Supporting long-term workforce planning with a dynamic aging chain model: A case study from the service industry

SUPPORTING LONG-TERM

WORKFORCE PLANNING WITH A

DYNAMIC AGING CHAIN MODEL:

A CASE STUDY FROM THE

SERVICE INDUSTRY

A N D R E A S G R Ö ß L E R A N D A L E X A N D E R Z O C K

This study demonstrates how a dynamic, aging chain model can support stra-tegic decisions in workforce planning. More specifi cally, we used a system dy-namics model (a modeling and simulation technique originating from supply chain management) to improve the recruiting and training process in a large German service provider in the wider fi eld of logistics. The key fi ndings are that the aging chain of service operators within the company is affected by a variety of delays in, for example, recruiting, training, and promoting employees, and that the structure of the planning process generates cyclic phases of workforce surplus and shortage. The discussion is based on an in-depth case study con-ducted in the service industry in 2008. Implications are that planning processes must be fi ne-tuned to account for delays in the aging chain. The dynamic model provides a tool to gain insight into the problem and to improve the actual hu-man resource planning process. The value of the paper lies in the idea of apply-ing a well-known and quantitative method from supply chain management to a human resource management issue. © 2010 Wiley Periodicals, Inc.

Keywords: workforce planning, aging chain, system dynamics, simulation, delays

Introduction

Sy stem dynamics modeling and simu-lation have long been applied suc-cessfully in the field of supply chain management. System dynamics, however, is not bound exclusively to

this area of management; indeed, the meth-odology and insights gained in one area can be transferred to other structurally similar

fields. In this paper, we use an aging chain model of employees “flowing” through an organization to demonstrate the principal usefulness of system dynamics for human resource management. An aging chain is a modeling structure that represents the different stages of a maturation process. The structure contains information about the re-tention times it takes physical entities to flow

Correspondence to: Andreas Größler, Institute for Management Research, Radboud University Nijmegen, P.O. Box 9108, 6500 HK Nijmegen, The Netherlands, Phone: +31 24 36 16287, Fax: +31 24 36 11933, E-mail: [email protected]

Human Resource Management,Human Resource Management, September–October 2010, Vol. 49, No. 5, Pp. 829 – 848

© 2010 Wiley Periodicals, Inc.

Published online in Wiley Online Library (wileyonlinelibrary.com).

DOI: 10.1002/hrm.20382

830 HUMAN RESOURCE MANAGEMENT, SEPTEMBER–OCTOBER 2010

Human Resource Management DOI: 10.1002/hrm

through respective stages. In the case of em-ployees, the stages represent employees’ vari-ous levels of training and productivity, de-pending on their seniority in the organization (Sterman, 2000). This paper shows how a dy-namic model and the process of generating such a model can enhance decision making for long-term workforce planning.

Our research goal aligns with Cappelli’s (2008) concept of using ideas, insights, and methods from supply chain management for what he calls talent management. For Cap-pelli (2008), this concept deals with employ-ing the correct number of appropriately qualified people necessary to run an organi-zation successfully. More particularly, we shed some additional light on how modeling and simulation can benefit workforce plan-ning, as briefly indicated in Cappelli (2008, pp. 137–138).

The elaboration in this paper is based on a case study conducted involving a large Ger-man service provider. The performance of the service provided is based solely on the em-ployees’ skills and motivation. Thus, selecting and developing employees are important and difficult tasks hampered by an often sub-opti-mal workforce planning process. This, of course, has a major effect on operational per-formance. For example, the seemingly trivial problem of having an appropriate number of operational employees appears to be highly non-trivial. We demonstrate the usefulness of insights gained from system dynamics re-search in supply chain management to shed some light on the issues that occurred in the case-study company and to provide solutions to these issues.

The paper is structured as follows. In the first section, we briefly summarize the system dynamics method and demonstrate its proven effectiveness for supply chain management. We explain why system dynamics can be transferred to human resource management. The second section begins by presenting the case-study company and its workforce struc-ture. The major issues of its workforce plan-ning process are discussed and the actual and potential effects of these issues are examined. Next, we describe the process of building a dynamic model of workforce planning and

provide an overview of the model. Here, we highlight the implications for the case-study company resulting from the modeling and simulation project and propose several changes to the workforce planning policies. In the third section, we discuss the implications of the case study for using system dynamics as a method in human resource management and identify appropriate fields of application. The paper ends with a brief outline of future exten-sions of the model and related projects.

System Dynamics for Supply Chain and Human Resource Management

System dynamics is a theory on the structure (and resulting behavior) of social systems and a method to represent such structures as dia-grams and mathematical equations (Größler, Thun, & Milling, 2008; see also for the meth-od’s limitations). Originally, Forrester (1958, 1961) developed system dynamics for analyz-ing industrial enterprises (it was therefore initially called “industrial dynamics”). Today, it is applied to many kinds of systems that change over time, in particular to socioeco-nomic systems (for an overview, see Lane, 1999, 2007). Practical system dynamics proj-ects comprise two phases: conceptualization/modeling and simulation/scenario analysis.

System dynamics relates to feedback loops, accumulation processes, and delays as fundamental building blocks of all social systems. The concept of feedback loops em-phasizes that actions and results are usually bidirectional: an action leads to an outcome and the outcome forms the basis for new actions (although the effect of an action may sometimes appear to be remote in time and space). Accumulation processes highlight the notion that the state of a system frequently does not change instantaneously but is devel-oped gradually over time. The focus on de-lays depends on the former two components and highlights that in each system, time lags exist between decisions and actions or be-tween actions and results. For example, new information on the state of a system does not immediately lead to new decisions because the information needs time to be perceived, considered, and decided on.

SUPPORTING LONG-TERM WORKFORCE PLANNING WITH A DYNAMIC AGING CHAIN MODEL 831


Non-linear behavior of systems results from these three building blocks. System dy-namics provides ways to design formal mod-els and generate their behavior over time, that is, for simulating the models. To do so, a graphical syntax is used, in which flow (rate) and stock (state or level) variables are distin-guished. These two types of variables com-bine to produce so-called stock/flow diagrams (Forrester, 1968; Lane, 2000). Graphically representing systems using these diagram-ming techniques is a valuable tool for under-standing complex issues. Technically, by the quantification of variables and links between variables, a system of differential equations is created that is simulated with the help of numerical algorithms (Sterman, 2000). A va-riety of tests must underpin the validity of the resulting models. This validity, however, can be judged only relative to the purpose of a model (Barlas & Carpenter, 1990). Addi-tionally, prototypical process models support the process of conducting a system dynamics study (Forrester, 1994).

Many authors emphasize the importance of the modeling process (not just the result-ing model) to gain insights into a problem (Lane, 1995; Sterman, 1988). This emphasis on the process results because system dynam-ics models are usually descriptive and not prescriptive in the sense that they do not provide an “optimal” solution. This has two implications. First, insights from the model can be deduced only indirectly by formulat-ing its equations and running different simu-lations (in the sense of quantitative scenario analyses). Usually, simulation models do not produce the right answer to solve a problem as a direct output (which might be the case for some operational research type problems and models). Second, while being a modeling endeavor, system dynamics research is sub-stantially empirical: it does not involve devel-oping ideal models, but rather it depicts real-ity as precisely as possible. Based on a valid picture of what is the case in real organiza-tions, attempts are made to gain insight into how these organizations can be improved.

