Volume 24 Issue 1 Article 4 
March 2022 
Innovation and Productivity: Is Learning by Doing Over? 
Ružica Bukša Tezzele 
Juraj Dobrila University of Pula, Faculty of Economics and Tourism "Dr. Mijo Mirković", PhD Candidate, 
Pula, Croatia, ruzica.buksa@gmail.com 
Follow this and additional works at: https://www.ebrjournal.net/home 
 Part of the Labor Economics Commons 
Recommended Citation 
Bukša Tezzele, R. (2022). Innovation and Productivity: Is Learning by Doing Over?. Economic and 
Business Review, 24(1). https://doi.org/10.15458/2335-4216.1297 
This Original Article is brought to you for free and open access by Economic and Business Review. It has been 
accepted for inclusion in Economic and Business Review by an authorized editor of Economic and Business 
Review. 
ORIGINAL ARTICLE
Innovation and Productivity: Is Learning by
Doing Over?
Ru
zica Buk
sa Tezzele
Juraj Dobrila University of Pula, Faculty of Economics and Tourism “Dr. Mijo Mirkovic”, PhD Candidate, Pula, Croatia
Abstract
Labour productivity is one of the key measures of economic performance. It represents the total volume of output
produced per unit of labour. This paper examines the influence of business investments in research and development
and education on labour productivity using system dynamics modelling. The results reveal that investments in edu-
cation and training activities generate higher labour productivity growth. The impact of innovations largely depends on
their diffusion and adoption that require educated and trained users. The new industrial era makes learning by doing
quietly disappear and demand a great flexibility of workers and their ability to rapidly acquire new and master the
existing knowledge.
Keywords: Labour productivity, Innovation, Research and development, Education, System dynamics modelling
JEL classification: J24, O30, O52
Introduction
I n macroeconomics, labour productivity is one ofthe most used indicators for dynamic measuring of
economic growth, competitiveness and living stan-
dard within an economy. Labour productivity is a
value that each employedperson creates per his or her
input. It measures the amount of real gross domestic
product (GDP) produced by an hour of labour or by a
person during a given reference period. Labour pro-
ductivity provides general information about the ef-
ficiency and quality of human capital in the
production process for a given economic and social
context (International Labour Organization, 2019).
The most important determinants of labour pro-
ductivity are physical and human capital and tech-
nological change. Physical capital implies the tools,
equipment and facilities available to workers to
produce goods and services. Human capital repre-
sents the accumulated knowledge, skills and
expertise that workers possess. Technological
change implies a combination of invention (ad-
vances in knowledge) and innovation (putting the
advancement into a new product or service). If
labour productivity is growing, it can usually be
attributed to the growth in one of these de-
terminants. However, if the real gross domestic
product is increasing while labour hours remain
static, it signals that the labour force has become
more productive. Such situations could be observed
during economic recessions as workers increase
their labour effort to avoid losing their jobs.
After the World War II, between 1950 and 1995,
labour productivity throughout Europe grew faster
than in the USA (by 3.1% in France, Germany,
Belgium, Italy and the Netherlands, while by 2.1% in
the USA). The trend reversed later on, as labour
productivity in Europe slowed down, while the pro-
ductivity in the USA rose and even with slowdowns
always remained higher than in Europe. Labour
productivity growth in the USA averaged 1.1% from
2007 to 2017, while in the EU-15 only 0.6% in the same
period, widening the already existing productivity
gap. According to the Information Technology and
Innovation Foundation (ITIF), this effect is largely
attributable to Europe's failure to invest in informa-
tion and communication technologies which drives
labour productivity (Atkinson, 2018).
Received 30 November 2020; accepted 8 September 2021.
Available online 17 March 2022.
E-mail address: ruzica.buksa@gmail.com
https://doi.org/10.15458/2335-4216.1297
2335-4216/© 2022 School of Economics and Business University of Ljubljana. This is an open access article under the CC-BY-NC-ND license (http://creativecommons.
org/licenses/by-nc-nd/4.0/).
