necessary to describe two aspects of Dutch schooling which differ significantly from Britain: (a) its high degree of differentiation at the secondary stage, including Junior Vocational Schools; (b) and the Dutch system of post-16 vocational colleges, and the legal obligation imposed on 16-17 year old school-leavers to attend part-time vocational classes. In the present limited space a full treatment is obviously not possible. We shall pay more attention to Dutch schooling provision for average and below-average pupils, in respect of whom--in an age of automation--it has become ever more important to raise standards and, as the Institute's previous international comparisons have suggested, this is particularly so in Britain.(11) The high actual attainments of Dutch secondary school pupils are considered in Appendix B. This section concludes with a comparison of the vocational qualifications attained by the Dutch workforce in comparison with Britain and other countries.

Variety in schooling and the vocational stream

The Dutch schooling system is remarkably varied as a result of that country's long-standing tolerant policy of allowing any group of parents, employers or other interested citizens to establish 'privately-run' (bijzondere) schools in conformity with their own religious or philosophical ideals, and local economic needs; they are then provided with state funding, subject to minimum criteria on size of school (80 pupils may be adequate, depending on the municipality's population), on the employment of certified teachers, and on compliance with the very broad framework of their national curriculum (much less detailed than in Britain). Such 'privately-run' schools accounted for three-quarters of all pupils in 1986;(12) no fees are payable during the ages of compulsory schooling (i.e. up to 16), but modest 'voluntary' contributions are requested (in the region of [Br pound]70 a year). In these, as in all other state-maintained schools, parents are responsible for textbooks and costs of stationery.(13) In general principle these 'privately-run' Dutch schools are similar to 'voluntary-aided' schools in Britain which account however for only 15 per cent of all pupils in British state-maintained schools.

Dutch schools have for long placed emphasis on early systematic preparation for a career, in contrast to the avoidance of vocational preparation in Britain for pupils under 16.04) Until the age of 12 all Dutch pupils are in comprehensive primary ('basic') schools similar to British primary schools. They then move to one of a great variety of secondary schools for a further 4-6 years, the first four years of which are compulsory. These schools can be treated as failing into four main types, though in reality there are considerable differences within each type as exemplified further below. The broad distribution of 15 year-old pupils amongst the four types of secondary school in 1986 was as follows(15):----

--about 35 per cent were in full-time Junior Vocational Schools (LBO);

-- about 30 per cent were in Middle (Junior) General Schools (MAVO, fairly similar to the German Realscbulen), leading--via intermediate vocational schools--to administrative or higher technician positions (eg hotel administration, engineering maintenance foreman);

-- about 15 per cent were in selective schools (HAVO) intended to lead to higher education at polytechnics (eg for qualified engineers);

--about 15 per cent were in 'grammar' schools (VWO) leading to academic courses at universities (divided into two categories--the Gymnasium teaches Latin and Greek, but the Atbeneum generally does not do so).

This differs markedly from the present British school system, in which some 90 per cent of all secondary school pupils are in comprehensive schools; in many ways the Dutch system is similar to the selective system of Germany-though the Dutch has more school-types than the German. In respect of its specialised Junior Vocational Schools the Netherlands is closer to France, where vocational schools (Lycees pro[essionnels) are available from the age of 13 or 14. An increasing number of adjacent 'school-types' (eg MAVO and VWO) tend now to be located under the same roof, with common curricula for the initial years; thereafter each type follows its own curriculum and standards (moving a pupil from one school-type, or 'stream', to another is thought to be easier than if they were on separate sites). Comprehensive schools of the British kind, where the lull ability-range is under the same roof and often taught in the same class, account for barely 1 per cent of Dutch secondary school pupils.(16)

The clearest difference between the two countries from the point of view of vocational preparation is that in the Netherlands serious thought has to be given to a career at a younger age than in Britain.(17) Decisions as to the best type of secondary school for a particular child are of course never easy. It is provisionally settled in the Netherlands by parents and teachers at the age of about 12, with the help of school tests at the end of primary school. The pupil then usually enters one of a number of 'transitional' classes feeding adjacent streams in secondary school; at the end of that year a final decision is made. Even that decision can be reconsidered at later stages, but usually at a cost in increased length of schooling(18); for example, a pupil who is doing well in a Junior Vocational School may transfer to a General School, provided he 'repeats' at his new school the year that he has just completed to ensure he is up to the requisite standard for the following year.

For many years there has been much debate in the Netherlands about the possible extension of comprehensive schooling till the age of 15; there are now moves in that direction, but at present it seems likely that such schools will remain 'multilateral' (i.e. streamed) rather than mixed-ability comprehensives. Vocational courses may then be split into a preparatory 'theoretical year' taken at the Junior General School, and a 'practical year' taken at the Junior Vocational School. Our concerns in this paper must however concentrate on the system as it has affected the present-day structure of vocational qualifications, that is, as it has operated hitherto. Junior Vocational Schools in the Netherlands cover a considerable range of specialisations, usually classed under five broad types: some specialise in technical subjects (LTS) and are attended mostly by boys; others specialise in commercial subjects (LEAO) and are attended to a greater extent by girls; others are concerned with agricultural work, domestic science (LHNO) or retailing. Each of these broad types comprises yet further specialisations, varying according to local needs. Within each school building a limited selection of specialisations is often provided; for example, one of the schools we visited included courses in electrical engineering, vehicle repairing, office work, restaurant management, hairdressing and skin-care.

The curriculum for the first two years at Junior Vocational Schools, at ages 12 and 13, covers the familiar school subjects; vocational subjects are taken only in the last two years of schooling, at ages 14 and 15. General ('academic') school subjects at this latter stage are studied for just over half the curriculum time. The syllabuses and teaching methods for these general subjects are adapted to the ability-range of this category of pupils, and to some extent are co-ordinated with vocational subjects. For example, in mathematics there is a greater emphasis on arithmetic in these schools and less on 'modern' topics such as set theory, prime numbers, probability and other abstruse aspects which have found a place in the British curriculum; the mathematical attainments of Dutch secondary school pupils appear exceptionally high in international tests (they are remarkably high in their junior technical schools; see Appendix B). Vocational study takes up the remainder of the time and includes the equivalent of two or three afternoons a week in workshops, leading to standards between parts I and II of our City and Guilds examination (further details are in Appendix B).

All this is similar to the arrangements in French vocational schools, which always provide a mix of general and vocational education. The approach in secondary vocational schools in both the Netherlands and France differs markedly from that in most schools in Britain, where the corresponding section of pupils crosses a great divide at age 16: before that age in Britain there is virtually no vocational instruction, and after that age there is virtually no general education for pupils taking vocational courses.(19)

Costs to society of providing vocational schools are inevitably higher than general schools because of more expensive equipment required in workshops, and lower class-sizes; Dutch statistics indicate that the excess is about 25 per cent per pupil-year. Parents' contributions to the costs of schooling are however the same, irrespective of whether a vocational or a general school is attended. In total, Dutch government expenditure on education as a proportion of GNP is slightly higher than in the UK, at 6-4 compared to 5.2 per cent.(20)

Vocational schooling and training after 16

All Dutch 16 and 17 year-olds who are not in full-time education are subject to a legal requirement to attend pan-time classes, usually for two days a week; these part-time classes are divided into vocational subjects and general educational subjects. Those on formal apprenticeships (discussed further below) are usually required to attend classes on only one day a week, devoted to vocational theory and general subjects. Employers are prohibited from employing anyone unless release is granted for obligatory college attendance.

These requirements appear similar to the arrangements current in Germany for under-18s; but the consequences are different. Whereas in Germany about two-thirds of all young persons are apprenticed and attend part-time vocational colleges, in the Netherlands some 75 per cent of all 17-18 year-olds continue in full-time education half in vocational schools or colleges, and half in general schools. It seems possible that the indirect impact of Dutch legal provisions requiring part-time schooling of 16-17 year-olds is greater than may be inferred from the small proportion affected. These legal constraints embody widely accepted social conceptions of what is desirable in preparing young persons for adult life, and over the years have cumulatively encouraged a greater proportion of young people to continue with further full-time education.(21)

In total only about a tenth of each cohort join formal apprenticeship schemes. Most do so after having completed their schooling at a Junior Vocational School; but there is also a fair proportion of apprentices from other types of secondary school, and some take up an apprenticeship only after some years of employment(22) Dutch apprenticeship schemes are organised on lines similar to those in Germany, with 350 specified occupations (380 in Germany) and require attendance at day-release courses for two or three years (two years for the basic level, and a further year for the advanced level). The associated written and practical examinations are however not as well regulated for those of low attainments as in Germany or France (discussed further in Appendix C). Apprentices who pass only the practical part are awarded a 'certificate of practical skills'. Training allowances are usually based on four days a week at legally minimum rates, say, [Br pound]50 a week for 16-17 year-olds (dfl. 700 a month for 16 year-olds, 800 for 17 year-olds) with the government contributing a subsidy of some [Br pound]12 a week (dfl. 2000 a year) plus all costs associated with apprenticeship colleges.