The first major publication in system dy-namics—Forrester’s article “Industrial Dy-namics: A Major Breakthrough for Decision

Makers” (Forrester, 1958)—is also one of the first published works in the field of supply chain management. Forrester described the flow of goods and orders in a four-tier supply chain. Investigating information feedback, he observed a dynamic behavior of the system—today commonly referred to as the bullwhip effect (Lee, Padmanabhan, & Whang, 1997; Sterman, 1989a)—in which stocks of inventory or order back-logs show amplifying oscillations over the tiers of a supply chain: while downstream, at the end-customer, the oscillations are modest, they grow bigger as they move upstream toward producers of components or raw materials. From 1958 to the present, research on supply chain management is-sues has made up a substantial part of the system dynamics liter-ature (e.g., Akkermans & Dellaert, 2005; Akkermans & Vos, 2003; Anderson, Fine, & Parker, 2000; Fowler, 1999; Morec-roft, 1983; Spengler & Schröter, 2003; Ster-man, 1989a; Towill, 1982, 1996; Wikner, Naim, & Towill, 1991).

For many years, system dynamics model-ing has also been employed for analysis of human resource issues (e.g., Abdel-Hamid & Madnick, 1991; Andersen & Emmerichs, 1982; Doman, Glucksman, Tu, & Warren, 2000; Gupta & Gupta, 1990; Hafeez & Abdelmeguid, 2003; Liou & Lin, 2008; Packer, 1964; Reid & Taylor, 1989; Runcie, 1980). For instance, Abdel-Hamid and Madnick (1991) analyzed the dynamics of large-scale soft-ware-development projects, paying particular attention to the processes of staffing (i.e., the number of people working on a project) and training (i.e., the quality of people working on a project). As another example, Liou and Lin (2008) discussed the implications of potential terrorist threats for the human resource management of companies, based on a system dynamics model.

A methodology used in supply chain management can be beneficially transferred to human resource management because sys-tem dynamics is capable of representing

Based on a valid

picture of what is

the case in real

organizations,

attempts are made

to gain insight

into how these

organizations can be

improved.



dynamic systems in general. In that sense, employees flowing through a number of organizational stages (for instance, represent-ing different hierarchical or training levels) possess operational characteristics that are

similar to those of material flow-ing along the many stages of a supply line. From this perspec-tive, the same abstract structures exist in the workforce as in mate-rial chains; for instance, delays, accumulation in stages, iterations to earlier stages, and information feedback between stages. Taking just one example, from a struc-tural/process-oriented point of view, the same factors are at work whether a finished product needs rework because of quality issues and reverts from final goods inventory to the assembly line or whether a staff member needs to brush up training on a specific topic to become operational in a particular service again. The same principal behavior results from the same structure in material chains as in workforce chains: overcrowding at certain stages, unfulfilled demand at other stages—in short, such chains are difficult to understand and man-age (Sterman, 1989a).

Of course, this similarity does not imply a materialistic view of people as being processed like parts in a production process, but certain phenomena arise simply because there is a certain number of people at one stage in an organizational aging chain, defined as a structure of consecu-

tive stages in which organizational mem-bers are considered together according to a specific criterion such as their level of train-ing, age, or hierarchy. Although all of the people in an organization are individuals, as a group, they can generate specific prob-lems, and it is these problems that are addressed by system dynamics projects. The use and utility of system dynamics for a

specific human resource management issue (workforce planning) are demonstrated in the following section based on a case study from the service industry. For that purpose, we chose a model structure (mainly consist-ing of an aging chain) that is commonly used in researching issues in supply chain management.

Case Study: System Dynamics Modeling of a Dynamic Chain of Operator Staff

Issues of Long-Term Workforce PlanningThe case-study company is a German service provider in the logistics industry that is the market leader in this industry in Germany and that has a quasi-monopolistic status. One of the major services the firm provides requires the availability of highly skilled op-erator staff. The company’s performance depends to a great extent on the timely and effective provision of this service; having the right number of operators constitutes a strategic capability and determines the per-formance of the company (Lopez-Cabrales, Valle, & Herrero, 2006). The intellectual and personal requirements that these operators have to meet are demanding, and the dura-tion of their training is long. Thus, selecting and developing employees are complicated matters resulting in a difficult and often sub-optimal workforce-planning process that has a major effect on operational perfor-mance (Ahmad & Schroeder, 2003). As a re-sult of preferable employment conditions and the company’s excellent reputation, however, there is no shortage of candidates willing to work in the company. To meet all the requests for its services, the company must recruit and train the right number of operators at the right times. This paper does not address how the company selects train-ees who have the ability to acquire the re-quired skills and are motivated to remain in the organization for a long time (cf. Barrick & Zimmerman, 2009) but focuses instead on the question of how many of such trainees should be hired.

The intellectual

and personal

requirements that

these operators

have to meet are

demanding, and the

duration of their

training is long.

Thus, selecting

and developing

employees are

complicated

matters resulting

in a difficult and

often sub-optimal

workforce-planning

process that has

a major effect

on operational

performance.



The workforce planning practice em-ployed at the case-study company was to assume continuous growth of demand for the future (for instance, assuming about 3% growth of demand) and, on the basis of this figure, to calculate for each geographi-cal division the deviation between the as-sumed operator capacity and their required operator capacity for a future period. The forecasting horizon was determined by the time required to train a newly hired opera-tor to be ready for operational services. In the case-study company, the human resource department considered this time lag to be 51 months without variation. If a geographical division had too few opera-tors, it requested new employees from the central staffing office, and, together with the central training facilities, the staffing office consolidated the requirements from the divisions and started the hiring and training process for new operator staff members.

The basic structure of the long-term work-force planning process shown in Figure 1 represents a control loop for the number of available staff members. The control loop in-corporates two delays established on the basis of the time it takes to recruit suitable candi-dates for in-house training (approximately 12 months) and the duration of the training. The latter incorporates two delays: the

duration of the basic training (approximately 15 months) and the on-the-job training (ap-proximately 24 months). A further complica-tion is caused by the fact that some staff members are needed for non-operational tasks such as training and projects. For in-stance, the company uses a trainer recruit-ment scheme that relies heavily on opera-tional staff members, but this approach can lead to further aggravation of a staff-shortage problem since times of shortage, when there are high numbers of trainees, require higher numbers of operational staff members for training purposes.

In the past few years, the case-study company has experienced a situation of overall growth in demand for its services, with brief, intermittent periods of growth decline or even stagnation. The case-study company’s managers described its long-term planning scheme over this period as sub-optimal primarily because they perceived the staff situation to be characterized by transient but prolonged periods of staff shortages, followed by periods of staff sur-pluses. Both situations are highly undesir-able as they can result in a declining service quality, excessive workforce costs, and lower productivity. In short, the seemingly trivial problem of providing just the right number of qualified people at the right time is actually highly non-trivial.

FIGURE 1. Schematic of the Long-Term Workforce Planning Task for the Service Provider



Because it concentrates on the question of providing the right number of the right people, this case study refers to the area of

strategic human resource plan-ning. According to Orpen (1993), strategic human resource plan-ning determines how many and what kind of people will be needed in the mid- and long term to achieve the company’s strategic objectives and designs efficient programs so workers can contribute to achieving the firm’s goals. Therefore, the scrutiny that is used when planning the use of equipment, material, and finan-cial resources is also applied to the planning of human resources (Ainsworth, 1995).