The concept of “learning by doing” is based on
learning from experiences resulting from one's own
actions (Reese, 2011). Learning takes place through
the attempt to solve a problem and it only takes
place during activity (Arrow, 1962). The role of
experience in increasing productivity has not gone
unobserved by scientists. In his work published in
1936, Wright analysed and concluded that the
number of labour hours expended in the production
of an airframe is a decreasing function of the total
number of airframes of the same type previously
produced. Furthermore, Lundberg introduced the
“Horndal effect” in 1961 that indicates the process of
gradual increase in output despite the lack of in-
vestments. Lundberg observed that productivity at
the Horndal steel works in Sweden increased on
average close to 2% per year for a period of 15 years,
despite the lack of significant capital investments
(Hendel & Spiegel, 2013). In his work on the theme
of “learning by doing”, a Nobel prize winner in
economic sciences Kenneth Arrow argued that the
sustained productivity growth at Horndal could
only be attributed to learning from experience
(Arrow, 1962). Arrow provided foundational work in
endogenous growth theory which holds that in-
vestments in human capital, knowledge and inno-
vation are significant contributors to economic
growth.
Early researches of the concept of “learning by
doing” have mostly been performed in the
manufacturing sector and through a progress function
determining variation in production costs over the
years (Hirsch, 1952; Montgomery, 1943; Wright, 1936).
In further studies (Bahk & Gort, 1993; Rapping, 1965;
Sheshinski, 1967), learning by doing has been
acknowledged as a productivity enhancing factor in
the production function. However, learning is not
continuous in the production process; it can be
depreciated or interrupted. When a technological
change occurs in the production, past knowledge and
experience can become irrelevant. Knowledge and
skills passed to workers can be lost when they leave
the job. Therefore, recent studies try to investigate the
effects of learning by doing and labour turnover on
productivity (Baffoe-Bonnie, 2016; Chiang, 2004; Da
Rocha, Pero, & Corseuil, 2019).
This paper contributes to the existing literature in
several ways. First, existing studies on labour pro-
ductivity and innovation relationship, as well as on
labour productivity and investments in human capital
have largely focused on production function estima-
tion. The purpose of this paper is to challenge and
advance these studies by using system dynamics
modelling. System dynamics modelling is based on
system thinking and enables viewing problems and
human actions as interconnected wholes, taking into
account also lagged feedback loops that is difficult to
do through traditionally used production functions. It
enables understanding the big picture around the
problem and predicting its long-term change. A sys-
tems approach enables to identify where to focus ac-
tions and with what intensity to reach the best
desirable results. System dynamics modelling has
been widely applied to better understand complex
system issues from organisational change to urban
and world dynamics, including climate change (Lane
& Sterman, 2011; Sterman et al., 2013). This paper
presents a dynamicmodel thatmay beused to analyse
the influence of investments in education and
research and development (R&D) on labour produc-
tivity and can easily be applied by companymanagers
and decisionmakers for long-term strategic planning.
What often prevents companies from achieving
higher performance is not a lack of resources,
knowledge or commitment to change, but a lack of
meaningful system thinking capability (Sterman,
2002). Therefore, this paper could also serve to stu-
dents to better understand system dynamics model-
ling and its application. Second, the model also takes
into consideration labour turnover and thus contrib-
utes to the recent studies on learning by doing, labour
turnover and productivity relationship. Third, a better
understanding of the effects of the above relationswill
make policymakers, managers and company owners
aware of the importance of investing in continuous
education and innovation and allow them to adjust
their mechanisms, improve their employment de-
cisions, provide support and work on key issues that
contribute to labour productivity growth. By
increasing labour productivity, they will directly
contribute to the growth of income per capita and
living standard in their countries.
The paper is composed of five chapters. After the
introduction, the first chapter gives a brief overview of
the relevant literature. The second chapter presents
the developed model. The results and discussion are
presented in the third chapter. The last chapter sum-
marizes the most important implications.
1 Theoretical background
Labour productivity is directly linked to economic
growth, as the latter most frequently occurs when
labour productivity increases. With growth in la-
bour productivity, an economy is able to produce
more goods and services for the same amount of
work. The additional production enables higher
consumption of these goods and services.
