D O C U M E N T O D E T R A B A J OW O R K I N G P A P E R S S E R I E S

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

HUMAN RESOURCEMANAGEMENT AND

OPERATIONAL PERFORMANCEIN SPANISH MANUFACTURING

INDUSTRY

Alberto Bayo MorionesJavier Merino Díaz de Cerio

DT 43/00

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Universidad Públicade NavarraNafarroakoUnibersitate Publikoa

C a m p u s d e A r r o s a d í a , 3 1 0 0 6 P a m p l o n a , S p a i nT e l / P h o n e : ( + 3 4 ) 9 4 8 1 6 9 4 0 0

F a x : ( + 3 4 ) 9 4 8 1 6 9 4 0 4

2

E - m a i l : w o r k i n g . p a p e r s . d g e @ u n a v a r r a . e s

3

Human Resource Management and Operational Performance in Spanish

Manufacturing Industry

Alberto Bayo Moriones

Departamento de Gestión de Empresas

Universidad Pública de Navarra

Campus de Arrosadía

31006 Pamplona – Spain

Fax number: 34 948 169404

E-mail address: abayom@unavarra.es

Javier Merino Díaz de Cerio

Departamento de Gestión de Empresas

Universidad Pública de Navarra

Campus de Arrosadía

31006 Pamplona – Spain

Fax number: 34 948 169404

E-mail address: jmerino@unavarra.es

4

Abstract

In recent years companies have begun to implement a number of human resource

management practices that are referred to in the literature as high performance or

innovative. One of the aims of this study is to provide descriptive details of the use of

these practices in Spanish industrial factories. In addition, we will attempt to determine

how and to what extent the adoption of this type of practices affects the firm's

performance record. We focus specifically on the impact HRM has on operational

performance. We examine whether there is any difference in the relationship between

HRM and the different kinds of operational results (efficiency, quality and time). We try

to assess if the effect of high-performance work practices is positive and significant for

all the operational results measures included. For this we will use a database covering

an initial sample of 965 factories each with a workforce of over 50 employees. We

begin with a review of the literature published so far, before going on to present the

descriptive statistics for the variables to be used and, finally, testing the impact of HRM

on operational performance by estimation of several logit models. Our results reveal the

presence of a positive, statistically significant correlation between the adoption of these

practices and improvements in performance, quality and flexibility.

Keywords: Human resource management, empirical, manufacturing, European

Community

5

1. Introduction

The belief that the people working for a firm are one of its main assets and one of

the decisive factors in determining its progress, is one that leaves little room for

argument. There is no question regarding the fact that workers' qualities, attitudes and

behaviour in the workplace go a long way to determining a company's success or lack of

it.

While this type of resource is one over which companies do not have complete

control, there do exist certain instruments to enable them to exert their influence on the

quality and performance of the human capital on which they rely. The human resource

management practices that they adopt will have a vital influence in this area and thereby

on the performance achieved by the firm.

In recent years references have begun to appear in the literature regarding a series

of human resource management practices that are dubbed high-performance, high-

commitment or innovative, and are said to help firms to achieve significant

improvements in performance. The aim of these practices is to achieve a more valuable

workforce, by selecting and retaining the more highly skilled individuals, and to

increase motivation in the existing workforce by getting its members to adopt the firm's

objectives as their own. This is accompanied at the same time by structuring and

organising tasks in such a way as to obtain the maximum benefit from the increased

capacity and motivation thus achieved among the workers. These practices include, for

example, the provision of training for the workers, broader job definition and

meticulous personnel selection processes.

This study will attempt to test the hypothesis that the way a firm manages

personnel influences its operational performance. Thus it will be in those firms where

the above mentioned practices are implemented that more outstanding improvement in

results will take place. The environment in which we intend to test this hypothesis is the

Spanish industrial sector. Using an initial sample of close to a thousand Spanish

6

manufacturing plants we will try to discover whether findings in similar sectors in other

parts of the world are also applicable to Spain.

This study aims to offer, first, a stock of useful descriptive data concerning the

implementation of a wide range of human resource management practices in a large

sample of Spanish industrial firms covering all sectors. This, in itself, marks an

important step forward in research into personnel management in our environment,

since, until now, very little empirical data of this nature has been published relating to

Spain. The analysis of the relationship between the general use of human resource

management practices in the operational context and the various performance

measurements (efficiency, quality and time-based) is also a significant contribution, in

view of the general scarcity of empirical studies with this sort of focus, the few that

exist applying mainly to the US context.

In the second section we will carry out a review of existing empirical findings

regarding the connection between combinations of human resources practices and

business results while section three we will be devoted to briefly explain the importance

of operational results in factory management. In section four we define the research

objectives of the article. Section five will begin with a few remarks regarding data

collection methods before going on to define the variables to be used in our subsequent

empirical analysis and defining the estimating framework. In section six we will

estimate the effect of personnel management on a plant's performance and analyse the

results obtained. We will finish with a discussion of the main conclusions to be drawn

from our study.

2. Human resource management and its impact on results

Workers are a key element in the running of a firm and as such play an important

part in the firm's success in reaching its fixed objectives. Human resources, taken to be

"the pool of human capital under the firm’s control in a direct employment relationship"

[43,p. 304] can provide the firm with a source of competitive advantage with respect to

its rivals.

7

This is possible because of the series of requirements that they fulfil [42,43]. The

first of these is the value added to the company's production processes, the contribution

made by each individual having its effect on the results obtained by the organisation as

a whole. Also, since individuals are not all the same, their characteristics are in limited

supply in the market. In addition, these resources are difficult to imitate, since it is not

easy to identify the exact source of the competitive advantage and reproduce the basic

conditions necessary for it to occur. Finally, this type of resources is not easily replaced;

though short-term substitutes may be found, it is unlikely that they will result in a

sustainable competitive advantage anything like that provided by human resources.

For a firm to find the human capital to provide it with a sustained competitive

advantage is not just a matter of luck. It depends rather on the action the firm is

prepared to undertake towards that end. It is through personnel management practices

that firms try to obtain the human resources that will give them the advantage when it

comes to holding their own against other companies. It is the human resources

themselves, however, and not the mere practices, that make up the source of competitive

advantage. Personnel management practices do not qualify as sources of sustainable

competitive advantage, since they are perfectly replaceable and quite likely to be copied

by other firms. They are, nevertheless, necessary, not only for the firm's human capital

to develop but also to enable it to be employed in such a way as to ensure improvements

in the company's performance.