Design of System Dynam-ics–Based Intervention

The problem of long-term capac-ity planning for service workers in the case-study company is highly complex with respect to its dynamics. Therefore, the case-study company decided to com-plement its regular planning process, which consisted of fore-casting future workforce demand through spreadsheet analyses, with a modeling approach based on system dynamics, which fo-cuses on the analysis of the com-plex temporal behavior of key variables of social systems. The goals that were set by the case-study company’s management

for this study were to conduct a structural analysis of the existing long-term workforce planning process for service operators, to provide a dynamic analysis of the exist-ing planning policies, and to construct a scenario tool to improve the existing planning policies and the established risk-management approach accompanying the existing processes. To meet these requirements, the company defined the fol-lowing framework for the project:

Phase 1: Interviews with members of the case-study company with all the depart-ments involved.Phase 2: Interactive construction of a basic system dynamics model with the mem-bers of the case-study company.Phase 3: Testing and validation of the con-structed model for one service center of the case-study company.Phase 4: Testing and validation of the model for a second service center.Phase 5: Definition and evaluation of fu-ture scenarios to demonstrate the capabili-ties of the new planning tool.Phase 6: Possible roll-out to all remaining service centers in Germany.

This multiphase approach highlights the fact that system dynamics–based modeling frequently is much more than a back-office modeling exercise with some expert involve-ment. The main reason for this approach is the necessity of building intensive involve-ment in the organization to foster commit-ment and trust in such a new approach to organizational planning (Rouwette & Ven-nix, 2006; Snabe & Größler, 2006; Vennix, 1996; Vennix, Andersen, Richardson, & Rohrbaugh, 1992). The case-study company informed us that their intensive involve-ment created additional value—even if the constructed model had been discarded after the project was completed—as a result of the quality of the discussions that are facilitated in the organization through the modeling performed. Therefore, in the following, we report on the usefulness of both the process and the model (Lane, 1995; Sterman, 1988). The description of a simplified version of the model also incorporates some generalized simulation results that demonstrate the use-fulness of the chosen modeling approach.

Insights From the Model Building Process

The modeling project was conducted be-tween March and November 2008. The inter-views in the client organization took about two months to complete, and the modeling process itself took place primarily between

•

•

•

•

•

•

This multiphase

approach highlights

the fact that system

dynamics–based

modeling frequently

is much more than a

back-office modeling

exercise with some

expert involvement.

The main reason

for this approach

is the necessity of

building intensive

involvement in the

organization to

foster commitment

and trust in such

a new approach

to organizational

planning.



May and July 2008, followed by a testing and validation period that ended in September 2008. The subsequent scenario analysis was completed by November 2008, when the project was reviewed in order to determine whether to roll out to the company’s other service centers in Germany.

Most of the relevant data on the existing planning paradigm could be collected through interviews with stakeholders in the workforce planning process. Interviews are also a pre-requisite for the acceptance of all further steps in the client organization because they not only provide the project team with knowledge about the considered organiza-tional processes, but also help to build trust between the project team and the client orga-nization. Important insights that were gath-ered about workforce planning through these interviews included the following:

Different parts of the organization focus on different parts of the overall plan-ning process so none of the departments focuses on the architecture of the pro-cess as a whole (no “systemic picture”).Different sources of information had dif-fering views of important characteristics of the process, such as lead times in the planning process.Uncertainties in the process and condi-tions were mainly considered in a qualita-tive fashion.Management followed a “best bet” ap-proach when determining the most prob-able outcome of the planning process in-stead of using target corridors that reflected the uncertainty in the planning approach.During recent planning activities, organi-zational knowledge about past problems and their causes was neglected because it resides in the experience of line employ-ees in the staffing office, who are often in their positions for much longer periods than their managers are.

By presenting these insights at the begin-ning of the modeling process, we were able to build credibility for the subsequent steps. All the staff members and managers concerned felt involved and were pleased to share these “organizational truths” that, although not

•

•

•

•

•

new to the organization, had never been for-mulated in such a comprehensive manner.

The modeling process involved several working sessions with a group of six to ten people who went through the modeling exercises step by step. In the context of these ses-sions, several additional insights were gained:

There were data inconsisten-cies in the planning databases used by the local service cen-ters of the case-study company and the centralized planning department. These incon-sistencies had never before been laid bare, and only the rigorous character of the for-mal modeling approach led to a comparison of the data-bases in such detail that the problem could be identified.The overall lead time between the request for new employees and those employees’ becom-ing operational was much lon-ger than had been assumed. This insight also showed that the implemented planning ho-rizon was actually not long enough, which also brought up questions about the ad-equacy of the existing forecast methods for this time horizon.

In the course of the modeling sessions, most participants re-ported that the systemic represen-tation of the planning process cre-ated during the sessions provided an integrative picture of the pro-cess that had never before been provided to the organization. They also said that the quality of the discussion in the sessions was very high and had led to interdepart-mental dialogue that had not taken place before the project started.

In addition to these qualitative insights, a number of quantitative aspects of the plan-ning process were found to be of considerable

•

•

In the course of the

modeling sessions,

most participants

reported that

the systemic

representation

of the planning

process created

during the sessions

provided an

integrative picture

of the process that

had never before

been provided to

the organization.

They also said that

the quality of the

discussion in the

sessions was very

high and had led to

interdepartmental

dialogue that had

not taken place

before the project

started.



interest to the project members from the or-ganization. One example is the dynamic con-sequence of the variations in the lead times of the recruitment process. The overall lead time of this process was divided into three blocks: the recruitment lead time (12 months), the basic training (15 months), and the on-the-job training (24 months). If one assumes that these lead times are only average times and that they display some degree of vari-ance, the situation represented in Figure 2 arises. With lead times being considered aver-age times, the analysis (and the system dynamics model, as described in the next sec-tion) covers the individual variability among staff members.

Figure 2 shows the flow of 100 trainees through the overall recruitment process. It demonstrates that although the average lead time is 51 months, not all of the recruited operators are fully productive after this time. Although some may finish their training ear-lier, quite a few take longer to become opera-tional. (In the simulation it takes 68 months until about 100% of the ordered operator ca-pacity is actually available.) Although this insight may be intuitive when one considers the meaning of the term “average lead-time,” the planning process did not take it into ac-count before this simple simulation outcome was presented.

Other quantitative insights gained in the modeling process included the influence of limited capacity at the local center for on-the-job training, the effects of different poli-cies for distributing trainees across the cen-ters, and the occurrence of cyclic behavior in workforce capacity. Most of these issues were analyzed using the system dynamics model that is discussed in the next section.