Labour productivity is important for everyone in
an economy: businesses, workers and government.
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 53
Increased labour productivity brings businesses
higher profit and thus an opportunity for more in-
vestments, while for workers it can translate into
higher wages, better working conditions, as well as
new job positions. For the government, increased
labour productivity results in higher tax revenues.
The amount of tax revenue received by the gov-
ernment is dependent upon the extent to which
productivity grows over a period and paying atten-
tion to trends in productivity is therefore vital for
policymakers. If productivity grows by less than
expected, fiscal deficits could occur, while if it grows
by more than expected, there may be fiscal sur-
pluses (Sprague, 2014).
According to the Bureau of Labor Statistics of the
United States Department of Labor, workers in the
United States business sector worked the same
number of hours in 2013 as they had in 1998
(approximately 194 billion labour hours). During
that period, the United States population increased
by over 40 million people and thousands of new
businesses were established. However, the United
States businesses managed to produce 42% ($ 3.5
trillion) more output in 2013 than in 1998 (adjusted
for inflation). Such additional output growth must
have come from productive sources other than the
number of labour hours, i.e. investing in technology,
equipment, hiring more high-skilled and experi-
enced workers (Sprague, 2014).
Drawing on the literature, the conceptual frame-
work is composed of two key determinants of labour
productivity: innovation, including research and
development and education.
1.1 Innovation, research and development and
labour productivity
Innovation can be defined as the implementation
of creative ideas in order to generate value (Baum-
gartner, 2009). R&D refers to innovative activities
undertaken by companies or governments in
developing new or improving existing products or
services. R&D is a crucial component of innovation
and a key factor in developing new competitive
advantages. It directly supports the development of
knowledge and technology.
Many studies have already been done on inno-
vation and labour productivity as well as investment
in R&D and labour productivity relationship. Many
of them reach the conclusion that innovation and
investment in R&D do matter. Crepon, Duguet, and
Mairesse (1998) gave one of the most influential
contributions in recent literature on economics of
innovation. They introduced a structural model that
explains productivity by innovation output and
innovation output by research investment. By the
use of data on innovation output in French
manufacturing and econometric methods which
correct for selectivity and simultaneity biases, the
authors revealed that innovation output rises with
research investment and productivity correlates
positively with a higher innovation output. L€o€of and
Heshmati (2006) investigated the relationship be-
tween innovation and company's performance in
the multidimensional sensitivity analysis, showing
that the model of Crepon, Duguet and Mairesse can
be estimated in a simple framework, applying the
instrumental variables approach.
Mohnen and Hall (2013) did a brief survey of the
empirical literature on innovation and productivity. In
spite of the usual positive relationship between
innovation and productivity, in the EUmember states
high costs of innovation often lead to lower com-
panies’ productivity. Moreover, in Central and
Eastern European countries, a negative feedback ef-
fect from productivity to innovation output has been
noticed which could indicate that companies rather
improve production of existing products than intro-
duce new products to the market (Hashi & Stoj
cic,
2013). Crespi and Zuniga (2012) have examined the
determinants of technological innovation and its
impact on labour productivity across six Latin Amer-
ican countries (Argentina, Chile, Colombia, Costa
Rica, Panama and Uruguay). They found out that
companies that invest in knowledge are abler to
introduce new technological advances and that those
that innovate have greater labour productivity.
There are two types of innovation output: product
and process innovations. Product innovation is the
main driver of labour productivity and the influence
of process innovation is rather insignificant (Bau-
mann & Kritikos, 2016; Mairesse & Robin, 2009).
However, process innovations are important, as
they improve the transformation process (Kemp et
al., 2003; Klomp & Van Leeuwen, 2001). Some
companies perform both, product and process in-
novations, especially in manufacturing. Stoj
cic and
Hashi (2014) investigated the influence of product,
process and both product and process innovations
on productivity across several East and West Euro-
pean countries. The results revealed that there is a
positive and statistically significant relationship be-
tween process innovations, as well as product and
process innovations together, and productivity.
Companies tend to be more successful in product
innovation if they use customers as a source of in-
formation and in process innovation if they use in-
formation from suppliers (Griffith et al., 2006).