If human resource management practices help to create better human capital in the

firm and, thereby, a sustained competitive advantage, it is reasonable to assume some

sort of connection between the way in which personnel is managed and the results

obtained by the firm. There are numerous empirical studies that deal with the influence

of individual personnel management practices on different performance measurements.

Among these we would find quality circles [26], recruitment [16], worker training [5],

profit-sharing schemes [38] and information sharing [31].

8

Other investigations have attempted to examine the individual impact of not one

but several of these practices. These would include investigations carried out by

Kalleberg and Moody [24], Delaney and Huselid [10], Black and Lynch [3,4], Cappelli

and Neumark [8], Harel and Tzafrir [15] and Fey et al. [12]. Finally, there are some

studies that take a global perspective on the relationship between personnel

management and performance, taking the view that practices are not applied separately

but in conjunction with one another. Instead of looking at the impact of one specific

practice, they study the total joint effect of the way in which the different aspects of

personnel management are dealt with, particularly the adoption of high performance

working practices1. Studies such as those of Huselid [18] and MacDuffie [28] are among

these.

This type of studies, and here we would include our own, analyse a wide range of

measurements. Among them we find, for example, such measurements of human

resource management performance as the ability to attract and retain employees [24],

management-worker relations [41], turnover [7,18,41], absenteeism [17,41], workers'

commitment to the company [17], job satisfaction among workers [17] and indices that

capture several of these [27].

In addition to these, we also see measurements of financial results, which are an

indication of the firm's overall performance. Among others we might mention

productivity [18,20], profitability [19], customer satisfaction [24], Tobin's q [18,20] or

the firm's market value [7].

Finally, we find studies which, like ours, are concerned with analysing the effect

of human resource management on aspects of operational performance, such as the

hours of labour required to manufacture a particular product [2,28], the percentage of

programmed time that the production line is in operation [22,23], the defects rate [2,28]

or the percentage of production that meets the required quality standards [22,23]. There

are also works that take as their dependent variable indices that capture the firm's

1 There are several works that offer an in-depth critical review of this literature. Among them we mustmention those of Dyer and Reeves [11], Becker and Gerhart [6], Huselid and Becker [19], Ichniowski etal. [21], Guest [14], Whitfield and Poole [39] and Becker and Huselid [7].

9

overall performance in this area. These would include Liouville and Bayad [27], which

incorporates aspects such as the defects costs, or Youndt et al. [44], that includes the

degree of utilisation of equipment, minimisation of spoilage, product quality and in-time

delivery. Broadly speaking, all the studies that address this issue, irrespective of the

type of operational result on which they are focused, show that the impact on the plant's

productive system, brought about by introducing high performance human resource

management practices, also has a beneficial effect on the business as whole.

3. Operational performance

The subject of the measurement, evaluation and conceptualising of operational

performance in a company is a recurrent theme in the different sections of the academic

literature. One of the first general classifications that has been widely used, is that of

Venkatraman and Ramanujan [36]. This adopts a Strategic Management perspective and

focuses on the measurement to establish a division between financial and operational

performance, with the emphasis on the latter. Following a similar line, Kaplan and

Norton [25] believe that the traditional measurements of financial performance are no

longer valid for today’s business demands. Therefore, they consider that operational

measurements for management are needed in relation to customer satisfaction, internal

processes and the activities concerning improvement and innovation in the organisation,

which lead to future financial returns.

Manufacturing performance, which includes part of the operational performance

previously mentioned, is commonly used in the field of Operations Management. This

type of results takes into account the company’s performance in reaching its basic

objectives, that is, productivity, quality and service. There are several studies which aim

to establish a classification of this kind of results [9,13,32]. For example, Corbett and

Van Wassenhove’s model [9] consider the measurements of the performance in three

dimensions: cost or efficiency, quality and time.

Efficiency refers to the best possible use of all available resources in order to

obtain the maximum output of them. This leads to the attainment of low cost products

10

through the reduction of the waste and allows the factory to give good value to

customers.

Traditionally quality has been defined in terms of conformance to specification

and hence quality-based measures of performance have focused on issues such as the

number of defects produced and the cost of quality. With the advent of total quality

management (TQM) the emphasis has shifted away from conformance to specification

and toward customer satisfaction. In any case, the firms must obtain high levels of

quality performance in order to improve or, at least, maintain their competitiveness.

The first dimension of time-based performance is reliability. It represents

fulfilling delivering promises. On-time deliveries may have a significant impact on

customer satisfaction, which makes it an issue to be seriously taken into account in

operations management. The second dimension related to time is that referred to the

speed of production processes. Time elapsed between materials reception and delivery

of product to the customer is frequently used as a measure of it. The development of

“just in time” (JIT) and other systems of production planning and control (e.g. OPT)

have as one of their main goals to improve the fluidity of production processes in order

to be able to give a faster response to customers demands.

4. Research objectives

As we have explained above, both the theoretical and empirical literature suggest

that the way in which the employees of a plant are managed has a significant impact on

its performance. The human resource management practises adopted contribute to a

great extent to explain organisational success. Those companies that have implemented

a bundle of high-performance work practices, that is, a human resource management

system aimed at improving the capabilities and motivation of people, outperform those

rivals that have not do so.

We have also stated that operational performance measures are the most adequate

to assess how a factory has been managed. In this article we intend to analyse the

11

impact of high-performance work practices on operational results. We understand that,

according to the above mentioned literature, the most immediate and direct effects of

human resource management in a factory should be on manufacturing performance, as

financial results can be affected by other variables that cannot be controlled from an

operations management point of view.

This article aims to contribute to the literature in some respects. First, it attempts

to ascertain whether the impact found for human resource management on firm

performance in other countries also holds in Spain. We investigate if high-performance

work practices lead to better performance. Second, we focus specifically on the impact

HRM has on operational performance using a large sample of manufacturing

establishments. Third, we examine whether there is any difference in the relationship

between HRM and the different kinds of operational results. We try to assess if high-

performance work practices are useful for factory managers in their efforts to achieve

each and every of their manufacturing goals or if their effect is not uniform for all the

operational results measures.