A Simplifi ed Dynamic Simulation Model of the Workforce Planning Process

A simplified diagram of the system dynamics model that was developed in collaboration with the case-study company is depicted in Figure 3 and built using the Vensim software package. The figure shows a regular stock/flow diagram (Forrester, 1961; Lane, 2000) based on the conceptual understanding of is-sues represented in Figure 1. The central chain represents the flow of trainees from basic training to on-the-job-training to be-coming operational staff. Between these stocks (or stages in which staff may find themselves) are flows that represent when trainees were promoted through the chain and ultimately became fully operational. These flows are controlled by information feedback loops that symbolize the decision-making process in workforce planning. In the

100

75

50

25S

taff

00 10 20 30 40 50 60 70 80 90 100

Time (Month)

number of new staff to be hired

number of recruited trainees

number of trainees in basic training

number of trainees in OJT

number of trainees becoming operational

FIGURE 2. Flow of Staff Through the Different Stages of the Recruitment Process of Operational Staff



simplest case, these structures determine how long, on average, operators remain in one stage and when they move to the next stage. Since the movement of staff is calculated in average terms, these structures also account for the variability in the aging chain that are due to the differences among the trainees. This aging chain model resembles to a great degree system dynamics models that have been used in supply chain management to analyze issues such as excessive or depleted inventories, long delivery times, or quality-of-service problems.

The most relevant part of the figure to the context of this paper is the comparison be-tween actual operator capacity and required operator capacity, as represented by the vari-able gap, which determines the hiring policy of the company for new operator staff. This simplified model compares the current values for actual capacity (Operational staff) with required capacity (Required staff), and the

requests for new employees are based on the difference between these two variables.

The flow of operational staff toward the bottom right of the diagram (and back) repre-sents the fact that operators frequently have to fulfill non-operational tasks, such as train-ing, and are not productive during these times. Depending on its strength, this facet of the system can lead to a “worse-before-bet-ter” phenomenon and amplify oscillations in the number of operators available because an increase in the number of trainees requires more operators to take over training activi-ties, which reduces the overall service capac-ity in the short run (Doman et al., 2000; For-rester, 1992). For the simulation results shown here, however, the effect of non-operational tasks such as training does not have a signifi-cant effect.

The model used for running simulations has a mathematical equation linked to each variable in the diagram that characterizes the

FIGURE 3. Simplifi ed System Dynamics Model of the Workforce Chain in the Case-Study Company

Note: Not all model details are shown. Please see the Appendix for full equation listing.

Required staff

Operationalpersonnel

OperationalPersonnel (activein non-Ops areas)

Staff in BasicTraining (BT)

Temporary suspension rate

Recruiting rate

Duration ofbasic training

Gap

Recruiting Delay

Exit-Rate

Average DutyTime

Reactivationrate

Number of trainersrequired

Trainees pertrainer

Trainer-Gap

Staff in OJTCompletionrate (OJT)

CompletionRate (BT)

Duration of OJT

(re)activationtimes

Rectangles symbolize stocks, double arrows symbolize flows, single arrows symbolize information links.

and represent two different forms of external demand as scenarios:

continuously growing demand for operators demand cycles around the expected growth trajectory



variable’s behavior over time in relation to the development of other variables in the model. (See the listing of equations in the Appendix.) Of course, the simulation results presented in this paper depend on this parameterization, so they must be considered explicatory, not precise; however, the numerical values are similar to the real company data.

The two small line graphs at the top of the figure characterize two simulation runs that are used as an example of the results that can be gained from scenario analyses based on a system dynamics model. The graphs show the externally determined future devel-opment of demand—in this case, demand for operator staff. With this scenario set-up, we analyzed the support that human resource planning can provide for the growth of a firm (Mohrman, 2007).

The variable Required Staff acts as a sce-nario variable and is the only input that is exogenously changed in the following analy-ses, although other developments of the vari-able could be tested as well. The simulation results for the two scenarios are shown in Figure 4, which depicts the behavior of the two model variables of Staff in Basic Training (BT) and Operational Personnel. For reasons of confidentiality, the numerical values are exemplary only and do not correspond to the actual number of staff in the case company.

While in both scenario runs the number of operators increases in accordance with the demand for new staff, Figure 4 shows one

counterintuitive result: in both scenarios—that is, when there are no demand cycles (Sce-nario 1) and when there are demand cycles (Scenario 2)—the number of employees does not grow continuously but fluctuates around the growth trajectory. Thus, operator capacity may be characterized by phases of shortages and oversupply even when the external input into the system does not show such cycles.

The cycles can be found in all stages of the workforce chain. For instance, the num-ber of operators in basic training fluctuates with a very high amplitude with ups three times as high as downs (depicted on the left-hand side of Figure 4). Note the different scales of the two parts of the figure: in abso-lute terms, amplitudes for Operational Per-sonnel are roughly as high as for Staff in Basic Training; in relative terms, the oscillations for Staff in Basic Training are higher. The oscilla-tion in the number of trainees in basic train-ing has profound consequences for the plan-ning of training capacity, for instance, in terms of the number of trainers, rooms, and technical equipment. In addition, the cycles in the two scenarios do not correspond, so further analysis of the causes and conse-quences of cyclical behavior in the industry is required. (For an example from another in-dustry, see Liehr, Größler, Klein, & Milling, 2001).

As suggested by the system dynamics and supply chain literature, the cycles are gener-ated by the delays in the system in connection

FIGURE 4. Simulation Output for Two Exemplary Scenarios (note the different scales on the y-axes)

Operational personnelStaff in basic training (BT)

6,000

Scenario 2Scenario 1

600

Scenario 2Scenario 1

4,500450

3,000

1,500

Staf

f

300Staf

f

0

150

0

1 91 181 270 360 1 91 181 270 360

Time (Month)Time (Month)



with inadequate hiring requests for new op-erator staff. This case resembles what has be-come known as the bullwhip effect in supply chain management (Lee et al., 1997). Ster-man (1989a, 1989b) demonstrated that the bullwhip effect is generated by a faulty as-sumption about what is in the “supply line” (i.e., the operators who are currently in train-ing but will become operative later) in com-bination with faulty assumptions about the time it takes to close a perceived gap (i.e., how long it takes until operators become fully productive once their employment is requested). In addition, the consolidation of requests from the different divisions into one central hiring process may amplify the bull-whip effect, as Lee et al. (1997) showed. The modeling process made clear that a concept (and possible remedies) known in supply chain management can be transferred to the human resource management issue of long-term workforce planning.

The simulation model can be used to test for the effects of different delay times and/or hiring policies on the cyclicity of staff avail-able. For instance, Figure 5 shows how the system reacts to the inclusion of operators in training (Staff in OJT) in the calculation of the gap between required and available operators. For this figure, we calculated hiring requests based on the deviation between actual opera-tors plus operators in training (since they will become available soon) and required opera-

tors. As the figure shows, a more complete consideration of the “supply line” of opera-tors leads to a decrease in oscillations.

In both cases (Figure 4 and Figure 5), there is a gap between required and actual staff because the comparison is made based on current situations, and future demand growth is not taken into account. While in the case-study company the real hiring policies were not so simple as in Scenario 1 or Scenario 2, the “supply line” of operators in training and the length of their training process were not fully comprehended and considered. Thus, the model demonstrates the possible effects of changes in the policies and the structure of long-term workforce planning. (The model has been simplified for presentation in this paper; in the actual project a more elaborate version of the model was used to capture real company policies.)