Dutch pupils who attend secondary schools of the middle and more academic types (as detailed above) move on, in large proportions, to become students at Intermediate-level vocational colleges. We need concern ourselves with only a few key figures, since the diverse pattern of Intermediate Vocational Colleges is similar to that just described for Junior Vocational Schools. Of those who left Middle General Schools (MAVO) in 1986 with a diploma, as many as 68 per cent went on to three-year full-time vocational courses (plus one year of employment as trainee) at Intermediate Vocational Colleges (MBO); they were joined at these colleges by about a third of the more successful pupils who had completed Junior Vocational Schools.(23) In total about a quarter of the 16-19 age-group now attend such full-time vocational colleges. The courses again cover both academic subjects (languages, mathematics, science, etc) and specialised technical vocational subjects; the latter are closer in standard to those provided on British technician courses rather than craft courses.

Students' attainments at most Dutch Intermediate Vocational Colleges (MBO) correspond to our BTEC (Ordinary) National Diploma; but standards in their Intermediate Technical Colleges (MTS) are higher, and our own comparisons suggest that those who continue to the fourth year reach a standard approaching our BTEC Higher Diploma. (24)

The next higher level of vocational education in the Netherlands is provided by their Higher Vocational Colleges (HBO), which are similar to British polytechnics in providing full-time courses mainly for those who have completed a full general secondary-school course at age 18-191251; entry is also possible after Intermediate Vocational College (accounting for about a quarter of HBO entry). About 10 per cent of all 20 year-olds now attend HBOs compared with 7 per cent who attend universities. The standards aimed for at HBO generally corresponds to BTEC Higher Diplomas; their Higher Technical Colleges (HTS) attain standards comparable to our university degrees.

Work force qualifications as a whole

Taken as a whole we see that the Dutch system of vocational preparation relies principally on full-time vocational colleges; the Dutch system is thus closer to the French schooling-based system than to the German apprenticeship system. The Dutch have however been more successful in promoting vocational qualifications throughout the whole spectrum of the workforce (both in manufacturing and elsewhere in the economy) than the French, and are closer to Germany in the final result. This is shown in table 2 which provides summary estimates of workforce qualifications in the four countries based on recent Labour Force Surveys. The workforce has here been divided into four broad qualification-groups: (a) those with university and equivalent diplomas; (b) those with higher vocational diplomas (corresponding to our BTEC Higher Diplomas); (c) those with 'craft' or lower technician qualification; (d) those without vocational qualifications, though they may have general educational qualifications (for example, CSE passes).

Taking a broad view, there is little difference amongst the four countries at the top level of qualification, where 7-11 per cent of the workforce has university diplomas: the important differences lie at the intermediate--'technician' and 'craft'--levels. The Netherlands seems particularly well supplied with higher technician-level personnel compared not only with the UK, but also when compared with France and West Germany. At the craft and lower-technician levels only 20 per cent of the workforce in Britain have acquired vocational qualifications that fall into this category, compared with 38 per cent in the Netherlands and and 56 per cent in Germany. France has recently much increased its vocational education and, in this respect, now lies between Britain and the Netherlands.(26) Those without vocational qualifications, correspondingly, account for 63 per cent of the British workforce--but only 35 per cent in the Netherlands. The larger category of middle-level qualifications in Netherlands can perhaps be regarded as part of a long-standing characteristic of Dutch society, much emphasised by social historians, namely, the breadth of its 'middle classes' (middenstand) which resulted from its geographical position and made it a prime trading nation.(27)

'Human capital endowments', as summarised in table 2, are obviously not the sole factor affecting industrial productivity; but it is of interest that our companion studies at the Institute based on Censuses of Production have shown that manufacturing output per employeehour in the Netherlands, France and Germany were some 20-40 per cent ahead of Britain in 1988.(28)

3. Industrial structure and productivity

In the remainder of this article we examine how the very substantial differences between the Netherlands and Britain in the vocational qualifications of their workforces manifest themselves in economic reality by comparing matched samples of manufacturing plants in the two countries; as explained in the Introduction, engineering and biscuit production were selected as our industries for comparison. This section outlines the recent background of these industries, and then assesses average productivity differences between the countries in the plants we visited.

Engineering

Britain's engineering industries have passed through a decade of extreme pressures, characterised by competition from high-quality producers in Europe (i.e. where quality rather than low price was the biting edge).and from mass-produced standard components commg m at low prices from developing countries. Under the threat posed to the survival of many firms, employment in British engineering fell by a third in 198 0-91; output rose slightly, and productivity rose by a remarkable average of 3.5 per cent a year cumulatively in this period (i.e. by nearly 50 per cent in the decade). The corresponding Dutch industry was in a more robust position, and was able to maintain its employment as between the beginning and end of that decade (there was a temporary fall in the middle); and output and productivity rose by an average of only just over 1 per cent a year.(29)

While the engineering industries of both countries are characterised by very many plants of moderate size (say, of 100-200 employees), the requirements of efficient production have led to narrower product-ranges and much specialisation; consequently, finding producers of closely-matched products in the two countries for the purposes of this study was not easy. Our previous comparisons of engineering firms in 1983--4 between Britain and Germany had also encountered high degrees of specialisation and similar problems in matching; the impression gained then was that Britain more often tended to manufacture a lower grade of product than Germany. The impression gained on our present visits in 1990-1 was that most surviving British firms were now producing at close to the same quality-level as in other West European countries, and that British and Dutch firms were often actively competing with each other in home and third-country markets. British firms were not however always as readily adaptable in meeting customers' specialised demands.

Notwithstanding Britain's recent more rapid growth in productivity, comparisons of the level of output per manhour (based on returns to the Censuses of Production for 1984 with subsequent extrapolations) indicate that the Netherlands in 1990 was ahead of Britain by about 30 per cent.(30) From the sample of plants that we visited producing pumps, valves and springs we were able to select 23 pairs of closely matched operations in the two countries for which we compared outputs and direct labour-inputs (the observations came from six pairs of plants with very similar products, each pair of plants providing between two and five matches). The difference in output per man-hour averaged some 27 per cent in favour of the Netherlands for assembly operations, 29 per cent for machining operations, and 60 per cent for set-up times. The unweighted average Dutch advantage is 43 per cent (to which is attached a sampling error of 9 percentage points); if these three groups of operations are weighted to reflect their approximate shares in total direct employment (say, 1: 2: 1), the average Dutch advantage reduces to 36 per cent, which is close to, and statistically consistent with, the Census-based estimate.

Biscuits

The biscuit industry has been less subject to international competition, but has been under pressure since at least the beginning of this century to consolidate to take advantage of economies of scale that arise in automated process industries. Probably the most important developments that have increased the potential size of the market for individual plant in recent decades have been: better packaging to preserve freshness and prevent breakage, quicker transport facilities, and the development of preservatives and cold storage to prevent spoilage of moister varieties.

In addition, two important non-technical factors have encouraged larger sizes of firm particularly in Britain, namely, advertising and the pressures of the capital market. Advertising based on national newspapers began before the First World War in Britain, and progressively encouraged the large-scale production of selected varieties under brand-names; these pressures intensified with the advent of commercial television. Access to industrial capital via stock exchange listing permits larger companies to acquire smaller firms on a favourable basis, and then re-organise the group's production to benefit from longer runs. Some large biscuit concerns in both countries now benefit from access to the capital market via stock exchange listings, but the process began earlier in Britain than in the Netherlands (where most biscuit firms were family-owned until very recently). This has provided the British industry with a greater opportunity to concentrate production into larger plants.(31)

Despite the cost-advantages of producing standard products in long runs, consumers in both countries continue to pay a premium for variety, as well as for boxes containing an assortment of varieties; the latter usually still require hand-packing, though some firms have developed machine-packing of limited assortments. With rising general prosperity, demand for variety has increased, and has revived the scope for more labour-intensive and more expensive varieties often produced by smaller firms.