5. Methodology

5.1. Data

The Spanish manufacturing industry constitutes the scope of our study. The

concept of manufacturing industry is clearly defined in the National Classification of

Economic Activity (NACE), which includes all the manufacturing industries (from code

15 to code 37) with the exception of oil refining and the treatment of nuclear fuel (code

23). The sector of production and distribution of electricity, gas and water (codes 40 and

41), as well as the mining industries (from 10 to 14) are also excluded.

Fixing the unit of analysis was an important matter to settle. Two possibilities

were initially open to us: to choose the company or plant as the organisation to be

considered under study. We opted for the latter. In the industrial sector, the plant

constitutes the business unit that is of strategic importance for the implementation of the

12

practices that make up our study. These practices are adopted in the plant, and therefore,

it is at this level where problems arise and where the results must be analysed.

Moreover, the answers to the different questions raised are expected to be more reliable

when taken from the plant; since the knowledge of these issues is greater even if only

because of greater proximity.

Another aspect of the field of application to determine was the size of the plants.

The industrial plants included in our sample employ fifty or more workers. This limit

has been used in other studies relating to this area (see [33]) and it serves to cover a

wide spectrum of the population employed in Spanish industry, what is more it

simplifies the fieldwork. With these criteria the reference universe was formed by 6013

units, a thousand units being the aim of the sample, these were stratified according to

sector and size.

In order to carry out the investigation a questionnaire was made up and after the

corresponding pre-test, it was modified in different ways to form the final questionnaire.

The questions concerning human resource management refer to blue-collar workers.

The fact that we refer to a specific group of workers creates problems, as far as the

generalisation of the results to other professions is concerned. However, limiting the

occupation under study means that it is easier to compare the different units we have

studied, as within the company there are possibly several internal labour markets with

substantial differences between them.

As we had foreseen at the beginning of the study, most of the questionnaires

(more than three-quarters of them) were filled in by either the plant manager of the

production manager. The questionnaire covered different issues, all linked to

production. It was meant to be filled in by a person with a broad understanding of the

organisational aspects of the plant as well as, though to a lesser degree, the technical

ones. Nevertheless, the questionnaire’s complexity did not mean it could not be

understood by any of the plant managers with knowledge of the areas under study.

13

After making 3246 telephone calls to make the necessary appointments, 965 valid

interviews were undertaken. This number represents a response rate of 29,72 percent

and constitutes the initial sample about which we have information. The sample was

size- and sector- stratified in order to create adequate samples within size and sector

categories.

5.2. Measures of variables

5.2.1. Measuring high commitment HRM practices

A look at the different empirical studies to be found in the literature dealing with

high-commitment HRM practices soon reveals a lack of unanimity surrounding the type

of practices that ought to be included under this general heading. It sometimes even

happens that practices considered by some authors to be synonymous with innovative

HRM systems, are taken by others to be more closely related to traditional personnel

management systems.

For the purpose of the present study, the choice of practices to be included in the

category of high-commitment management was made with various points in mind. We

first turned our attention to those practices most often included in investigations into

innovations in HRM. A second consideration were the restrictions imposed upon us by

the questionnaire we were using; we could obviously only include in our study those

practices about which we had data. Taking these details into account, we included the

practices most often mentioned in the literature and which figured in our database. We

then added a number of other variables which, though not included in other studies, we

feel to fit into the management philosophy that we wish to investigate.

The use of employees already working for the organisation to fill vacancies

further up in the hierarchy is a variable common to all the different studies of high

performance HRM practices. Respondents were asked to show on a scale of one to five

how many of their current supervisors and skilled technicians had previously had jobs

14

on the shop floor. The PROMOTION variable assumes a value of one when more than

half the supervisors are in this position and zero if this is not the case.

Another factor that encourages workers' identification with the firm is their

confidence in being able to keep their jobs. The SECURITY variable assumes a value of

one when less than a five percent of the workforce is on a fixed-term contract of

employment and zero otherwise.

The criteria applied in the selection and recruitment of new workers give some

indication of how the firm is being run, since they reveal what type of qualities and

what kind of behaviour is being sought after in candidates for jobs in the firm. The

SELECTION variable tells us whether one of the two main factors considered when

selecting and taking on new workers gives priority to personality characteristics or the

ability to work as part of a team and learn new skills, rather than focussing on higher

qualifications and a closer match between the technical requirements of the job and the

abilities and technical skills possessed by the candidate.

The effort made by the firm to train its workers also serves as an indication of its

concern for the future of its workforce. TRAINING is equal to one if the workers have

received any formal training within the past year. As for the content of that training,

GROUPTRAIN assumes a value of one if the workers have received any training in

teamwork and problem solving techniques.

There are several issues relating to wage systems that must be included.

WAGELEVEL is a binary variable that indicates whether the workers' wage level is

above average for the same type of workers' in the same sector and the same region.

PLANTINC shows whether or not the plant uses incentive schemes based in any way

on its operations or financial performance. KNOWPAY indicates whether the main

factor in determining workers' wages is their level of professional skill, as opposed to

other criteria, such as the nature of their job or their length of service in the firm.

15

Several of the aspects included refer to the nature of the job. On the one hand,

there is the existence of autonomous working teams, which is measured by the variable

TEAM. The use of task rotation among shop-floor workers is captured by the binary

variable ROTATION. Self-inspection by workers as a quality management technique is

represented by the SELFINS variable. Respondents were asked in the interview to

indicate on a scale of nought to ten the degree to which their workers were allowed to

plan and organise their own work. For all those observations in which this assessment

was between five and ten, both inclusive, AUTONOMY assumes a value of one.

Specific action designed to increase the participation of workers in the running of

the firm may take the form of schemes to collect individual suggestions or the creation

of groups of workers to meet periodically in order to identify problems related with

their work and propose possible solutions. These two instruments are reflected in the

binary variables SUGGESTION and GROUPS, respectively.

The use of methods to allow communication to take place between workers and

company management is an issue that needs to be taken into account when dealing with

the attempt to get workers to identify with the firm. While periodical meetings with

employees in order to inform them about company issues creates a downward flow of

communication, conducting surveys in which employees are asked about their degree of

job satisfaction sets up a flow of communication in the opposite direction. This type of

action is captured by the variables MEETINGS and SURVEYS. OPENDOOR,

meanwhile, indicates whether or not the firm has organised open-door days in an

attempt to involve employees and the surrounding community and to create links

between the workers' private environment and the firm that employs them.