Implications of Modeling and Simulation for Workforce Planning

Based on the simulation runs, the company is currently revising its workforce hiring poli-cies in light of the insights they gained dur-ing the course of the project. In particular, the variation in training times is being con-sidered, as well as the consolidation of re-quests from divisions into a central need for capacity. An actual comparison of results

6,000

Operational personnel

600

Modified hiring policyScenario 1Staff in basic training (BT)

Modified hiring policyScenario 1

4,500450

3,000

Staf

f

300Staf

f

1,500

0

150

0

1 91 181 270 360 1 91 181 270 360

Time (Month) Time (Month)

FIGURE 5. Simulation Output for Initial Scenario 1 and Modifi ed Hiring Policy Considering the “Supply Line” of Operators in Training (note the different scales on the y-axes)



achieved with the “old” workforce planning system and these revised policies cannot be made at this point, however.

When asked about the benefits of the described project, the representatives of the case-study company reported the following value propositions, which cover more as-pects of the process than originally were an-ticipated at the beginning of the project:

Using the implemented tool in the frame-work of the long-term planning process, a more detailed planning paradigm could be implemented at the departmental level instead of the central level to avoid the negative effects of consolidating the dif-ferent requests for new operator staff.The possibility of speeding up the planning process significantly allows for repeating the workforce planning cycle several times per year instead of only once per year.The risk management approach that ac-companies the long-term planning pro-cess can now be complemented by quan-titative scenarios that are provided almost in real time.The established systemic perspective on the planning process problem that inten-sified the communication among stake-holders is credited with starting a new organizational dialogue targeted at the workforce planning process.Some representatives from the case-study company expect that the new scenario tool can also act as a learning platform for the company to integrate the experi-ence and perspective of several depart-ments, providing a “holistic” view of the company (Senge, 1990).

Finally, within the framework of the dis-cussions at the end of the project, possible future applications of the tool were men-tioned by the project team members from the case-study company. The following applica-tions have good potential:

Use of the scenario tool in the context of strategic negotiations (e.g., in man-agement workshops, in discussions with trade unions);

•

•

•

•

•

•

Use of the scenario tool for training pur-poses, such as for “management flight sim-ulators” that allow management—similar to what pilots do in flight simulators—to experience different organizational situa-tions and test the effects of possible poli-cies (Bakken, Gould, & Kim, 1992; Maier & Größler, 2000); andUse of the scenario tool as part of a consul-tancy service offered to other companies in the industry (e.g., strategic partners).

These potential uses of the simulation model focus on the idea of having an in-strument to conduct quantitative scenario analyses. In principle, such studies are un-dertaken with the aim of exploring possible future developments in an organization or their effects on an organization, rather than with the purpose of exact forecasting (Schoe-maker, 1993; Schwartz, 1996; Wack, 1985a, 1985b). In contrast to conventional sce-nario studies, however, analyses based on simulation models provide quantitative in-sights (Lane, 1992; Morecroft & van der Heijden, 1992). While quantification in-creases the burden in scenario formulation because every variable needs a numerical value or a mathematical expression that de-termines its value, quantitative results am-plify the usefulness and acceptability in or-ganizational settings since many managers perceive qualitative methods as being too vague and remote from daily practice to be practical. The danger of having too much trust in quantification, however, should be dealt with as well (Mintzberg, 1994) by clarifying that the accuracy of simulation results depends on the quality of the quan-tification efforts and that—partially because models are, by definition, simplified repre-sentations of reality and partially because we do not have perfect foresight—not all possible influence factors and not all possi-ble future developments can be addressed. Nevertheless, providing quantitative sce-nario capabilities proved to be a relevant benefit offered by system dynamics–based simulation studies.

As with conventional scenario studies, the simulation option provides an additional

•

•



benefit over using other diagramming-based methods such as process mapping or Six Sigma approaches. While such approaches can result in a similar process to that described in the case study and are certainly beneficial in a variety of cases, having a dynamic model that can be simulated allows for extra out-comes regarding the actual or potential be-havior of the system. For instance, the insight that current hiring and training practices in the case organization naturally lead to oscilla-tion in the number of employees in each of the different stages could hardly be gained using static diagramming techniques.

Following the idea of providing an option for running scenario analyses, it is evident that system dynamics models usually cannot provide optimal solutions in the sense of the one and only best alternative to handling a problem; whenever the complexity of a sys-tem is too great, such an optimal solution can no longer be calculated. In organizational situations of relevance to practitioners, this limitation is nearly always encountered. Therefore, instead of trying to calculate an optimal solution to the workforce planning problem, the simulation model can help to improve the company’s understanding of its system and to design robust policies that may help to achieve success under a variety of scenarios (Forrester, 1961).

In summary, analysis of the workforce planning process of the case-study company based on a system dynamics modeling pro-cess yields the potential for significant changes in the organization and strategic le-verage for the competitiveness of the enter-prise. Thus, the system dynamics–based ap-proach presented here can provide a key capability in the field of strategic human re-source management, which is the focus of the next section.

Discussion: System Dynamics and Strategic Human Resource Management

Workforce planning models based on system dynamics aim to improve company perfor-mance and are part of what has been called strategic human resource management (see

Anthony & Norton, 1991; Lengnick-Hall, Lengnick-Hall, Andrade, & Drake, 2009; Sala-man, Storey, & Billsberry, 2005; Schuler & Jackson, 2007). Therefore, improvements are expected not only in terms of “classical” HR concepts of performance (such as employee-retention rate) but also for SCM-related perfor-mance measures, or even for the performance of the organization in the marketplace (giving such models strategic implications). For ex-ample, when one changes the workforce planning process in such a way that requested operators ar-rive at the locations where they are needed more quickly, we improve the typical SCM performance score of delivery time and, at the same time, guarantee that the quality of the service provided (as another SCM performance score) remains at an appropriate level by having enough operators to handle cus-tomer requests. Finally, while hav-ing an appropriate number of op-erators may increase labor costs in the short run, it does ensure a steady stream of revenues coming in from satisfied clients in the long run.

System dynamics as a model-ing and simulation method could also be fruitfully applied to areas of human resource management other than workforce planning. For instance, Vancouver (2008) de-scribed a dynamic theory of work motivation using concepts such as feedback loops and causal relationships between variables. Sys-tem dynamics offers a way to operationalize these ideas within a quantifiable model that allows for numerical analyses of the theory. On a more abstract level, Chadwick and Dabu (2009) presented a causal model of the links between human resource management and the firm’s competitive advantage. This causal model can be made dynamic and quantified using system dynamics, although the abstract nature of the concepts involved presumably allows only for a conceptual model.

Because system dynamics models consist of representations of real-world objects and relationships between objects, assumptions

Because system

dynamics models

consist of

representations of

real-world objects

and relationships

between objects,

assumptions about

the structure of

systems are made

transparent.



about the structure of systems are made transparent. Therefore, as suggested by Roeh-ling et al. (2005), system dynamics projects can be used to open up the black box and formulate theories on how workforce plan-ning can support the performance of a com-pany, enabling a formalized approach to strategic human resource management (Box-all, 2003; Fleetwood & Hesketh, 2008; Har-ney & Dundon, 2006; Lam & Schaubroek, 1998). As with applications in supply chain management, system dynamics models provide testable hypotheses on workforce

planning processes that may be used to improve the management of human resources.

In a similar way, by conducting system dynamics studies in workforce planning, Dipboye’s (2007) “outrageous statement #3” (HR science has not advanced understanding or application) can be addressed. The understanding of human resource issues can be improved through the use of system dynamics models because they are open to inspection; system dynamics models can be easily

scrutinized, advancing a rigorous approach to science. At the same time, they aim at improving real problem settings by providing a way to test alternative human resource management policies.