The econometric results obtained so far indicate
the existence of a positive and significant
54 ECONOMIC AND BUSINESS REVIEW 2022;24:52e63
relationship between R&D and company and labour
productivity growth (Solomon et al., 2015; Tsai &
Wang, 2004; Wakelin, 2011). Both applied and
experimental R&D, as well as R&D from private
sources tend to have higher productivity impacts
compared to the basic R&D and publicly funded
R&D (Coccia, 2011; Solomon et al., 2015). Although
many studies have shown positive correlation be-
tween investment in R&D and labour productivity,
Benavente (2006) analysed the impact of spending
on R&D and innovation on labour productivity
using data from Chilean companies. He found out
that in the short term, productivity is not affected by
innovation or by spending on R&D. The explanation
for such results he found in the fact that there were
no lags between the implementation of innovations
and impact on productivity. Similar results were
obtained in the study of Erdil, Cilasun, and Eruygur
(2013) in which the contribution of R&D expendi-
tures to labour productivity in 22 OECD countries
was analysed. The results showed that the initial
impact of the growth in the ratio of total R&D
expenditure to GDP on labour productivity is
negative and insignificant. However, that situation
ceases to exist after a period of time and labour
productivity starts to rise. In this respect, the pol-
icies that favour R&D could help to increase the
productivity.
1.2 Education and labour productivity
Human capital is a very important contributor to
the productivity growth (Engelbrecht, 1997; Pietrzak
& Balcerzak, 2016). Rapid technological change
makes many skills obsolete quite quickly, while
constantly creating demand for new ones. Even
though the willingness to participate in adult
learning increases with the worker's socio-economic
status and the given level of education (Desjardins,
2015), workers usually struggle to improve their
knowledge and skills, especially because of their
age, lack of time, will and finance resources. In order
to improve their ability to absorb new technologies
coming out of R&D and keep pace with the
competition, companies often invest in education
and training of their workers.
The U.S. National Center on the Educational
Quality of the Workforce (EQW) studied the rela-
tionship between education and productivity at
more than 3100 U.S. workplaces. In a report pub-
lished in 1995, it showed that, on average, a 10%
increase in workforce education level led to an 8.6%
gain in total factor productivity, while a 10% rise in
physical capital (equipment) increased productivity
by just 3.4%. Therefore, the marginal value of
investing in human capital is almost three times
greater than the value of investing in machinery
(Stewart, 2010). Further studies (Almeida & Car-
neiro, 2009; Galor & Moav, 2004; Stauvermann &
Kumar, 2018) confirmed the importance of investing
in human capital. However, according to Duran and
Rillaers (2002), investing less in education does not
necessarily lead to less growth, as the accumulation
of physical and human capital displays some degree
of substitutability.
Some recent studies (Afrooz et al., 2010; Benos &
Karagiannis, 2016; Maciulyte-Sniukiene & Matuze-
viciute, 2018; Tabari & Reza, 2012) have shown that
human capital in terms of education has a positive
impact on labour productivity growth. Human
capital is composed of general and company-spe-
cific knowledge. General knowledge is acquired
during schooling, while company-specific knowl-
edge is acquired through worker's on-the-job
experience, training or through specialised courses.
High rate of turnover (loss of experienced workers)
has negative effects on learning and labour pro-
ductivity (Chiang, 2004; Da Rocha, Pero, & Corseuil,
2019). With the high turnover, companies lose
company-specific human capital and it becomes
harder to convert “doing” into “learning”. General
human capital is easy to compensate by employing
new workers, but only by retaining experienced
workers can companies convert their “doing” into
“learning” (Chiang, 2004).
Bartel (1995) studied the relationship between on-
the-job training and worker productivity and found
that training has a positive and significant effect on
job performance, thereby confirming the relation-
ship between training and productivity. It is partic-
ularly important that companies increase their
investments in on-the-job training and learning by
doing of workers with secondary school diploma
(Yunus, Said, & Hook, 2014). The skills acquired
through training should be applicable across all
possible departments in the organization to engage
the workplace and increase productivity.