The consequences of these varied pressures in the past decade 1980-90 have been that the industry's output (measured simply in tonnage) increased by 16 per cent in Britain, productivity rose by 84 per cent, and the industry's employment fell by 37 per cent. In the Netherlands the industry's output rose more rapidly, by 43 per cent; productivity rose less rapidly than in Britain, but still by a respectable 49 per cent; and employment declined very slightly, by some 4 per cent.

The more rapid rise in Britain's productivity has to be understood as part of a 'catching-up' process: in terms simply of tonnage, the Dutch sample as a whole produced 21 per cent per man-hour more than the British sample (without standardising for differences in the size-distributions of the populations). This corresponds fairly closely with an estimated 15 per cent advantage in productivity for the Dutch biscuit industry as a whole derived from the Censuses of Production of the two countries for 1988.(32) Our samples thus seem to be not unrepresentative of the industries in the two countries. However, as we observed on our visits, the products made by our British sample included a greater proportion of simpler undecorated varieties and a smaller proportion of more complex (filled or chocolate-coated) varieties--involving more production processes--than in the Dutch sample; this probably reflects long-standing difference in national tastes since imports, though growing, are still not very important (they accounted for only 5 per cent of British domestic sales and 20 per cent of Dutch domestic sales in 1988; exports accounted for a tenth of British production and about 40 per cent of Dutch production). A quality-adjusted measure of productivity-if it could be derived so as to reflect differences in product complexity and not merely differences in prices of ingredients--may be expected to show a greater Dutch productivity advantage for the biscuit industry as a whole.(33)

Productivity and plant-size

In our samples of engineering plants we could not detect any variation in productivity with size (that is to say, within the size-range of plants we visited). But this was not so in biscuits, where we visited a wider range in order to attain an overlap of sizes in the two countries. The very largest British biscuit plants were very highly automated, and their productivity--if measured simply in terms of tons per man-hour--was as high or higher than any of the plants we visited in the Netherlands. But for medium-sized and small plants (producing, say, under 300 or 400 tons a week) we found average labour productivity in Dutch plants verged on double that of the corresponding British plants.

The apparent advantage of large plants in tonnage produced per employee arises partly from genuine economies in producing longer runs of substantially the same varieties, and partly from a shift towards more standardised varieties. Taken size for size, plants in the two countries do not of course produce precisely the same breadth in their mix of varieties: but that assumption provides a better basis for understanding productivity-differences than taking the samples as a whole without adjustment for plant-size. Accordingly, in order to compare productivity at similar levels of output, we carried out a regression analysis for our whole sample of observations (reported in Appendix D) which related productivity in each plant (tonnage per employee-hour, including both direct and indirect employees) to the total weekly output tonnage of that plant and to a term representing average inter-country productivity-differentials.

The results suggested (a) that at the same total weekly tonnage per plant, Dutch plants produced about twice as much per employee-hour as British plants; and (b) that a doubling of plant-size was accompanied by a 24 per cent greater tonnage produced per employee-hour (similar to Pratten's earlier conclusions on scale-economies in bread baking--but part of the increase in our case, we believe, reflects the greater emphasis in larger plants on simpler and more standardised biscuits).(34) With our small sample of observations there is a substantial margin of uncertainty attached to these estimates; nevertheless, our calculations indicate (at the conventional 95 per cent level of confidence) that for the same total weekly output Dutch plants required at the most only 80 per cent of the working-hours that were required in Britain.

In short: different as the engineering industry is from the biscuit-producing industry, their histories in the past decade in Britain and the Netherlands share several important features. First, in both industries there has been a greater rise in productivity in Britain than the Netherlands but the Netherlands is still ahead in productivity in both industries--more so in engineering (30-40 per cent) than in biscuits (15-20 per cent). Secondly, 'catching-up' has been at the expense of very heavy reductions in employment in Britain in both industries (by 30 per cent in engineering, 40 per cent in biscuits in the decade), while in the Netherlands employment in 1990 was near to its 1980 level in both industries. Thirdly, contrasting the different British experience in small-batch production, typical of our engineering sample, with the successful mass-production of large British plants concentrating on standard varieties of biscuits, the question arises whether it is precisely in the ability to produce flexibly with the help of modern machinery that Britain is at a disadvantage---whether because of inadequacies in equipment or in the technical competence of the workforce.

4. Machinery and its maintenance

The machinery in our sample plants in the two countries may next be described, at least broadly--in terms of age, technical sophistication, country of manufacture, and state of maintenance--so that we can better understand the range of skills required to operate it.

Age

The average ages of the main items of machinery lay in the range 7-14 years, as shown in table 3. The main difference between the countries was in engineering, where Dutch machinery was considerably younger (average age of 7 years as against 12 years in Britain). Some of the older machines in Britain no longer functioned as precisely as they used to, with ostensible symptoms of reduced efficiency: trouble in 'holding tolerances', longer setting-up times to get trial products to conform to specification, and restricted running speeds to reduce strains (over-heating) and help maintain accuracy. Probably more important, the Dutch plants had a greater proportion of modern CNC machinery: 35 per cent of the main items of Dutch machinery were CNC compared with 20 per cent in Britain. Allowing for the greater rate of usage of this equipment, the relative advantage of Dutch engineering plants was greater than these figures suggest. Change-overs on these modern machines are more rapid, so contributing to more efficient production of small runs.

In biscuit plants there was more new machinery in Britain than in the Netherlands (average ages of 12 and 14 years respectively). The differences arose particularly at the earlier stages of production, namely, mixing and forming equipment and tunnel ovens. However, apart from the need to update control systems, this type of equipment tends to be long-lived; it was apparent on our visits that any penalties for slower Dutch re-equipment at these stages in biscuit production were not serious (not comparable to the penalties for lower re-equipment observed in British engineering). Automated wrapping equipment undergoes more wear and tear, and has also been subject to continued technical advances; in both countries the machinery at these final stages was equally new (60 per cent installed in the past ten years; see Appendix E).

National sourcing

CNC machinery was largely imported by our sample engineering plants in both countries, with a slight difference in provenance: British engineering plants relied mainly on Japanese and Continental European suppliers, while Dutch plants relied mainly on Japanese and Korean producers--the latter were cheaper and thought adequate by the Dutch firms we visited. The other technologically-interesting items of equipment were modern high-speed automated wrapping machines for the final stages of biscuit production. In our British sample these machines came from well-known precision engineering firms in Switzerland and Italy; the Dutch sample took just over half their wrapping machinery from those same suppliers, and the remainder from Dutch machinery-makers (machinery for the food processing industry is one of the sectors of specialisation of Dutch engineering).

Much of the other types of machinery in British plants was of British origin, both in our engineering and in our biscuits samples. The use of locally-made equipment used to be regarded as a great advantage in coping with teething and maintenance problems, since expert help from local machinery-makers can be obtained more rapidly; buying from local machinery-makers also eases the development of special-purpose adaptations. From comments received in the course of the present inquiry it seems that after-sales advice, maintenance and repair services for foreign machinery have improved greatly in recent years, and spare parts for foreign machinery can now be obtained almost as readily as for domestically-produced equipment; the substantial easing of importing formalities in recent years and the growth of specialist international air-forwarding agencies have clearly helped to promote international competition in machinery.

In the Dutch plants (both engineering and biscuits) machinery was almost entirely imported, with Germany as the main supplier. In Dutch biscuit plants there were also a substantial number of British mixing machines and tunnel ovens, and they were operating efficiently; this seems to be a branch of machinery-making where Britain has kept at the forefront of technology. Even so, British equipment accounted for only about a fifth of the total, usually in second place behind German suppliers.

Machine downtime

Breakdowns and malfunctioning of machinery observed on our visits provided an important clue to sources of differences in productivity. Machinery breakdowns were observed in both countries; but they posed a greater problem in British plants, both in engineering and in biscuit-production, even when the equipment was relatively new and locally-made. This is similar to findings for other industries previously compared by the Institute. In most British engineering plants breakdowns were regarded as of little consequence; when they occurred spare capacity was available, and allowed transfer of production to other machines. However, such transfers still involved costs and delays, particularly if re-setting of the replacement machine was required. The extent of emergency downtime (when equipment should have been working, but could not do so because of unexpected breakdowns) was recorded more consistently by biscuit producers: on average there was 10 per cent of machinedowntime in Britain compared with 4 per cent in the Netherlands. The difference between these averages probably underestimated the true differences in cost. Malfunctioning of one machine often led to stoppages of other machines downstream, and thus not only to immediate loss of output but also to wastage of work in progress (half-baked or burnt biscuits) and excess labour costs for temporary stacking of intermediate products.

