Workforce forecasting models: A systematic review

• Date: 01 November 2018
• Author: Anahita Safarishahrbijari
• DOI: http://doi.org/10.1002/for.2541
• Published date: 01 November 2018

RESEARCH ARTICLE

Workforce forecasting models: A systematic review

Anahita Safarishahrbijari

Department of Computer Science,

University of Saskatchewan, Saskatoon,

Saskatchewan, Canada

Correspondence

Anahita Safarishahrbijari, Department of

Computer Science, 110 Science Place,

University of Saskatchewan, Saskatoon,

SK S7N 5C9, Canada.

Email: anahita.safari@usask.ca

Abstract

Workforce analytics involves using models that integrate internal and external

data to predict future workforce and help organizations in any industry exam-

ine factors that have a prognostic effect. This paper assesses workforce model-

ing and prediction methods by examining their rationale, strengths, and

constraints. It aims to identify enhancements for further development of work-

force forecasting models and compares the capacity and reliability of different

forecasting methods. Past and present modeling trends are described and cri-

tiqued based on their relevance to current requirements. Several approaches

are reviewed, such as time series modeling and system dynamics simulation.

Sensitivity analysis in models is assessed. The models are decomposed into

three modes: supply‐based, demand‐based, and need‐based, which in some

cases provide substantially different estimates of future workforce need. The

chronological progression of models'development is analyzed. The articles

are also classified based on the countries and the sectors that have paid great

attention to workforce prediction research. Consideration of the use of work-

force models and the inputs into such models is not within the scope of this

paper.

KEYWORDS

labor forecasting, manpower planning, organizational studies, workforce, workforce modeling

1|INTRODUCTION

Workforce supply and demand modeling can help inform

human resource (HR) planning and decision making.

One of the principal reasons to use workforce forecasting

models is to identify the potential gap between the pres-

ent workforce and future requirements in order to predict

training and recruitment needs. The question as to

whether the statement “Rising labor costs in labor‐inten-

sive industries”is true has increased the concern about

the cost effectiveness of the delivery of services. Since

workforce models can involve a high degree of complex-

ity, tradeoffs across multiple objectives, and numerous

uncertainties, the forecasting and modeling methods in

this area have evolved over time, becoming more complex

and taking into account external forces and factors.

Depending on the purpose of the model and the questions

to be answered, system factors, data requirements, and

modeling methods will vary. The scope of such models

can be categorized into three major perspectives: supply

based, demand based, and need based. In supply‐based

models the goal of projection is determining the number

of staff available at a particular time or period of time

based on training, retirement, and promotion or demo-

tion. Supply‐based models provide a response to “How

many providers will there be?”Demand‐based models

deal with prediction of future service requirements and

likely changes in the demand and provide an answer to

“How many supplies will we need?”In addition, needs‐

based estimates use an exogenous benchmark to judge

the adequacy of the number of staff required to meet

the defined targets. Need‐based models respond to

Received: 20 October 2017 Revised: 23 June 2018 Accepted: 30 June 2018

DOI: 10.1002/for.2541

Journal of Forecasting. 2018;37:739–753. © 2018 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/for 739