INSERT TABLE 1

Table 1 contains the most relevant descriptive statistics of the seventeen human

resource management practices that we have considered as high performance. This table

shows that the incidence of the different practices in Spanish industry varies

considerably. As far as the spread is concerned, at one extreme we find plant incentives,

16

which are applied by only twelve percent of the plants in the sample. Another two

practices with a limited spread of around twenty percent are related to increasing worker

involvement. These are, namely, surveys to detect job satisfaction and the organisation

of open days. In addition, only a quarter of the plants interviewed attach special

importance to the workers' level of knowledge and skills when fixing wages.

At the other extreme, we have workers' self-inspection of the quality of the

product they are manufacturing. This occurs in ninety percent of the plants. There is

also very widespread provision by the firms of some kind of formal training for its

production workers (eighty percent of the survey). Though to a more moderate extent,

we should also mention the widespread application of personality-based criteria in

personnel selection processes, as well as frequent use of internal promotion in order to

fill posts as job vacancies occur within the organisation.

Associations between the different practices and how they interrelate is an issue of

great relevance. A close look at the correlation table will provide some very interesting

information in this respect. As was to be expected, the correlation coefficients are

overwhelmingly positive, and among the negative only one is significantly distinct from

zero. In all, of the 136 coefficients that make up the matrix, only seventeen are negative.

It can be seen how some of the practices considered possess an individual

diffusion profile that falls far wide of the trend followed by most of the rest. This casts

serious doubts on the theory that human resource management practices are

implemented in such as way as to remain consistent with one another. However, this

phenomenon may also be interpreted as an indication of error in the theoretical

definition of the practices that make up the subject of our investigation. In other words,

it may be that some of the practices included fail to complement the rest.

In relation to this issue, there is evidence to suggest that paying above the going

rate, providing job security, remuneration for knowledge and the use of internal

promotion are practices with very little connection with other aspects of human resource

management. The case of the PROMOTION variable is particularly remarkable, in that

17

it fails to register even a single positive correlation significantly distinct from zero with

any of the remaining sixteen practices analysed. This echoes the results obtained in

Becker and Huselid [7] and points to the need for further theoretical investigation into

the relationship of this feature of internal labour markets and the rest of the human

resource management system.

When there is a great number of practices to be analysed, as in the case in hand,

the investigator is faced with two possible ways of evaluating the degree to which HRM

can be considered to be based on high commitment. The first of these involves taking

the variables that indicate the presence of individual practices and creating an index that

measures the extent to which high-performance employment practices are being applied

(e.g. [30,34,40]). The second option is to classify firms according to their level of

application of the whole set of practices under consideration [1]. It is thus possible to

opt for a more or less continuous measurement or for using a nominal variable. For the

purposes of the present study, we will apply both options, since this will help in

ascertaining whether the results obtained are robust in relation to the procedure used to

evaluate the way in which human resources are managed.

The index we intend to use for the present study is the total number of high-

performance practices taking place in the establishment, which we will call

NUMPRAC.

In order to organise the plants into groups according to the degree to which they

have introduced high-performance practices into their management of human resources,

we have classified them by cluster analysis. We began by carrying out an agglomerative

analysis using a Ward hierarchical procedure and the square of the Euclidean distance.

From this we were able to deduce that the optimum number of groups to be formed was

two. Taking the results of this first analysis as our initial centroids, we then performed a

non-hierarchical cluster analysis using two groups.

INSERT TABLE 2

18

The final result of this process was two groups. Table 2 shows the number of

plants making up each group, the mean value in each group for the variables, the chi-

square statistic and the statistical significance of the correlation between each of the

variables and membership of one group or the other. Table 2 enables us to appreciate

that the two groups resulting from the cluster analysis are similar in size, since there are

only four observations of difference between them. Judging from the frequencies of the

different HRM practices examined, the first of the groups is the one that can be classed

as containing the firms where high performance HRM practices have been introduced.

The second group, meanwhile, is made up of plants that continue to practise traditional

methods of personnel management in which the emphasis is placed on the strict control

of everything done by the workers. HRMG is a dichotomous variable which has a value

of one when the plant belongs to the first group taken from the cluster analysis, that is,

those plants that include high performance practices in their human resource

management, and a value of zero otherwise.

It can be clearly seen how the first group contains a greater percentage of plants

where the seventeen practices considered have been introduced. However, it must also

be emphasised that differences between the two groups are not always significant. This

is the case for two of the variables; on the one hand, no major differences can be

observed in the use of internal promotion of manual workers to fill vacancies for jobs as

supervisors and skilled technicians. Neither are there any differences significantly

distinct from zero to be seen in the scores on our employment security measurement.

These findings coincide reasonably closely with the conclusions drawn from our

previous analysis of correlation in the implementation of these practices. On all the

remaining variables, however, there is absolutely no question regarding the fact that

averages for the plants in the second group are substantially higher.

5.2.2. Operational performance measures

It is important to explain the two characteristics of the measurements of

performance we have used in this study. First of all, they are relative measurements of

the improvement in the results of the plant in relation to the situation three years earlier.

19

The different manufacturing performance, which are measured in an absolute way,

depend a great deal on the technology and type of process found in the plant. Therefore,

it becomes difficult to establish comparisons when the data is obtained from a group of

heterogeneous plants, even when the sector variable is introduced as a control variable.

The other noteworthy characteristic is the subjectivity of the information used. Results

of a subjective kind are often used in research on organisation. Some studies have

shown that there is an important relationship between the objective financial

performance and the same measured in a subjective way. This may serve as a

justification for the use of this kind of performance [35,37].

The indicator for efficiency (cost performance) we use is EFIC, that refers to the

percentage of productive hours in relation to the total number of hours the workforce is

directly present. It reflects the waste and inefficiency of the productive system and

states the unproductive moments owing to organisational problems (lack of material,

breakdowns, problems with quality, etc.).

The two indicators of improvement in quality performance correspond to a

concept of product quality as conforming to specifications and they are defined as the

percentage of defective products. We have included aspects related to finished products

(QUAL1) as well as unfinished products (QUAL2).