Ferris et al. (2007) and Mendenhall et al. (1998) suggested that non-linear relationships should be taken into account when conduct-ing research in the field of human resources. System dynamics models do not rely on the linearity of phenomena, and researchers are not forced to assume linear relationships. By their very nature, system dynamics comprises of non-linear behavior patterns that are caused primarily by feedback loops or non-linear links between variables. Instead of linear rela-tionships, functional relationships can either be formulated in an arbitrary mathematical form or graphically approximated, providing the opportunity to deviate from assumptions about linearity.

In the supply chain literature, there is a slowly growing interest in human resource

issues (e.g., Carter, Smeltzer, & Narasimhan, 2000; Large & Gimenez, 2006; Othman & Ghani, 2008; Shub & Stonebraker, 2009). Since many issues from both fields are struc-turally similar, they can be addressed using system dynamics, forging a bridge between the two disciplines. (By including financial variables, a link with finance might also be established.) Based on this view of system dynamics as a potentially common method for the fields of supply chain management and human resources management, system dynamics could help to identify research questions that are relevant to both fields.

Outlook

Although system dynamics stems from the area of supply chain management and has many applications there, it is also a valuable method in human resource management issues. We support this claim with the help of a case study in the workforce planning process of a service provider. For the case-study company, both the model and the modeling process proved useful in mitigating their problems.

Our future research efforts will concen-trate on transferring the insights gained in this modeling project to other industries and companies. We observed (and began to inves-tigate) a similar process within the workforce chain of a large transport-provider. In addi-tion, the simulation model can be extended to other issues in order to make it a more gen-eral test-bed for workforce planning policies:

The impact of differences between national legal systems or company cultures (cf. Tay-lor, Beechler, & Napier, 1996) that affect the length of the workday or the ease with which employees can be laid off in down-turns (in contrast to the virtually lifelong commitment of operators and a no-layoff policy in the case-study company; cf. Steel & Lounsbury, 2009);Differences in job complexity that could include studying low-profile jobs with short training times but high staff turn-over (as opposed to the long training times and nearly zero fluctuation in the case-study company);

•

•

For the case-study

company, both

the model and the

modeling process

proved useful in

mitigating their

problems.



Including the quality of training as a vari-able in the model, which could be related to the duration of training and/or the characteristics of the trainees (as opposed to assuming that every trainee will even-tually learn what needs to be learned; cf. Saks & Belcourt, 2006);Other exogenous demand patterns in ad-dition to the growth scenarios that were investigated, such as decreasing demand (cf. Feldman, 1996).

We hope to generalize our findings on how delays in the workforce planning process, com-bined with delays in the workforce aging chain, lead to systematic problems with human resource capacity. Here, it may be useful to combine such an analysis with questions con-cerning how other departments interact with the HR department and the implications of these kinds of interactions. Such a combined perspective may foster the idea of an integrated strategic resource management policy that in-cludes staff members. Reducing the complex-ity of the models even more may be a useful way to analyze this subject.

Another future strand of research deals with the relationship between workforce chains as discussed in this paper and service

•

•

supply chains. In the literature, human resources are identified to play a central role in securing service chain performance (Balta-cioglu, Ada, Kaplan, Yurt, & Kaplan, 2007; Ellram, Tate, & Billington, 2004; Heskett, Jones, Loveman, Sasser, & Schlesinger, 1994). Workforce chains are not the same as service supply chains, however; while the company that we described in our case study offers a service (and, thus, is establishing a service supply chain), the workforce chain, as such, does not constitute a service that the com-pany provides to its customers. Rather, the availability of qualified staff is a main require-ment of service quality, so managing the dynamics of the workforce chain is a require-ment for successfully offering a service. In addition, workforce chains combine features of physical supply chains and of service sup-ply chains, an issue that requires further investigation in subsequent papers (cf. also Chopra & Lariviere, 2005, for their idea of service inventory, and Frei, 2006, on the trade-off between efficiency and service).

Finally, workforce planning models should contribute to the understanding of a more general research issue: the explanation and management of cyclical behavior in in-dustries.

ANDREAS GRÖßLER is an associate professor at the Nijmegen School of Management, Radboud University, The Netherlands. He teaches operations management and method-ology courses at the undergraduate, master, and executive level. His research interests lie in the fi elds of strategic and international operations management, in applying sys-tem dynamics modeling and simulation in business and not-for-profi t organizations, and in investigating individual and organizational dynamic decision making. Dr. Größler is a managing editor of the System Dynamics Review.

ALEXANDER ZOCK, Ph.D., is currently acting as a managing director at the European Center for Aviation Development—ECAD GmbH, located in Darmstadt, Germany. He is responsible for all activities in the area of aviation management at ECAD, which comprise research, consultancy, and teaching at universities. His research interests lie in the fi eld of operations strategy and decision modeling in organizations, with special emphasis on the aviation industry. In addition, he is interested from a methodological point of view in using modeling and simulation techniques in the framework of organizational interven-tions. Before his engagement with ECAD, Dr. Zock worked seven years as a manager in the aviation industry.



References

Abdel-Hamid, T., & Madnick, S. (1991). Software project dynamics: An integrated approach. Engle-wood Cliffs, NJ: Prentice-Hall.

Ahmad, S., & Schroeder, R. G. (2003). The impact of human resource management practices on operational performance: Recognizing country and industry differences. Journal of Operations Man-agement, 21(1), 19–43.

Ainsworth, C. (1995). Strategic human resource plan-ning at Zeneca Pharmaceuticals. Management Development Review, 8(2), 11–15.

Akkermans, H., & Dellaert, N. (2005). The dynamics of supply chains and networks. System Dynamics Review, 21(3), 173–186.

Akkermans, H., & Vos, B. (2003). Amplifi cation in service supply chains: An exploratory case study from the telecom industry. Production and Operations Management, 12(2), 204–223.

Andersen, D. F., & Emmerichs, R. M. (1982). Analyzing US military retirement policies. Simulation, 39(5), 151–158.

Anderson, E. G., Jr., Fine, C. H., & Parker, G. G. (2000). Upstream volatility in the supply chain: The ma-chine tool industry as a case study. Production and Operations Management, 9(3), 239–261.

Anthony, P., & Norton. L. A. (1991). Link HR to corpo-rate strategy. Personnel Journal, 70(4), 75–86.

Bakken, B., Gould, J., & Kim, D. (1992). Experimenta-tion in learning organizations: A management fl ight simulator approach. European Journal of Opera-tional Research, 59(1), 167–182.

Baltacioglu, T., Ada, E., Kaplan, M. D., Yurt, O., & Kaplan, Y. C. (2007). A new framework for service supply chains. Service Industries Journal, 27(2), 105–124.

Barlas, Y., & Carpenter, S. (1990). Philosophical roots of model validation: Two paradigms. System Dynam-ics Review, 6(2), 148–166.

Barrick, M. R., & Zimmerman, R. D. (2009). Hiring for retention and performance. Human Resource Man-agement, 48(2), 183–206.

Boxall, P. (2003). HR strategy and competitive ad-vantage in the service sector. Human Resource Management Journal, 13(3), 5–20.

Cappelli, P. (2008). Talent on demand. Boston: Harvard Business School Press.