It was generally accepted by British biscuit plants that no more than 5 per cent emergency downtime should be tolerated, but only one of the ten British plants in our sample attained that level. Breakdowns were observed particularly in complex wrapping machinery, conveyor belts, and some automatic control equipment. Some malfunctioning observed in British plants was clearly due to inadequate diagnosis and inadequate corrective action following previous faulty functioning.

Maintenance and repair

We noted that Dutch plants gave much higher priority to prescribed maintenance procedures than did British plants. Routine 'preventative' maintenance was carried out in two-thirds of the engineering plants in our Dutch sample compared with only a sixth of our British sample. In biscuit production, four-fifths of Dutch plants undertook preventative maintenance compared with only one-fifth of British plants.

Machine-tool suppliers who sold in both countries told us they were busier in Britain than on the Continent with 'urgent call-outs', and with the supply and fitting of replacement parts which would not have needed replacement if machines had been maintained properly. The reasons proffered by British managers for avoiding routine preventative maintenance were familiar: they were too busy dealing with emergencies to cope with routine maintenance, they were too busy on equipment change-overs, and there was too little spare time between shifts. Our own judgement, consistent with our previous international comparisons, is that these proffered reasons are closely related to underlying shortages of relevant skills.

5. The workforce and its qualifications

The stronger emphasis in Dutch schools on preparation for work and on subsequent acquisition of vocationally relevant qualifications had consequences that were clearly manifest in the workforce-qualifications of the plants we visited. Not surprisingly, engineering plants in both countries employed more vocationally-qualified personnel than biscuit plants; and in both countries those working at higher levels, as well as those concerned with machinery maintenance, were more technically-qualified than those engaged on simpler repetitive tasks. But comparing like with like (that is, comparing personnel engaged in similar activities in the same industry in the two countries) it was clear that the Dutch plants employed a substantially greater proportion of qualified personnel. The following examples illustrate the size of the gap between the countries, distinguishing those employed on (a) the main production processes, (b) technical support, (c) maintenance and (d) supervision.

Production processes

In the engineering plants we visited, the main production processes were carried out by turners, millers, fitters, toolmakers, press-operators and so on: they accounted for about half the total workforce in both countries. In Britain, 40 per cent of this group of employees were regarded as craftsmen having served their time as apprentices, some having also attained City and Guilds or Engineering Industry Training Board qualifications. In the Netherlands nearly 80 per cent had reached a comparable craft-standard, having started with a diploma from their Junior Technical Schools (LTS) followed by about a year's on-the-job training. In both countries craft-level training is considered an essential prerequisite for setter-operators controlling a range of machines, such as centre-lathes and milling machines, to high standards of precision. On 'conventional' (non-CNC) machines this requires not just efficient initial selection of tools, jigs and fixtures, but also subsequent manual adjustments for tool-wear and for variability in quality of materials processed (castings are especially variable); even when modern CNC machines are used, some intervention during the machining cycle may be necessary to maintain tolerances. In both countries skilled setter-operators were frequently responsible for on-machine programming in the first place--or at least for modification of programmes which had been prepared by office-based technicians.

Those who had not attained craft-level qualifications were usually confined in both countries to simpler machines or to supervising fewer machines, and worked on a limited range of operations (e.g. drilling). As noted above, such 'semi-skilled' operators accounted for a much larger share of machine shop employment in the British plants than in the Netherlands: the majority had received a similar amount of on-the-job training as Dutch machinists--approximately 12 months--but, lacking the 'head start' of prior vocational education and training as received in Junior Vocational Schools in the Netherlands, British operators remained less versatile and less selfreliant than their Dutch counterparts.

In biscuit production the main process-workers operate, supervise and control automated lines of mixing machines, forming machines and ovens. Technical understanding is required in order, for example, to monitor and make appropriate adjustments to the dough's moisture-content, the dimensions of the product, and oven-temperatures.(35) In our British sample none of the process-workers had any vocational qualifications; initial on-the-job training averaged two months and was mostly restricted to a single task (e.g. mixing). In our Dutch sample about 40 per cent of process-workers had acquired Junior Vocational School diplomas (some had additionally acquired intermediate-level vocational diplomas from an MBO school, or apprenticeship certificates) followed by about half a year's on-the-job training covering the full range of these processes.

About half of all employees in biscuit plants are involved with relatively simple post-baking processes such as collating, wrapping, check-weighing, and packing. They require less technical expertise; housewives often take on such work on a part-time basis. In both countries most of the older long-serving employees in these activities had no vocational qualifications. But there was a marked difference in those recruited in the past decade or so. Most Dutch plants now look for a certificate of basic vocational competence in the form of a leaving certificate from a Junior Vocational School. A lower grade certificate (A or B) was acceptable, but an applicant with a higher grade (C or D, especially from a Junior Technical School) had better prospects of subsequently undertaking specialised training and promotion to process-worker.

In general the present-day need for flexibility at work, and for trainability to cope with changing technology, has increasingly led Dutch employers to rely on a certificate from a Junior Vocational School as an important indicator of a new recruit's potential. Dutch engineering employers valued the prior exposure of such schoolleavers to machinery and to workshop practice: it was not simply the importance of the skills acquired in this way, but that pupils had been helped to become clearer on their aptitudes and ambitions, and as a result were better motivated. Dutch engineering employers were consequently more reluctant to engage school-leavers who had attained only general--rather than vocational --school-leaving qualifications (i.e. MAVO rather than LTS). In the biscuit industry a similar attitude applied with respect to those seeking employment even in the simpler post-baking processes. The approach of British firms was that formal qualifications were irrelevant for employment in such basic operations. Most vocational qualifications available in Britain are acquired at a later age than in the Netherlands, and are at a higher level than required for this work. General school-leaving qualifications--CSEs or low-level GCSEs--were not mentioned as relevant; as explained, there is now virtually nothing equivalent to a vocational stream for under 16s in British secondary schooling from which an employer could receive a certificate of examined attainments relevant for work at this level.

Technical and maintenance staff

Technical support staff in our engineering firms were responsible for R & D, production planning, quality control, testing, etc.; they accounted for about a third of all indirect staff (that is, for about a sixth of all employed) in our sample firms in both countries. In the British plants only about 45 per cent had acquired technician or graduate qualifications, while in the Netherlands nearly 80 per cent had attained comparable levels (MTS or HTS qualifications).

Most of those engaged on routine machinery maintenance in both engineering and biscuit plants had acquired vocational qualifications, but had done so at different average levels in the two countries: those in Britain were mainly qualified craftsmen, 'diluted' as to about a fifth with semi-skilled workers; while in the Netherlands just over half had a craft-level qualification, and nearly half had gone further and attained the equivalent of a technician diploma (MTS). For most major repairs, half the plants in our engineering samples in both countries relied on external service contractors. In biscuits, by contrast, there was a greater reliance on plants' own full-time maintenance personnel, who also undertook changeovers on complex wrapping machinery.

Supervisors

Differences between the countries in the qualification of supervisory staff paralleled those just described for maintenance and shop-floor staff. In engineering plants in both countries almost all supervisors had acquired vocational qualifications--in Britain they were mostly at craft-level, while in the Netherlands they were more often at technician level.(36) In biscuit production, only a sixth of British supervisors held any form of vocational qualification, compared with two-thirds in the Dutch plants. In both industries the incidence of continuing or updating training for supervisors in British plants compared favourably with their Dutch counterparts. However, as was also the case for shopfloor worker training, such efforts rarely seemed able to bridge the overall gaps in vocational education and training levels between the two countries.

6. Productivity and skills

Productivity is clearly affected by a range of factors related to the quality and utilisation of inputs. In our engineering comparisons, the Dutch productivity advantage of some 30-40 per cent must be partly attributed to the greater use of new machinery in Dutch plants. In biscuit production, the Dutch industry has undertaken much less new investment than in Britain; but Dutch efficiency levels were enhanced by the use of spare capacity to implement full preventative maintenance programmes, and to ensure that machinery changeovers took place outside normal working time--something which was regarded as less feasible in multi-shift working British plants.

What both industries had in common were the marked disparities in human capital endowments between the two countries. As outlined in the previous section, higher standards of training and of vocational qualifications of the Dutch workforce are widely diffused throughout the various levels of plant-operation and plant-control. The following more detailed illustrations of how differences in productive efficiency relate to workforce flexibility and equipment reliability are perhaps linked as much to differences in the skill-spectrum as a whole as to skill-differences of particular sections of the workforce.