We also use two indicators for time performance. TIME1 indicates the

improvement in the percentage of delivery dates complied with is a typical

measurement of punctuality, considered a basic aspect of customer attention. TIME 2

indicates the improvement in the reduction in the time taken from the moment the

material is received to the moment the product is delivered to the customer. That serves

as an indicator of the speed of the processes (lead time).

The five performance variables are dichotomous variables. They have the value of

0 for those plants whose manufacturing results have not improved in the last three years,

they takes the value 1 for those plants whose results have improved.

20

5.2.3. Control Variables

The first control variable is the size of the plant measured by the natural logarithm

of the number of employees that work in the factory (LNSIZE). It is a usual way of

measuring establishment size.

Technology has a great significance when trying to explain the operational results

achieved by the plant. Technical features are measured by the variable AUTOMAT,

which aims to capture the degree of automation in the plant. The questionnaire enquired

after four technical features that we felt to be directly related to the degree of

automation: namely, robots or programmable automatons, automatic materials storage

and retrieval systems (AS/RSs), computer integrated manufacturing (CIM) and

computer networks for the processing of the plant's production data. By applying factor

analysis to these four variables, a single factor is obtained with an eigenvalue greater

than one, which accounts for just over 47% of the variance. The factor loadings on these

variables are greater than 0.44. AUTOMAT is defined as the average of the four

variables mentioned.

Competitive pressure can force firms to make special efforts in trying to obtain

better operational performance if they wish to survive in the market or maintain and

improve their financial results. The level of competition being faced by the firm is

captured by the variable COMPET. This assesses on a scale of one to five the evolution

of competition levels over the last three years in the sector in which the plant operates.

A score of one on this scale indicates a large decrease, whereas a score of five

represents a large increase.

QUALASSUR is a binary variable that assesses if the plant has set up a quality

assurance system. This tells us if the factory is following a quality strategy by

establishing organisational routines whose aim is to avoid defective products arriving to

the customer.

21

Table 3 shows the average and standard deviation on the dependent variables,

control variables and HRMG and NUMPRAC. Table 3 also shows the correlation

between HRMG, NUMPRAC, performance variables and control variables.

INSERT TABLE 3

This table enables us to see how the plants included in the survey remain at

intermediate levels of automation, though slightly below the exact mid-point. However,

the plants that participated in the study claimed that the competition they had to face

had increased over the three years previous to the interview. It is also worth mentioning

the effort that is going into increasing quality assurance at the plants, since almost

seventy percent of the manufacturers claim to have set up systems to deal with this.

As was to be expected, the vast majority of the plants report improvements in the

different plant performance areas considered, that is, efficiency in the use of resources,

quality, and the speed at which they complete the different stages in the productive

cycle. Averages for the variables that measure these aspects fall between 0,64 for

EFFIC and 0,72 for TIME2. Analysis of the correlation matrix brings us to the

conclusion that no incompatibilities are present in the performance attaining processes

in the different areas of production management at the plant. Plants that have improved

their management in one of the three areas considered are more likely to have improved

in the remaining areas also.

Table 3 also offers a simple profile of the plants that have opted to introduce

innovative practices in their personnel management style. The larger the size of the

plant, the greater the likelihood of its introducing this type of practices. Likewise, in the

area of technology, there is clear evidence to show that these tend to be plants with

more highly automated processes and where it will be more usual to find quality

assurance schemes in progress.

22

5.3. Estimating framework

The question of whether human resource management leads to differences in

operational results will be tested, because of the dichotomous character of our

dependent variables, by estimating logit models [29]. For each dependent variable

(performance measurement) three different models are constructed. The first of these

includes only control explanatory variables, the second incorporates the HRMG variable

and the third substitutes HRMG with the NUMPRAC variable which is another

measurement of the prevalence of high performance human resource management

practices. In this way we aim to obtain a clearer picture of the impact of the practices

under consideration on plant performance. Although this will be checked in all cases by

activity sector (eleven dummy variables), the coefficients do not appear in the

corresponding tables.

6. Results

Tables 4, 5 and 6 show results for the logit models estimated to examine the

determinants of plant performance. Table 4 is an analysis of the efficiency of the

productive system, table 5 deals with quality and table 6 time-related performance.

6.1. Efficiency

Table 4 shows the results of the logit model estimated for improvement in

efficiency, measured in terms of changes in the proportion of productive hours over the

total number of hours of direct labour. Though the model proves significant, results are

not entirely satisfactory, owing to the low pseudo-R2 value.

INSERT TABLE 4

Out of all the control variables included, both the level of automation and the

installation of quality assurance systems register coefficients significantly distinct from

zero. These, therefore, are factors that increase the manufacturing plant's chances of

23

improving the efficiency of its productive processes. This likelihood also increases with

size of the plant. The drop in significance of these three variables in the second and

third models was to be expected in view of their strong correlation with HRMG and

NUMPRAC. In the second model it can be seen how the fact of belonging to the high

performance HRM practitioners group enhances a factory's efficiency results, though

not to a significant degree. Results obtained on model 3, in contrast, show that if the

level of application is measured in terms of the number of practices installed

(NUMPRAC) we see the emergence of a positive and significant impact on

improvements in efficiency. For the case in hand, therefore, we are unable to draw any

definite conclusions regarding the relationship analysed.

6.2. Quality

Table 5 shows the results of the logit models estimated for the two variables that

capture the plants' quality performance. The first three models refer to QUAL1, changes

in the percentage of product defects, while the remaining three refer to QUAL2,

changes in the percentage of processing defects. All six models are statistically

significant and give a substantially better overall impression than those obtained when

attempting to explain manufacturing efficiency.

INSERT TABLE 5

As in the previous table, both the level of automation and the installation of

quality systems have a strong influence on the firm's progress in the pursuit of quality.

The two factors have a similar type of effect on the firm's capacity to improve the

quality of its products. In the case of QUAL1, however, a significant impact is also

brought about by the evolution of the firm's competitive position in the market. It would

appear that plants feeling themselves exposed to increasing competition are forced to

improve the quality of the final products that they take to market.