Carter, J. R., Smeltzer, L. R., & Narasimhan, R. (2000). Human resource management within purchasing management: Its relationship to quality manage-

ment success. Journal of Supply Chain Manage-ment, 36(2), 52–62.

Chadwick, C., & Dabu, A. (2009). Human resources, human resource management and the competitive advantage of fi rms: Toward a more comprehensive model of causal linkages. Organization Science, 20(1), 253–272.

Chopra, S., & Lariviere, M. A. (2005). Managing service inventory to improve performance. MIT Sloan Management Review, 47(1), 55–63.

Dipboye, R. L. (2007). Eight outrageous statements about HR science. Human Resource Management Review, 17(2), 96–106.

Doman, A., Glucksman, M. A., Tu, N.-L., & Warren, K. (2000). The talent-growth dynamic. McKinsey Quar-terly, 1, 106–115.

Ellram, L. M., Tate, W. L., & Billington, C. (2004). Un-derstanding and managing service supply chains. Journal of Supply Chain Management, 40(4), 17–32.

Feldman, D. C. (1996). Managing carriers in downsiz-ing fi rms. Human Resource Management, 35(2), 145–161.

Ferris, G. R., Perrewé, P. L., Ranft, A. L., Zinko, R., Stoner, J. S., Brouer, R. L. et al. (2007). Human resources reputation and effectiveness. Human Resource Management Review, 17(2), 117–130.

Fleetwood, S., & Hesketh, A. (2008). Theorising under-theorisation in research on the HRM-performance link. Personnel Review, 37(2), 126–144.

Forrester, J. W. (1958). Industrial dynamics: A major breakthrough for decision makers. Harvard Busi-ness Review, 36(4), 37–66.

Forrester, J. W. (1961). Industrial dynamics. Cam-bridge, MA: MIT Press.

Forrester, J. W. (1968). Principles of systems. Cam-bridge, MA: MIT Press.

Forrester, J. W. (1992). Policies, decisions, and infor-mation sources for modeling. European Journal of Operational Research, 59(1), 42–63.

Forrester, J. W. (1994). System dynamics, systems thinking, and soft OR. System Dynamics Review, 10(2/3), 245–256.

Fowler, A. (1999). Feedback and feedforward as sys-temic frameworks for operations control. Inter-national Journal of Operations and Production Management, 19(2), 182–204.

Frei, F. X. (2006). Breaking the trade-off between ef-fi ciency and service. Harvard Business Review, 84(11), 93–101.



Größler, A., Thun, J.-H., & Milling, P. M. (2008). System dynamics as a structural theory of strategic issues in operations management. Production and Opera-tions Management, 17(3), 373–384.

Gupta, Y. P., & Gupta, M. C. (1990). A process model to study the impact of role variables on turnover intentions of information systems personnel. Com-puters in Industry, 15(3), 211–238.

Hafeez, K., & Abdelmeguid, H. (2003). Dynamics of human resource and knowledge management. Journal of the Operational Research Society, 54(2), 153–164.

Harney, B., & Dundon, T. (2006). Capturing complexity: Developing an integrated approach to analyse HRM in SMEs. Human Resource Management Journal, 16(1), 48–73.

Heskett, J. L., Jones, T. O., Loveman, G. W., Sasser, W. E., & Schlesinger, L. A. (1994). Putting the service-profi t chain to work. Harvard Business Review, 72(2), 164–174.

Lam, S. S. K., & Schaubroek, J. (1998). Integrating HR planning and organisational strategy. Human Re-source Management Journal, 8(3), 5–19.

Lane, D. C. (1992). Modelling as learning: A consultancy methodology for enhancing learning in manage-ment teams. European Journal of Operational Research, 59(1), 64–84.

Lane, D. C. (1995). On a resurgence of management simulations and games. Journal of the Operational Research Society, 46(5), 604–625.

Lane, D. C. (1999). Social theory and system dynamics practice. European Journal of Operational Research, 113(3), 501–527.

Lane, D. C. (2000). Diagramming conventions in system dynamics. Journal of the Operational Research Soci-ety, 51(2), 241–245.

Lane, D. C. (2007). The power of the bond between cause and effect: Jay Wright Forrester and the fi eld of system dynamics. System Dynamics Review, 23(2/3), 95–118.

Large, R. O., & Gimenez, C. (2006). Oral communica-tion capabilities of purchasing managers: Meas-urement and typology. Journal of Supply Chain Management, 42(2), 17–32.

Lee, H. L., Padmanabhan, V., & Whang, S. (1997). The bullwhip effect in supply chains. Sloan Manage-ment Review, 38(3), 93–102.

Lengnick-Hall, M. L., Lengnick-Hall, C. A., Andrade, L. S., & Drake, B. (2009). Strategic human resource management: The evolution of the fi eld. Human Resource Management Review, 19(1), 64–85.

Liehr, M., Größler, A., Klein, M., & Milling, P. M. (2001). Cycles in the sky: Understanding and managing business cycles in the airline market. System Dy-namics Review, 17(4), 311–332.

Liou, D.-Y., & Lin, C.-H. (2008). Human resources planning on terrorism and crises in the Asia-Pacifi c Region: Cross-national challenge, reconsideration, and proposition from Western experiences. Human Resource Management, 47(1), 49–72.

Lopez-Cabrales, A., Valle, R., & Herrero, I. (2006). The contribution of core employees to organizational capabilities and effi ciency. Human Resource Man-agement, 45(1), 81–109.

Maier, F. H., & Größler, A. (2000). What are we talking about? A taxonomy of computer simulations to support learning. System Dynamics Review, 16(2), 135–148.

Mendenhall, M. E., Macomber, J. H., Gregersen, H., & Cutright, M. (1998). Nonlinear dynamics: A new perspective on IHRM research and practice in the 21st century. Human Resource Management Re-view, 8(1), 5–22.

Mintzberg, H. (1994). The rise and fall of strategic plan-ning. New York: Free Press.

Mohrman, S. A. (2007). Designing organizations for growth: The human resource contribution. Human Resource Planning, 30(4), 34–45.

Morecroft, J. D. W. (1983). Managing product lines that share a common capacity base. Journal of Opera-tions Management, 3(2), 57–66.

Morecroft, J. D. W., & van der Heijden, K. A. J. M. (1992). Modelling the oil producers: Capturing oil industry knowledge in a behavioural simulation model. European Journal of Operational Research, 59(1), 102–122.

Orpen, C. (1993). Managing human resources strategi-cally. Work Study, 42(5), 15–17.

Othman, R., & Ghani, R. A. (2008). Supply chain management and suppliers’ HRM practice. Sup-ply Chain Management: An International Journal, 13(4), 259–262.

Packer, D. (1964). Resource acquisition in corporate growth. Cambridge, MA: MIT Press.

Reid, W. M., & Taylor, R. G. (1989). An application of absorbing Markov analysis to human resource issues in public administration. Review of Public Personnel Administration, 10(1), 67–74.

Roehling, M. V., Boswell, W. R., Caligiuri, P., Feldman, D., Graham, M. E., Guthrie, J. P. et al. (2005). The future of HR management: Research needs and directions. Human Resource Management, 44(2), 207–216.



Rouwette, E. A. J. A., & Vennix, J. A. M. (2006). Sys-tem dynamics and organizational interventions. Systems Research and Behavioral Science, 23(4), 451–466.