Flexibility

The success of Dutch engineering plants in recovering from the downturn in the early 1980s very often relied on their ability to introduce improvements in product-design in rapid response to customer 'feedback'. They were assisted by better-qualified design and planning departments (concerned with the selection of both optimum production processes and optimum scheduling of production) and organising production more thoroughly on the basis of modularised sub-assemblies. Dutch machine operators, being more broadly skilled, could transfer more readily from one task to another with less loss of time and of wasted materials because of 'teething problems'. Semi-skilled operatives, who account for 60 per cent of direct workers in British plants, 'needed patience' when transferred to new operations and, as we observed, often required the assistance of skilled setters and extensive supervision before they attained normal production levels. The same often applied to assembly departments in British plants when more complex, customer-specific products were to be assembled by semi-skilled operatives.

Managers and highly qualified technical staff in British engineering plants often seemed to be 'over-extended' by day-to-day production problems, and had less time to devote to the selection of new technology and to the development of innovations in products or processes. In fairness it must be recognised that some British engineering plants were as well-staffed with technically-qualified personnel as the Dutch, and were equally successful; but such plants were undoubtedly exceptional. In short, workforce skills in our Dutch sample as a whole seemed better suited to exploit the advantages of modern machinery which can produce efficiently in small batches to customers' specialised requirements.

Efficient production of very long runs o[ standard products was seen at its best in the large biscuit production plants in Britain. Dutch, as other Continental producers, have found their customers prefer variety; and Dutch producers are flexible enough to produce in their smaller plants a mix of short runs more efficiently than similar-sized British plants. A preference for variety by British consumers seems likely to grow in the longer term with a growth in prosperity and in international competition; and large British producers may then face problems in producing shorter runs efficiently.(37)

The advantages of greater operator-flexibility in Dutch biscuit plants were manifest in lower manning-levels in a number of processes. For example, at the pre-baking stages a single team was often responsible for two parallel conveyor lines, compared with usually only one line in Britain (a team consists usually of a mixer-operator, a forming-machine operator and an oven-operator); there also tended to be fewer stand-by relief personnel to cover tea-breaks and other absences, because of a greater ability to exchange tasks. In British plants, because of the narrow skill range, it was typically assumed that each operative needed to 'stick to' his or her 'own job' in order to be able to deal with problems as they arose. Greater reliability of well-maintained machinery in Dutch plants helped to reduce the reserve manning-levels needed to cope with possible defects in operation.

Reliability

As noted in the previous section, breakdown rates were lower in Dutch plants both in engineering and in biscuit production. In part this can be attributed to better maintenance on a planned preventative basis. Faults in machine operation seemed also to be corrected more rapidly. A more fundamental point, often noted, was that where operatives as a whole had been technically trained to a higher standard they were capable of 'sensing' faults before they had become serious, making appropriate adjustments to machinery at an early or convenient time ('a stitch in time ...'), and were better in assisting maintenance staff in the diagnosis of more serious problems; this was especially important when engineering firms relied largely on external maintenance contractors.

The greater diagnostic skills of Dutch technician-qualified maintenance staff contributed to a reduction in downtime, especially on more complex machinery (for example, biscuit-wrapping). Too many interruptions to production in British engineering plants were simply caused by machine parts having 'burnt out' (inadequate early diagnosis of the source of over-heating). Even on simple repairs, diagnostic skills were important; for example, if a securing bolt fails on a machine it is necessary not only to replace it, but to discover why it failed so that the failure does not recur the next day (as it did in one of the British plants that we visited). Sadder still: lack of supervision of chimney-sweeping routines on a British biscuit tunnel-oven led to a chimney fire--and to the closing down of a whole production line.

The issue of reliability concerns product quality as well as the functioning of machinery. In the engineering sectors we examined, no products can leave the factorygates without conforming to specified standards of accuracy; in British plants we often heard that machines had to be slowed down, and work had to be repeated, to achieve or maintain required tolerances. Dutch plants, with their relatively high proportion of craft-trained machinists, seemed better able to avoid these problems and 'repeat-work' was unusual. In biscuit manufacturing, lack of consistency in product-quality showed itself directly in products failing to pass smoothly through wrapping machines, or being rejected in significant proportions by electronic check-weighers. In British biscuit plants more operatives waited along the conveyor belt to correct misfeeds, more were engaged in unwrapping faulty packages, and more were shovelling biscuits that had fallen onto the floor into waste sacks (for sale as pig-food).

7. Summary and conclusions

This study has attempted to cast additional light on the relation between vocational education and productivity, with the help of visits in the Netherlands and Britain to schools and vocational colleges, and to matched samples of 36 manufacturing plants in selected branches of engineering and food manufacturing. Our findings can be summarised as follows:--

1 Dutch education places great emphasis on vocational preparation during the period of compulsory schooling. It has a much greater variety of secondary schools than its neighbouring countries, with highly differentiated curricula for pupils of different career-ambitions and aptitudes.

About a third of all secondary pupils attend schools with a distinct vocational emphasis; vocational subjects are taught at these schools from the age of 14, combined with general educational subjects. This is considerably earlier than is possible in Britain, where 16 is the normal minimum age for vocational education. Mathematical attainments of Dutch secondary-school pupils are outstanding by international standards, and provide a better foundation for technical and commercial training.

2 To promote vocational qualification after compulsory full-time schooling, the Dutch require obligatory release from work for attendance at part-time colleges for all 16-17 year-old school-leavers for usually two days a week (reduced to one day a week for those on integrated apprenticeship courses). These measures however now affect only a minority of school-leavers; three-quarters of all 17-18 year-olds remain in full-time education, of whom about half are at intermediate-level vocational colleges.

3 In consequence the Dutch workforce has acquired nationally-recognised vocational qualifications to an extent nearly as high as Germany's, and above that of France. The precise point where the Netherlands lies between France and Germany is not as important as that all three countries are distinctly ahead of Britain. Differences lie not amongst those with university degrees--where Britain, the Netherlands, France and Germany all now have some 7-11 per cent of the workforce with such qualifications: but at technician and craft-levels, where Britain has 27 per cent of its workforce with such qualifications and the other three countries have 40-63 per cent.

4 In engineering and biscuit production--the specimen industries examined empirically in the present study--the Dutch productivity advantage has narrowed in the past decade to some 15-40 per cent, as estimated both from the Censuses and from our sample plants. In the past decade employment fell by roughly a third in both these industries in Britain (as it has in total manufacturing in Britain), while in the Netherlands the corresponding industries more or less maintained their total employment.

5 The Dutch engineering plants we visited were equipped with more new and sophisticated (CNC) equipment than their British counterparts making similar products. This helped the Dutch to meet customers' demands for specialised products more rapidly and more efficiently. While those British engineering plants that survived the intense competitive pressures of the past decade have raised their productivity impressively, we suspect that further progress in output and productivity will depend on a faster rate of re-equipment.

6 Some British biscuit plants were of a much greater size than the largest in the Netherlands; they produced very long runs of standard varieties, and achieved productivity levels (in terms of tons per employee-hour) higher than any in our sample in the Netherlands. But British medium-sized plants, which produced shorter runs of a greater variety of products to meet consumers' varied demands, required considerably more labour per ton of output than plants of similar total output in the Netherlands. British plants--especially the very large British plants--were at an advantage in comparison with the Dutch in the newness of their machinery. Yet in the past decade the total output of the Dutch biscuit industry grew more rapidly than that of the British industry; it remains to be seen whether increased consumer demand for variety can be met efficiently by Britain's large-scale plants.

7 Substantially higher proportions of vocationally qualified personnel were found at virtually all levels in Dutch plants in both industries. Better planning of product improvements and of innovations in manufacturing processes in Dutch engineering plants depended on a greater proportion of staff trained to technician and higher levels. Amongst Dutch maintenance staff and operatives, there were also significantly higher levels of vocational qualifications; we were in little doubt that these differences contributed to fewer breakdowns of machinery, greater workforce flexibility, lower manninglevels and higher efficiency.