For both product and processing defect rates, it is apparent that an understanding

of the impact of the human resource management policies being implemented helps to

24

explain the evolution of the plant's quality performance. In the case in hand, results

remain conclusive whatever method is used to measure the implementation of these

practices. The explanatory capacity of the two models increases substantially by

introducing the HRMG variable (for QUAL1, the pseudo-R2 increases by 51,8% and for

QUAL2 the increase is 39,5%) a similar effect also takes place with the NUMPRAC

variable. The implementation of high performance HRM practices in a factory has a

beneficial effect on reducing the defect rate.

6.3. Time-based performance

Table 6 shows the results of the logit models estimated to explain trends over the

last three years in the two indicators used to evaluate speed of action, which are the

percentage of on-time deliveries and the time that elapses between receiving the

materials and delivering to the client. It can be seen that all six models are significant

and help to explain the tendencies of the variables being analysed.

INSERT TABLE 6

Once again we find that quality assurance systems and high levels of automation

in product processing help firms to achieve improvements in time-based performances.

In the case of TIME2 it is also possible to see the effect of competition on the results

obtained by the firms in this respect. Plants that come under higher levels of competitive

pressure make a greater effort to reduce the time that elapses between receiving

materials at the start of the production cycle and the moment when the product finally

reaches the client, irrespective of whether or not they accompany this with action in the

technological area.

The conclusion that emerges yet again is that the way workers are managed has its

effect on a plant's performance. Correlation between HRMG and NUMPRAC and both

dependent variables is highly significant and, as with the quality results, the explanatory

capacity of the models increases noticeably with the introduction of these variables. The

implementation of a framework of innovative practices in the area of human resource

25

management leads not only to an increase in the percentage of on-time deliveries but

also to a reduction of the amount of time the firm takes to manufacture the product and

deliver it to the client.

7. Conclusions

This study enables us to confirm that the way in which human resources are

managed influences a company's performance. Generally speaking, our findings support

the theory that by using high performance HRM management systems, organisations

can improve their chances of reaching objectives as long as they do not lose sight of

other aspects, mainly of a technological nature, the explanatory capacity of which is

considerable. These findings coincide largely with those of researchers into this issue in

other contexts.

It can not be said, however, that the impact of the type personnel management

practices used has a significant impact in all of the three measurements of

manufacturing performance analysed. When it comes to efficiency, for example, we can

not be absolutely certain that the development of high performance human resource

management practices in the plant noticeably increases the likelihood of improvements,

since results differ depending on how the implementation of such practices is measured.

In the case of quality and time-based performance, however, there is conclusive

evidence that high performance work practices can bring about substantial

improvements in company performance.

One of the limitations of this study is a result of the context in which the empirical

analysis was carried out. The fact that we have focused on the manufacturing industry,

with the wide range of activities which that implies, and have concentrated on only one

of these, albeit the most frequent, means that our conclusions can not be made

applicable to all professions and sectors.

In subsequent studies we look further into issues that have been neglected in this

one. One of these is the possibility of a complementary relationship between the

26

practices comprised in personnel management. Although we have looked on the system

as a whole and not on individual practices, it is worth looking into the question of how

these practices are interrelated and in what way this affects company performance, since

this could reveal the importance of internal consistency in the field of personnel

management.

It will also be necessary to ascertain how far human resource management

practices complement other areas of plant management. While our findings support the

theory of the universal impact of these practices, it remains to be seen whether the

extent of this positive effect depends on other decisions taken in the factory. The degree

to which HRM complements strategy and technology will require special attention.

Acknowledgements

The authors would like to thank Fundación BBVA and the Spanish Ministry of

Education (PB 98-0550) for financial support.

References

[1] Arthur JB. The link between business strategy and industrial relations systems in

American steel minimills. Industrial and Labor Relations Review 1992;45:488-506.

[2] Arthur JB. Effects of human resource systems on manufacturing performance and

turnover. Academy of Management Journal 1994;37:670-687.

[3] Black SE, Lynch LM. How to compete: the impact of workplace practices and

information technology on productivity. NBER Working Paper Series 1997;6120.

[4] Black SE, Lynch LM. What’s driving the new economy: the benefits of workplace

innovation. NBER Working Paper Series 2000;7479.

[5] Bartel AP. Productivity gains from the implementation of employee training

programs. Industrial Relations 1994;33:411-425.

[6] Becker B, Gerhart B. The impact of human resource management on organizational

performance: progress and prospects. Academy of Management Journal

1996;39:779-801.

27

[7] Becker, BE, Huselid MA. High performance work systems and firm performance In:

Ferris G, editor. Research in Personnel and Human Resources Management. vol. 16.

Greenwich, CT: JAI Press, 1998. p. 53-101.

[8] Cappelli P, Neumark R. Do “high performance” work practices improve

establishment-level outcomes. NBER Working Paper Series 1999;7374.

[9] Corbett C, Van Wassenhove L. Trade-offs? Why trade-offs? Competence and

competitiveness in manufacturing strategy. California Management Review

1993;35:107-122.

[10] Delaney JT, Huselid MA. The impact of human resource management practices on

perceptions of organizational performance. Academy of Management Journal

1996;39:949-969.

[11] Dyer L, Reeves T. Human resource strategies and firm performance: what do we

know and where do we need to go? International Journal of Human Resource

Management 1995;6:656-670.

[12] Fey CF, Björkman I, Pavlovskaya A. The effect of human resource management

practices on firm performance in Russia. International Journal of Human Resource

Management 2000;11:1-18.

[13] Filippini R, Forza C, Vinelli A. Trade-off and compatibility between performance:

definitions and empirical evidence. International Journal of Production Research

1998;36:3379-3406.

[14] Guest DE. Human resource management and performance: a review and research

agenda. International Journal of Human Resource Management 1997;8:263-276.

[15] Harel GH, Tzafrir SS. The effect of human resource management practices on the

perceptions of organizational and market performance of the firm. Human Resource

Management 1999;38:185-200.

[16] Holzer HJ. Hiring procedures in the firm: their economic determinants and

outcomes. In Kleiner MM, Block RN, Roomkin M, Salsburg SW, editors. Human

Resources and the Performance of the Firm. Washington, DC: BNA Press, 1987.

[17] Hoque K. Human resource management and performance in the UK hotel industry.

British Journal of Industrial Relations 1999;37:419-443.

28

[18] Huselid MA. The impact of human resource management practices on turnover,

productivity, and corporate financial performance. Academy of Management

Journal 1995;38:635-672.