Runcie, J. F. (1980). Dynamic systems and the quality of work life. Personnel, 57(6), 13–24.

Saks, A. M., & Belcourt, M. (2006). An investigation of training activities and transfer of training in or-ganizations. Human Resource Management, 45(4), 629–648.

Salaman, G., Storey, J., & Billsberry, J. (2005). Stra-tegic human resource management (2nd ed.). London: Sage.

Schoemaker, P. J. H. (1993). Multiple scenario develop-ment: Its conceptual and behavioral foundation. Strategic Management Journal, 14(3), 193–213.

Schuler, R. S., & Jackson, S. E. (2007). Strategic hu-man resource management (2nd ed.) Malden, MA: Blackwell.

Schwartz, P. (1996). The art of the long view. New York: Currency Doubleday.

Senge, P. M. (1990). The fi fth discipline: The art and practice of the learning organization. New York: Currency Doubleday.

Shub, A. N., & Stonebraker, P. W. (2009). The human impact on supply chains: Evaluating the impor-tance of “soft” areas on integration and perform-ance. Supply Chain Management: An International Journal, 14(1), 31–40.

Snabe, B., & Größler, A. (2006). System dynamics modelling for strategy implementation. Systems Research and Behavioral Science, 23(4), 467–481.

Spengler, T., & Schröter, M. (2003). Strategic manage-ment of spare parts in closed-loop supply chains: A system dynamics approach. Interfaces, 33(6), 7–17.

Steel, R. P., & Lounsbury, J. W. (2009). Turnover process models: Review and synthesis of a con-ceptual literature. Human Resource Management Review, 19(4), 271–282.

Sterman, J. D. (1988). A skeptic’s guide to computer modeling. In L. Grant (Ed.), Foresight and national

decisions (pp. 133–169). Lanham, MD: University Press of America.

Sterman, J. D. (1989a). Modeling managerial behavior: Misperceptions of feedback in a dynamic decision making experiment. Management Science, 35(3), 321–339.

Sterman, J. D. (1989b). Misperceptions of feedback in dynamic decision making. Organizational Behavior and Human Decision Processes, 43(3), 301–335.

Sterman J. D. (2000). Business dynamics: System thinking and modeling in a complex world. Boston: McGraw-Hill.

Taylor, S., Beechler, S., & Napier, N. (1996). Toward an integrative model of strategic international human resource management. Academy of Management Review, 21(4), 959–985.

Towill, D. R. (1982). Dynamic analysis of an inven-tory and order based production control system. International Journal of Production Research, 20(6), 671–687.

Towill, D. R. (1996). Industrial dynamics modelling of supply chains. Logistics Information Systems, 9(4), 43–56.

Vancouver, J. B. (2008). Integrating self-regulation theories of work motivation into a dynamic process theory. Human Resource Management Review, 18(1), 1–18.

Vennix, J. A. M. (1996). Group model building. Chich-ester, England: John Wiley & Sons.

Vennix, J. A. M., Andersen, D. F., Richardson, G. P., & Rohrbaugh, J. (1992). Model-building for group decision support: Issues and alternatives in knowl-edge elicitation. European Journal of Operational Research, 59(1), 28–41.

Wack, P. (1985a). Scenarios: Uncharted waters ahead. Harvard Business Review, 63(5), 72–80.

Wack, P. (1985b). Scenarios: Shooting the rapids. Har-vard Business Review, 63(6), 139–148.

Wikner, J., Naim, M., & Towill, D. R. (1991). Smoothing supply chain dynamics. International Journal of Production Economics, 22(3), 231–249.



A P P E N D I X Model Listing

Variable Defi nition

Unit of

Measurement Comments

(re)Activation times 1 Month Time it takes to (re)activate operators to conduct (non)operational tasks

Amplitude 0 Dimensionless Cycle generation in Scenario 2

Average Duty Time 360 Month Time operators work in their job

Completion Rate (BT)

Staff in basic training (BT)/duration of basic training

Staff/month Trainees fi nishing their basic training

Completion rate (OJT)

Staff in OJT/duration of OJT Staff/month Trainees fi nishing their on-the-job training

Cycle SIN(2*3.1415*(Time-Offset)/Period)*Amplitude*Demand

Activities Cycle generation in Scenario 2

Demand INTEG (Growth Rate, 3e+006) Activities Activities demanded from operators

Duration of basic training

16 Month

Duration of OJT 24 Month

Exit-Rate Operational personnel/average duty time

Staff/month Operators leaving the organiza-tion (retiring)

Final Time 360 Month The fi nal time for the simula-tion—technical variable

Fractional growth in %

0.0025 1/month Growth fraction of demand per month

Gap IF THEN ELSE (Required staff-Opera-tional personnel > 0, Required staff-Operational personnel, 0)

For modifi cation, Fig. 5: IF THEN ELSE(Required staff-Operational personnel-Staff in OJT+(Operational personnel+Staff in OJT)/Average Duty Time*Duration of OJT > 0, Required staff-Operational per-sonnel-Staff in OJT+(Operational personnel+Staff in OJT)/Average Duty Time*Duration of OJT, 0)

Staff Calculates gap between actual and required operator capacity.

Growth rate Fractional growth in % *Demand Activities/Month

Growth in demand

Initial time 1 Month The initial time for the simula-tion—technical variable

Number of trainers required

Staff in basic training (BT) /trainees per trainer

Staff

Offset 96 Month Cycle generation in Scenario 2

Operational personnel

INTEG (Completion rate (OJT)+Reactivation rate-Temporary suspension rate-Exit-Rate, 1800)

Staff Staff really being operational (i.e., working on demand activi-ties).

Appendix Continued on Next Page



A P P E N D I X Model Listing (Continued)

Variable Defi nition

Unit of

Measurement Comments

Operational per-sonnel (active in non-ops areas)

INTEG (Temporary suspension rate-Reactivation rate, 5)

Staff Staff fulfi lling non-operational activities, in particular training of new hires

Period 132 Month Cycle generation in Scenario 2

Productivity of personnel

1666 Activities/per-son

How many activities one operator can fulfi ll

Reactivation rate IF THEN ELSE(Trainer-Gap < 0, -Trainer-Gap, 0)/(re)activation times

Staff/month Trainers becoming operational again

Recruiting delay 12 Month How long it takes to hire new operators on average

Recruiting rate Gap/Recruiting Delay+Exit-Rate Staff/month Hiring policy

Required staff Demand/Productivity of personnel+Cycle/Productivity of per-sonnel

Staff Operator capacity required

Staff in basic training (BT)

INTEG (Recruiting rate-Completion Rate (BT), 50)

Staff New hires being in basic training

Staff in OJT INTEG (Completion Rate (BT)-Completion rate (OJT), 8)

Staff Trainees being in on-the-job training

Temporary suspension rate

IF THEN ELSE(Trainer-Gap > 0, Trainer-Gap, 0)/(re)activation times

Staff/month Operators becoming non-operational (trainers)

Time step 0.0625 Month Time step for the simulation—technical variable

Trainees per trainer

10 Dimensionless [staff/staff]

Trainer-gap Number of trainers required-opera-tional personnel (active in non-ops areas)

Staff Number of trainers needed to train new hires

Note: Variables are listed in alphabetical order.

Documents

Supporting long-term workforce planning with a dynamic aging chain model: A case study from the service industry