8 Taken to its most general level the present investigation carries the suggestion that--with Britain's present structure of skills--its comparative advantage is greater (or its international disadvantage is less) in automated 'process' industries, exemplified by large-scale biscuit-production investigated here, than in industries such as precision engineering where operatives' craft-skills are more crucial (other examples of the latter are the production of wood furniture and fashion clothing, examined in previous National Institute comparisons with Germany). In process industries success depends on complex linked machinery maintained by relatively few skilled technicians, though there is also scope for a considerable employment of operatives with lower-level (or no) vocational qualifications to deal with repetitive simple activities such as loading and packing. Competitive pressures of their own accord may be expected gradually to shift resources in these directions. But it is questionable whether such tendencies will be adequate to provide the greater numbers who are vocationally under-qualified in this country with real incomes at the levels of the European mainland. The examples of the more widespread provision of vocational education and training in the Netherlands, France and Germany will thus continue to merit very careful study and emulation by Britain.

ACKNOWLEDGEMENTS

We are grateful to the many firms in the Netherlands and Britain who co-operated with this inquiry and made the study possible. In addition we would like to thank the following trade and educational experts who were generous with their advice and comments: - In the Netherlands: J Gravemaker (LSBL, Apeldoorn); J van Hoof (SISWO, University of Amsterdam); B Htvels (ITS Nijmegen);W Groot (University of Leyden); S Kelder, J V Fledderus, H J Spronkers (LBO De Nieuwe Vaart, Zaandam); G Kraan, A van der Wielen, S Broeksma (Scholengemeenschap De Hamrik, Groningen); H J Mulder (Middelbare Vakschool Wageningen); Professor W J Pelgrum (Technical University, Twente);A J Timmers (Damland College, Sector Techniek, Zaandam); P Verhoeven (SOM, Woerden); drs P van Vliet (Stichting Studiecentrum Zoetwaren, Zeist); P Wolters (Centrum innovatie beroepsonderwijs bedrijfsleven, 's-Hertogenbosch).

In Britain: A Benns, L Frew (Glasgow College of Food Technology); M Brennan, P Hiscock (City College, St Albans); M Foy (Blackpool and The Fylde College); 0 Liisenga, Dutch school, Cobham; D Manley (food industry consultant); C Mansell (City and Guilds); D Raffe (University of Edinburgh); U Spencer (NEDO); T Wilson-Hooper, N Burgess (Bromley College of Technology); M Webber, W Newman (Biscuit, Cake, Chocolate and Confectionery Alliance); I Wheal.

We are glad to acknowledge financial support for this inquiry provided by the Training Agency (Department of Employment), the Economic and Social Research Council and the Scottish Council Development and Industry; but they are not responsible in any way for the views expressed in this paper.

NOTES

(1) From a letter translated by R F Young, Comenius in England (Oxford, 1932), p.65.

(2) Quoted in an essay by B Suchodolski, Comenius and teaching methods, in C H Dobinson (ed.), Comenius and Contemporary Education: An International Symposium (commemorating the tercentenary of the death of Comenius, Unesco, Hamburg, 1970, p.43; the late Professor Dobinson was SJP's headmaster at King Edward's Grammar School, Five Ways, Birmingham, before Dobinson became Professor of Education at the University of Reading).

(3) A compendium of previous articles in this field, originally published in the National Institute Economic Review, has recently been issued under the fide, Productivit, Education and Training: Britain and other Countries Compared (NIESR, December 1989).

(4) The great seventeenth century European educational thinker Comenius (1592-1670) lived in the Netherlands for fourteen years and 'made a great name for himself by the publication of various books' on the principles and practice of schooling (see W Rood, Comenius and the Low Countries, Gendt, Amsterdam, 1970, p.13).. He visited England only briefly. His main concerns, from our point of view here, were (a) to extend schooling to the whole population (not just to the wealthier classes), and (b) to ensure that schooling was based on substantial practical elements. These views were controversial at that time, and were not unopposed; their great impact on the schooling systems of the European mainland are evident from the Unesco publication cited above.

(5) The Institute's latest statistical comparisons based on Censuses of Production put Germany's output per employee-hour at 19 per cent above that of the UK in 1990; the gap narrowed sharply in the 19 8 Os--from a peak of 51 per cent in 1979--following a heavy 'shake-out' of employment in UK manufacturing (Mary O'Mahony, National Institute Economic Review, February 1992). The Institute's census-based comparisons with the Netherlands suggested the latter country was even further ahead--by as much as 44 per cent above Britain in 198 8 (B van Ark, Comparative levels of labour productivity in Dutch and British manufacturing, National Institute Economic Review, February 1990, p. 74); about 8 percentage points of that gap is attributable to the Netherlands having more employees in capital-intensive industries than the UK.

(6) Productivity, machinery and skills in a sample of British and German manufacturing plants, by A Daly, D W M N Hitchens and K Wagner, National Institute Economic Review, February 1985.

(7) For the benefit of technically interested readers: the pumps were generally with outlet diameters of about 10-20 cm; the valves were of similar diameter; and the springs were generally made from 2 mm steel wire.

(8) The firms were approached on the basis of trade lists, including classified telephone directories; small craft-bakeries were excluded from our sample of biscuit manufacturers. Initial telephone interviews were held with a considerably larger sample to ensure that those of whom we requested a visit manufactured the main varieties in which we were interested (rather than highly specialised products; eg in biscuit manufacturing we excluded firms making only ice-cream cones). To save travelling costs, the plants selected for visits in Britain were geographically clustered.

(9) In comparing our sample with national statistics, we have taken the industries defined as metal products and mechanical engineering (British SIC 31 and 32, and Dutch industry classification SBI 34 and 35) throughout this paper, though for convenience we refer only to 'engineering'.

(10) By way of general background: the population of the Netherlands is approximately a quarter that of the UK, and its area about a seventh. There have been few published comparisons between Britain and the Netherlands dealing with education, and none that we know deal with comparisons of manufacturing plants. A small sample inquiry into productivity in Dutch and British hotels, footnoted in the Institute's previous study of German and British hotels (Prais, Jarvis and Wagner, National Institute Economic Review, November 1989, p. 70, n. 21 ), indicated that Dutch and German hotels were equally efficient in their use of labour, and significantly more efficient than British hotels. The interest of the Dutch system of vocational preparation for Britain was recognised thirty years ago in the Crowther Report 15-18 (Ministry of Education, HMSO, 195 9; Appendix III by H A Warren, pp. 498-502); and twenty years ago in a report prepared for the Council of Europe by HM Staff Inspector D Porter, Technical and Vocational Education: Six Area Studies (Strasbourg, 1970, pp. 75-89 on the Dutch system); see also Vocational Education in Five Countries (Further Education Staff College, Coombe Lodge Report, Vol. 19, no. 5, Bristol, 1988, chapter on the Netherlands by L Blokzijl, T Knobbout and R Russell, pp. 281-8). An official account in English of Dutch schooling is available in their Ministry of Education document no. 332E, The Dutch Education System (Zoetermeer, 1988); current problems and plans were discussed in a substantial report for OECD prepared by Dutch education officials, Richness o[the Uncompleted: Challenges Facing Dutch Education (Ministry of Education and Science, State Printing Office, The Hague, 1989).

(11) See TN Postlethwaite, The bottom half in lower secondary schooling, in Education and Economic Per[ormance (ed. GDN Worswick, Gower, 1985, p.931; and our two artides comparing Britain with Germany and Japan respectively in National Institute Economic Review, May 1985 and February 1987 (also in Compare: A Journal o[ Comparative Education, 1986, nos. 1 and 21. Scotland has slightly different provisions; where no issue of substance is at stake, we refer for convenience throughout to 'Britain' rather than to 'England and Wales' (similarly, we omit references to Scotvec when referring to BTEC).

(12) Of those schools known as Junior Vocational Schools (LBO - to be described below), nine-tenths are financed as privately-run schools.

(13) After the age of 16 parents at all schools, 'privately-run' and other state-maintained, are required to make a contribution to costs amounting to an average of dfl. 1030 a year in 1988-89 per pupil (about [Br pound]300; there is a relief scheme for parents with several children at school, and for those with low incomes).

(14) The situation has been changed only slightly by the Technical and Vocational Education Initiative which provides funds--out of the Department of Employment's budget--to encourage a more positive approach in schools towards industry, without however leading to anything resembling the specific technical and vocational courses available to secondary school pupils in the Netherlands or France. The traditional attitude in Britain was well expressed in a submission to the Royal Society by the Trades Union Congress at the beginning of the 1980s: 'The TUC has always opposed any attempt to make the school curriculum 'more relevant' by using schools as institutions of narrow vocational preparation. It is not the job of the schools...' (Science Education 11-18, Royal Society, 1982, p. 13). This however misunderstands the broad foundations of vocational competence that are provided by Continental secondary schools.