[19] Huselid MA, Becker BE. Methodological issues in cross-sectional and panel

estimates of the human resource-firm performance link. Industrial Relations

1996;35:400-422.

[20] Ichniowski C. Human resource management systems and the performance of U.S.

manufacturing businesses. NBER Working Paper Series 1990;3449.

[21] Ichniowski C, Kochan TA, Levine D, Olson C, Strauss G. What works at work:

overview and assessment. Industrial Relations 1996;35:299-333.

[22] Ichniowski C, Shaw K. The effects of human resource management systems on

economic performance: an international comparison of U.S. and Japanese plants.

Management Science 1999;45:704-721.

[23] Ichniowski C, Shaw K, Prennushi G. The effects of human resource management

practices on productivity. American Economic Review 1997;87:291-313.

[24] Kalleberg AL, Moody JW. Human resource management and organizational

performance. American Behavioral Scientist 1994;37:948-962.

[25] Kaplan RS, Norton DP. The balanced scorecard - measures that drive performance.

Harvard Business Review 1992;70:71-79.

[26] Katz HC, Kochan TA, Gobeille KR. Industrial relations performance, economic

performance, and QWL programs: an interplant analysis. Industrial and Labor

Relations Review 1983;37:3-17.

[27] Liouville J, Bayad M. Human resource management and performances: proposition

and test of a causal model. Human Systems Management 1998;17:183-192.

[28] MacDuffie JP. Human resource bundles and manufacturing performance:

organizational logic and flexible production systems in the world auto industry.

Industrial and Labor Relations Review 1995;48:197-221.

[29] Maddala GS. Limited-dependent and qualitative variables in econometrics.

Cambridge: Cambridge University Press, 1983.

[30] McNabb R, Whitfield K. The distribution of employee participation schemes at the

workplace. International Journal of Human Resource Management 1999;10:122-

136.

29

[31] Morishima M. Information sharing and firm performance in Japan. Industrial

Relations 1991;30:37-61.

[32] Neely A, Gregory M, Platts K. Performance measurement system design. A

literature review and research agenda. International Journal of Operations

Production Management 1995;15:80-116.

[33] Osterman P. How common is workplace transformation and who adopts it?

Industrial and Labor Relations Review 1994;47:173-188.

[34] Pil KF, MacDuffie JP. The adoption of high-involvement work practices. Industrial

Relations 1996;35:423-455.

[35] Powell TC. Total Quality Management as competitive advantage: a review and

empirical study. Strategic Management Journal 1995;16:15-37.

[36] Venkatraman N, Ramanujan V. Measurement of business performance in strategy

research: a comparison of approaches. Academy of Management Review

1986;11:801-814.

[37] Venkatraman N, Ramanujan V. Measurement of business economic performance:

an examination of method convergence. Journal of Management 1987;13:109-122.

[38] Weitzman M, Kruse DL. Profit sharing and productivity. In Blinder AS, editor.

Paying for productivity. a new look at the evidence. Washington DC: The

Brookings Institution, 1990.

[39] Whitfield K, Poole M. Organizing employment for high performance: theories,

evidence and policy. Organization Studies 1997;18:745-764.

[40] Wood S, Albanese MT. Can we speak of a high commitment management on the

shop floor? Journal of Management Studies 1995;32:215-247.

[41] Wood S, de Menezes L. High commitment management in the U.K.: evidence from

the Workplace Industrial Relations Survey, and Employers’ Manpower and Skills

Practices Survey. Human Relations 1998;51:485-515.

[42] Wright PM, McMahan GC. Theoretical perspectives for strategic human resource

management. Journal of Management 1992;18:295-320.

[43] Wright PM, McMahan, GC, McWilliams A. Human resources and sustained

competitive advantage: a resource-based perspective. International Journal of

Human Resource Management 1994;5:301-326.

30

[44] Youndt MA., Snell SA, Dean JWJr, Lepak DP. Human resource management,

manufacturing strategy, and firm performance. Academy of Management Journal

1996;39:836-866.

31

Table 1

Mean, standard deviation and correlationa matrix for HRM practices (N=758)

Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1. Promotion 0,66 0,48

2. Training 0,80 0,40 0,01

3. Grouptrain 0,36 0,48 0,04 0,38***

4. Wagelevel 0,42 0,49 0,04 0,05 0,12***

5. Plantinc 0,12 0,32 -0,01 0,08** 0,11*** 0,04

6. Knowpay 0,25 0,43 0,00 0,09*** 0,04 0,03 -0,05

7. Team 0,47 0,50 -0,05 0,11*** 0,11*** -0,01 0,02 -0,04

8. Rotation 0,46 0,50 0,03 0,12*** 0,13*** 0,05 0,10*** 0,01 0,02

9. Selfins 0,90 0,31 -0,01 0,06* 0,06* -0,05 0,07* 0,01 0,05 -0,01

10. Autonomy 0,27 0,44 -0,02 0,04 0,05 0,00 0,07** -0,01 0,19*** 0,02 0,03

11. Groups 0,39 0,49 0,00 0,28*** 0,30*** 0,09** 0,11*** 0,03 0,18*** 0,07* 0,03 0,15***

12. Suggestion 0,57 0,50 0,00 0,29*** 0,19*** 0,01 0,05 0,01 0,09*** 0,00 0,10*** 0,12*** 0,29***

13. Meeting 0,59 0,49 0,01 0,33*** 0,18*** 0,10*** 0,10*** 0,10*** 0,14*** 0,08** 0,09*** 0,09** 0,34*** 0,34***

14. Survey 0,21 0,41 0,00 0,17*** 0,16*** 0,09** 0,12*** 0,05 0,15*** 0,07** 0,10*** 0,13*** 0,32*** 0,24*** 0,28***

15. Opendoor 0,19 0,39 0,01 0,15*** 0,18*** 0,03 0,06* 0,01 0,14*** 0,03 0,07** 0,06* 0,23*** 0,17*** 0,22*** 0,28***

16. Selection 0,67 0,47 0,02 0,07** 0,13*** 0,01 0,01 -0,04 0,03 -0,01 -0,01 0,02 0,03 0,07** 0,09*** 0,10*** 0,04

17. Security 0,28 0,45 -0,06* 0,06* 0,08** 0,01 -0,04 0,02 0,00 0,08** 0,00 0,04 0,01 -0,00 0,00 -0,00 0,00 0,09***

a *** p<0,01, ** p<0,05, * p<0,10 (N=758)