(15) Zakboek Onderwijsstatistieken (hereafter ZOW), CBS, 1988, pp. 32-3, 74; those shown there as being in 'combined grades' relate to younger ages, and have been re-distributed for the purposes of the present estimates (see ibid., p. 29).

(16) No less than 48 different mixtures of Dutch secondary school-types were listed in a study of some ten years ago, though the largest ten types accounted for 80 per cent of pupils (WJ Pelgrum and THJM Eggen, Tweede Wiskunde Project: Opzet en Uitvoering, TH Twente, 1983, p.251.

(17) See P vd Dool, The Netherlands: Selection for vocational education starts early, Eur. ]. Education, 1989, p. 127; a more recent survey of vocational schools in the Netherlands and their current problems was published in the Dutch educational magazine School (SDV, The Hague), March 1990.

(18) Not surprisingly, parents tend tobe over-ptimistic in relation to the type of school initially chosen, and most changes tend to be to a lower category without involving repetition (R J Bosker, The middle school in the Netherlands, Int. J. Educational Res., 1988, p. 498).

(19) The opportunities in Scottish schools for taking Scotvec odu es at ages 15 and 16 do not, in practice, amount to anything substantially corresponding to Dutch vocational courses (see the references to classes $3 and $4 by D Weir and L Kydd, The National Certificate and Highers: a case of market forces, Scot. Ed. Rev., May 1991, esp. pp. 15 and 17).

(20) See Netherlands Ministry of Education, Richness (op. cit.), pp.168,171; the latter figures are from OECD compilations and warrant further examination for comparability of scope.

(21) ZOW, 1988, p.75.Unemployment rates among unqualified young perssons in the Netherlands have been very high, and have no doubt also encouraged longer full-time education.

(22) Descriptions of a number of apprenticeship schemes have been given by BWM Hovels, Arbeidsorganisaties en bet Leerlingwezen (Ministerie van Sociale Zaken en Werkgelegenheid, The Hague, 1985).

(23) Usually those with C-grades from LBO; for statistics see ZOW, 1988, p. 32.

(24) The high standards attained at Dutch MBO schools were noted at generation ago by Warren (1959), p. 499. A recent comparison by a professor of engineering in France, who had originally qualified in the Netherlands, put the MTS leaving standard as equivalent to the French Bac (F2) plus one year of industrial training; this also supports the view that the BTEC Higher Diploma is the appropriate comparison (see L K J Vandamme, Electrical engineering in the Netherlands education system, Eur. ]. Engineering Education, 1990, p. 48).

(25) At HAVO or VWO.

(26) As discussed in Hilary Steedman's analysis of trends in workforce qualifications in France (National Institute Economic Review, August 1990).

(27) See the comment on the brede meddenstand in the seventeenth century by S. Schama, The Embarrassment o[ Riches: An interpretation of Dutch Culture in the Golden Age (Collins, London; Knopf, New York, 1987), p. 174.

(28) See the studies by van Ark (National Institute Economic Review, February and August 1990) and O'Mahony (National Institute Economic Review, February 1992).

(29) Based on Censuses of Production, gross value-added deflated by producer price indexes; latest years from indexes of output and employment linked to most recent published Census.

(30) Based on detailed calculations similar to those published in the study by van Ark, National Institute Economic Review, February 1990, p. 74.

(31) While the present study was in progress, newspapers reported the takeover of the large Dutch biscuit producer Verkade by United Biscuits, a large international concern with its main interests in Britain. For a fuller account of the various pressures affecting the structure of the biscuit industry, see S J Prais, Productivity and Industrial Structure (Cambridge, 1981), pp.

(32) Based on average unit-value-ratios applied to value-added per employee (see footnote 30 above), treating all types of biscuit together as a single commodity for this calculation.

(33) The Institute's international comparisons of this industry are currently being extended to Germany and France; it appearsso far that quality-differences are of even greater significance for the latter countries, and require the development of explicit measures of the contribution of higher quality to productivity in its fuller sense. For the Netherlands as compared with Britain, a rough indication of the differences in product-mix is provided by the following figures: those varieties that account for the most expensive 15 per cent of the tonnage produced in the Dutch sample accounted for only 5 per cent of the tonnage in the British sample. The top varieties sell at over three times the price of basic varieties in both countries.

(34) C F Pratten, Economies of Scale in Manufacturing Industry (Cambridge, 1971, pp. 77-80, 270) suggested that in breadbaking a doubling of plant-size (from half to full 'minimum effective scale') was accompanied by a saving of 15 per cent of total unit-costs including materials (excluding delivery costs), equivalent to 30 per cent of net production costs (net of materials and delivery costs).

(35) The following details illustrate the kind of adjustment required. Both flour and fats have an inherent degree of natural variability which affects the consistency of the dough and its flow characteristics. Standard recipes and automated (often electronic) control provide the starting point. Temperature charts, control-limits and flow meters need to be observed and understood by the oven-operator. Precise dimensions and weight of the product need to be checked. Visual colour checks on the degree of baking have to be carried out routinely. Subsequent adjustment is then sometimes necessary to the dough-mix to provide satisfactory flow through the automated forming machines; and to the speed of the baking conveyor or the oven temperature to compensate for variations in moisture content.

(36) In more detail: 85 per cent in Britain had craft-qualifications, and the remainder had no qualifications; in the Netherlands 40 per cent had technician (MTS) qualifications, and 60 per cent had LTS certificates.

(37) This is similar to the problem noted in the Institute's previous study of clothing production: German manufacturers were able to produce efficiently a greater variety of styles for the upper end of the market, while British producers specialised in long runs of standard styles which fetched lower prices (see H Steedman and K Wagner, NIER, May 1989.)

Appendix A.

CHARACTERISTICS OF PLANTS VISITED

The tables below amplify the information on the distribution of the samples of plants as described in Section 1 (especially table 1, qv for sources).

Appendix B.

STANDARDS OF SECONDARY SCHOOLING IN THE NETHERLANDS

This appendix provides a brief introduction to the standards of attainments by pupils in the Dutch schooling system.

As in most parts of the European Continent, and in contrast to Britain, minimum attainments are required in each class before moving up to the next class; otherwise the class is repeated, or the pupil moves to a less demanding type of school. On average, only one or two pupils are affected in each classi38); this seems to be sufficient both to encourage pupils to reach required standards, and to encourage teachers to concentrate on those standards. Final leaving examinations provide implicitly the only detailed centrally-specified attainment targets; these have influenced the standards required in each class as a result of lengthy experience.

We may take mathematics to illustrate actual attainments. In the most recent international mathematical tests set to representative samples of 14 year-old pupils in twenty countries, Dutch pupils were the leaders in Europe, second only to the Japanese, and considerably ahead of England. For example, the sum 2/5 + 3/8 was answered correctly by 86 per cent of Japanese pupils, 80 per cent of Dutch pupils, but only by 28 per cent of English pupils.(39) The average scores on 157 mathematical questions were as follows:

 % correct
Japan 52
Netherlands 46
England and Wales 34

These tests were carried out in 1981; since then the Dutch have continued to revise their curricula and methods of teaching with a view to improving attainments in primary schools, and particularly those of low-attainers.(40) As far as could be judged from our comparisons of test papers for public examinations in mathematics, standards in the Netherlands are above those now set in the British National Curriculum, with a greater emphasis in the Netherlands on arithmetic, and on distinct courses differentiated according to ability. Because a grounding in mathematics is so important for subsequent scientific, technical and commercial education, great importance is attached by the Dutch to this subject at secondary school. The high attainments of Dutch pupils in international tests in mathematics have been attributed by Professor Pelgrum, the Dutch educational expert, precisely to the differentiation of the Dutch schooling system which, in his view, permits instruction-levels to be better adapted to the varying capabilities of pupils.(41) '

Our visits to Dutch Junior Technical Schools suggested a level of practical work in vocational subjects at ages 14-16 similar to that in our Colleges of Further Education at ages 16-17. Engineering teachers in both Britain and the Netherlands compared corresponding test papers, and judged the standards attained in theoretical vocational studies by successful pupils at these Dutch schools at age 16 to be well above our City and Guilds part 1, and approaching part II (usually not reached before the age of 19 in Britain).(42) It usually requires a year or two of subsequent work-experience, combined with on-the-job training, before a Dutch youngster with a pass certificate from a Junior Vocational School can be considered comparable to an English apprenticed craftsman with a City and Guilds qualification (or the German craftsman with his Berufsabscbluss at age 18 or 19); but it is clear that very strong foundations have been laid during their second.ary schooling. The !reportant advantage of the Dutch system, deriving from its vocational option during compulsory schooling, is that this standard is the norm for those who are mainly in, say, the lower third of the ability-range, whereas in Britain it is mainly those between, say, the top third and top half of the ability-range who are expected to be able to achieve such vocational standards.

In this way the Dutch schooling system has considerably broadened the section of pupils attaining vocational qualifications. Some 37 per cent of all 16 year-olds passed basic vocational school-leaving examinations in 198 7.1431 Hardly any pupils in Britain now attain anything approaching that level of vocational knowledge at that age; British pupils from the corresponding slice of the ability-range usually leave school with GCSE certificates in only a few general (ie non-vocational) subjects at low grades.(441

NOTES

(38) Higher rates occur in terminal or transitional years: for example, at the end of the final year of Junior Vocational Schools (LBO 4), an average of six pupils per class repeat (see CBS, Statistiek van bet Lager Beroepsonderwijs, 1988/89, Scbolen en leerlingen; Statistiek van bet MA VO, HA VO en VWO 1983/84, In-, door- en uitstroom van leerlingen).

(39) Derived from D A Robitaille and R A Garden, The lEA Study of Mathematics 11 (Pergamon, Oxford, 1989), p.125, and Appendix G3. The tests were based on multiple-choice questions with five options, so that 20 per cent could obtain the correct answer purely by chance, knowing nothing. The percentages quoted here have been adjusted for guessing (using the conventional simple formula (np-1)/(n-1), where p is the percentage marked correct, and n is the number of options). It is sometimes said by British experts in apologetica that British pupils compensate for their backwardness in routine calculations of this sort by doing better in 'problem solving'; but it is difficult to find any clear evidence for this.

(40) Netherlands Scientific Council for Government Policy, Basic Education (The Hague, 1987), pp. 26-7.

(41) Professor WJ Pelgrum was responsible for the Dutch component of the lEA mathematics comparisons; his comments appeared in the Dutch newspaper, De Volkskrant, 27 January 1990. Dutch mathematical attainments are higher not only for their best pupils (where England has always done well) but also for average and below-average pupils. Lack of space prevents a detailed account here; but our comparisons suggest, broadly, that a 16 year-old at the lower quartile in the Netherlands attains roughly the standard in mathematics of the median pupil in England.

(42) Comparisons were based on City and Guilds mechanical engineering maintenance (course 205) and electrical installation (course 236); and on Dutch Junior Technical School leaving examination papers (mainly at level B, which is reached by most pupils; level C is reached by about a tenth of Dutch LTS pupils, and the standards were judged to be closer to those expected of British BTEC National students). Excellent illustrated textbooks are published specifically for these courses in Dutch Junior Vocational Schools (eg the series published by Educaboek, Culemborg) in accordance with national curricular guidelines, and confirm the general high standards.

(43) A fail of approximately a tenth in the absolute numbers passing final LBO examinations is recorded for the most recent five years (ZOW, 198 8, p. 3 5, and vd Dool, op. cit. ); much of this fail can be attributed to a similar proportionate fail in the size of total age-group. On a medium-term view there has been little change; for example, in 1974 there were some 87,000 passes compared with 92,000 in 1987. Nevertheless there remains a longer-term worry amongst employers, especially small and medium-sized employers, who value the vocational grounding provided by the LBO; the concern is that a greater proportion of pupils leave primary schools and move to MAVO, rather than to LBO (this parallels the growth of German Realscbulen). The more immediate threat to the future of LBOs is that posed by current legislative changes; if put into practice, these would shorten the length of LBO courses, probably reduce the value of the qualification to employers, and consequently probably reduce the demand by pupils for such schools.

(44) By way of reservation it perhaps needs to be said that about a tenth of all Dutch school-leavers (about a third of all leaving LBO) fail to attain a leaving certificate, or attain only a grade A, which is not highly regarded by employers; nevertheless the substantial contrast with Britain remains. For a discussion of the decline of practical technical subjects taught in Britain at GCSE, see Pre-vocational Schooling in Europe today, by S J Prais and E Beadle, NIESR Research Report (new series no. 1), October 1991.

Appendix C.

ASSESSMENT IN DUTCH QUALIFYING EXAMINATIONS AT DUTCH VOCATIONAL SECONDARY SCHOOLS

The Dutch system of final qualifying examinations for say, the top two-thirds of the ability-range relies considerably on centrally-set examinations. But at lower levels of vocational qualification (LBO grades A and B) the Dutch rely more on internal assessment by teachers than on externally-marked examinations.(45) The ensuing variability in standards is widely recognised as a limitation by both employers and teachers. In this respect the Netherlands is deficient as compared with the external vocational examination systems of France and Germany, where--as we found on our visits to employers and colleges-the nationwide acceptability of the standard of their vocational qualifications was rarely questioned (minor variations in coverage of specialist topics were recognised by employers, and did not affect standards).

Having this in mind in drawing up table 2, we removed from the total with intermediate vocational qualifications half of those shown in the official labour force survey with LBO qualifications (based roughly on the proportion attaining grade A, and half of those attaining grade B, as shown in the detailed LBO examination results for vocational subjects in a recent year(46)); this adjustment affected about 10 per cent of the workforce.

Without that deduction, the Dutch and German distribution of qualifications would have appeared as virtually identical. The extent of deduction may seem unduly severe; we were however satisfied by the comments received that this weakness in the Dutch system needed recognition for the purposes of our present comparisons.

NOTES

(45) Official inspectors visit school routinely and inspect the results of international tests to ensure uniformity; but we were told that the effectiveness of this was questionable. See AJM Luiiten, Internal versus external assessment in Dutch examinations at 16-1- and 184-, Educ. Psycb., 1988, p. 237.

(46) Statistiek van bet Lager Beroepsonderwijs: Eindexamens 1987 (CBS, 1988), tables 11, 16, 20, 24 and 29 (praktijk, vaktheorie and vaktekenen).

Appendix D.

REGRESSION ANALYSIS OF PRODUCTIVITY AND PLANT-SIZE IN BISCUIT PRODUCTION

As explained in section 3 larger plants in biscuit production in both countries appeared more productive-in terms of tonnages (ignoring quality of product)--than smaller plants. To obtain an average measure of the gap between the countries for plants of a given size, we carried out a regression analysis of our observations on the 15 plants. It yielded the following results: - ln y/n: 1-25 + 0.31([+ or -]0-12)ln y + 0.80([+ or -].28)D (R2 = 0.47),

where y is tonnage produced, n is the number of employee-hours and D is a dummy variable set at unity for Dutch plants (standard errors are in brackets).

Two expository technical points may be mentioned on the robustness of this estimated equation (following questions raised by readers of a previous version of this paper). First, if we estimate an alternative equation, relating In n to In y and D (so as to eliminate the apparent inclusion of In y on both sides of our equation), then related mathematically-identical estimates of the regression coefficients are obtained; this is because only a linear transformation of the determining variables is involved. Secondly, bias may arise because of errors of measurement in the determining variable. This bias falls with an increase in the variance of the determining variable. In cross-sectional analysis the variance of the determining variable (here the size of plant) is very large; this problem is thus much less of a worry than in the analysis of economic time-series. In gathering our present observations we always asked for the usual output of the production line, rather than the output for that particular day or week; this reduces the variance of the measurement errors of the variables in the relationship.

In an attempt to eliminate the uncertainty introduced by the variety of biscuits, we made separate observations wherever possible on output and manpower for production-lines that made only the main standard ranges of dry undecorated biscuits ('Marie' 'Morning Coffee', etc). The central estimate of the gap in productivity between the countries derived from the ensuing regression analysis was only slightly smaller than mentioned above but, because of the smaller size of the sample for which these observations could be made, the sampling error increased (instead of the coefficient on the dummy-variable of 0.80([+over-]0-28) reported in the equation above, the new estimate was 0.66 ((+ or -) 0.3 9)). The two results could not therefore be distinguished statistically.

Appendix E.

DETAILED AGE-DISTRIBUTIONS OF MACHINERY IN SAMPLES OF BISCUIT-MANUFACTURING PLANTS

The following table gives more detailed information on the ages of the three main types of machinery installed in our samples of biscuit plants; a summary, combining the three types of machinery, is given in table 3 of the main text. In drawing up the table below, the age-distributions for the individual plants were weighted by their total employment; because of the limited size of our samples, the results were then rounded to the nearest 5 per cent.