32

Table 2

Association between HRM practices and groups resulting from cluster analysis

Variable Mean MeanGroup 1

MeanGroup 2

chi-2p-value

1. Promotion 0,66 0,67 0,65 0,573

2. Training 0,80 0,97 0,62 0,000

3. Grouptrain 0,36 0,58 0,14 0,000

4. Wagelevel 0,42 0,48 0,36 0,001

5. Plantinc 0,12 0,17 0,06 0,000

6. Knowpay 0,25 0,29 0,20 0,003

7. Team 0,47 0,62 0,31 0,000

8. Rotation 0,46 0,53 0,40 0,000

9. Selfins 0,90 0,93 0,86 0,001

10. Autonomy 0,27 0,36 0,18 0,000

11. Groups 0,39 0,68 0,08 0,000

12. Suggestion 0,57 0,84 0,29 0,000

13. Meeting 0,59 0,88 0,29 0,000

14. Survey 0,21 0,38 0,04 0,000

15. Opendoor 0,19 0,32 0,06 0,000

16. Selection 0,67 0,75 0,60 0,000

17. Security 0,28 0,29 0,26 0,422

N 758 381 377

33

Table 3

Mean, standard deviation and correlationsa between results variable, control variables, HRMG and NUMPRAC (N=621)

Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10

1. Lnsize 4,88 0,85

2. Automat 4,18 2,35 0,315***

3. Compet 3,48 0,89 -0,032 0,040

4. Qualassur 0,71 0,45 0,283*** 0,291*** 0,016

5. Efic 0,64 0,48 0,103** 0,129*** 0,014 0,146***

6. Qual1 0,65 0,48 0,072* 0,155*** 0,054 0,237*** 0,400***

7. Qual2 0,65 0,48 0,109*** 0,207*** 0,026 0,214*** 0,416*** 0,777***

8. Time1 0,68 0,47 0,079** 0,130*** 0,022 0,165*** 0,498*** 0,445*** 0,483***

9. Time2 0,73 0,45 0,095** 0,228*** 0,047 0,184*** 0,224*** 0,263*** 0,323*** 0,296***

10. HRMG 0,51 0,50 0,237*** 0,311*** 0,014 0,336*** 0,121*** 0,227*** 0,224*** 0,183*** 0,201***

11. NUMPRAC 7,64 2,80 0,316*** 0,347*** -0,020 0,351*** 0,139*** 0,181*** 0,208*** 0,157*** 0,219*** 0,765***

a *** p<0,01, ** p<0,05, * p<0,10

34

Table 4

Logit analysis for efficiency performancea

Model 1 Model 2 Model 3

b s.d. b s.d. b s.d.

Constant -1,356** 0,634 -1,70** 0,731 -1,921*** 0,738

Lnsize 0,210** 0,103 0,222* 0,120 0,202** 0,122

Automat 0,062* 0,034 0,070* 0,040 0,065* 0,040

Compet 0,008 0,086 0,019 0,097 0,025 0,097

Qualassur 0,444* 0,179 0,338* 0,206 0,323 0,206

HRMG 0,247 0,186

NUMPRAC 0,063** 0,035

Pseudo-R2 7 8,8 9,1

Chi-square 43,33*** 44,10*** 45,56***

Log L -514,18 -409,70 -408,96

N 666 666 666

a *** p<0,01, ** p<0,05, * p<0,10

35

Table 5

Logit analysis for quality performancea

QUAL1 QUAL2

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

b s.d. b s.d. b s.d. b s.d. b s.d. b s.d.

Constant -1,230* 0,727 -1,172 0,736 -1,552** 0,741 -1,335* 0,740 -1,228* 0,123 -1,667** 0,752

Lnsize -0,047 0,119 -0,093 0,120 -0,099 0,121 0,126 0,121 0,080 0,123 0,063 0,123

Automat 0,114*** 0,041 0,089** 0,042 0,094** 0,042 0,161*** 0,041 0,137*** 0,042 0,140*** 0,042

Compet 0,235** 0,101 0,245** 0,102 0,246** 0,101 0,122 0,100 0,131 0,101 0,134 0,101

Qualassur 0,836*** 0,208 0,698*** 0,213 0,745*** 0,211 0,639*** 0,204 0,494** 0,210 0,523** 0,208

HRMG 0,739*** 0,195 0,731*** 0,193

NUMPRAC 0,095*** 0,037 0,113*** 0,037

Pseudo-R2 18,4 21,1 19,7 15,4 18 17,1

Chi-square 94,48*** 109,05*** 101,22*** 78,19*** 92,65*** 87,68***

Log L -377,12 -369,83 -373,75 -387,37 -380,13 -382,63

N 659 659 659 664 664 664

a *** p<0,01, ** p<0,05, * p<0,10

36

Table 6

Logit analysis for based-time performancea

TIME1 TIME2

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

b s.d. b s.d. b s.d. b s.d. b s.d. b s.d.

CONSTANT -0,837 0,725 -0,786 0,731 -1,128 0,737 -0,978 0,755 -0,884 0,759 -1,362* 0,769

LNSIZE 0,067 0,120 0,036 0,121 0,021 0,121 -0,053 0,126 0,088 0,127 -0,109 0,128

AUTOMAT 0,101** 0,041 0,081* 0,041 0,082** 0,041 0,237*** 0,046 0,217*** 0,047 0,213*** 0,047

COMPET 0,100 0,099 0,106 0,100 0,108 0,100 0,176* 0,105 0,174 0,106 0,185* 0,106

QUALASSUR 0,595*** 0,204 0,466** 0,209 0,499** 0,208 0,584*** 0,208 0,461** 0,213 0,464** 0,212

HRMG 0,598*** 0,192 0,603*** 0,201

NUMPRAC 0,090** 0,036 0,115*** 0,139

Pseudo-R2 12,4 14,2 13,5 18,5 20,1 20,1

Chi-square 62,69*** 72,55*** 69,01*** 96,28*** 105,32*** 105,12***

Log L -392,17 -387,23 -389,01 -364,33 -359,81 -359,90

N 680 680 680 707 707 707

a *** p<0,01, ** p<0,05, * p<0,10