Although most of the countries have increased
efforts in adult learning, the share of adults who
participated in education programs and courses for
certain European countries was in the range be-
tween 24.2% (Italy) and 66.7% (Denmark) in 2012.
The Nordic countries, more precisely, Denmark
(66.7%), Sweden (65.9%), Finland (65.4%),
Netherlands (64.9%) and Norway (64.7%) registered
the highest rates of adult participation in education
activities (Desjardins, 2015). Despite their high level
of economic development, the Nordic countries
have a strong record of public policy that aims to
promote adult learning, target various barriers to
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 55
participation and ensure equal opportunity to edu-
cation for all.
Based on the above mentioned considerations and
with the goal to enhance the knowledge on the
relation between business expenditure on R&D and
labour productivity as well as business expenditure
on education and training and labour productivity, a
system dynamics method of modelling is applied in
the next section.
2 Model
System dynamics modelling is a method that fa-
cilitates the study and analysis of dynamic feedback
systems. By using the system dynamics method of
modelling, it is possible to look at the whole system
and examine the interaction between various fac-
tors, rather than examining each factor as isolated.
System dynamics modelling offers a more compre-
hensive view of the problem and helps explaining
which factors and policies lead towards improved
system performance.
System dynamics modelling is a highly abstract
method of modelling which ignores the fine details
of a system and produces a general representation
of a complex nonlinear dynamical system (Sterman,
2002; The AnyLogic Company, 2019). Originally
developed at the Massachusetts Institute of Tech-
nology in the 1950s by Jay Forrester (1958), it has
been applied in production management, educa-
tion, urban planning, public policy, agriculture, en-
ergy policy and other areas. System dynamics
models are designed to improve decision making
and are mainly used for “what-if” scenarios, policy
testing and optimisation. Although the method is
challenging and time-consuming, one of its main
advantages is its suitability to work as a learning
laboratory, since it offers the possibility to gain
experience with various systems that is often
infeasible or costly in the real world (Groff, 2013;
Uriona Maldonado & Grobbelaar, 2017).
The central part of system dynamics is system
thinking (Sterman, 2002). In system dynamics, the
real-world processes are represented in terms of
stocks (e.g. people, knowledge, money), flows be-
tween these stocks (processes that directly add or
subtract from a stock), and information that deter-
mine the values of the stocks and flows. Stocks can
only change through their flows. System dynamics
postulates that dynamic processes in systems run in
feedback loops and that the history of systems ac-
cumulates in defined variables. The accumulated
history influences the future development of a sys-
tem (Fang et al., 2018).
The process of modelling in system dynamics
consists of several phases. It starts with the problem
formulation, i.e. identification of a specific problem
to be addressed, definition of appropriate time ho-
rizon and selection of important variables. In the
second phase, it is necessary to identify the stocks,
flows and feedback structures that can explain the
problematic behaviour. The third phase comprises
formulation of a simulation model with stock and
flow structure and decision rules. In the fourth
phase, it is necessary to test the model for validation
and behaviour. The final phase refers to policy
formulation and evaluation. It comprises the eval-
uation of new decisions and strategies that could be
implemented in the real world through “what-if”
scenarios (Duggan, 2016; Sterman, 2002).
Based on the considerations previously
mentioned in the introduction and theoretical
background, the following model has been devel-
oped using the free online software for modelling
and simulation Insight Maker (2019):
The presented model shows the relationship of
diverse variables, stocks and flows that determine the
Fig. 1. Labour productivity simulation model. Source: author's model.
56 ECONOMIC AND BUSINESS REVIEW 2022;24:52e63
value of labour productivity. The latter is expressed as
labour productivity per hour worked, as it provides a
better picture of productivity developments in the
economy than labour productivity per person
employed (it eliminates differences in the full time/
part time composition of the workforce).
In order to determine the value of labour produc-
tivity per hour worked, some variables have been
taken into consideration based on the literature
overview and data availability, such as business
expenditure on research and development (BERD),
business expenditure on education, adult participa-
tion in learning and total number of workers.Worker,
skilledworker and value added (company's profit) are
stocks, while newworkers, worker leaving, education,
skilled worker leaving, income and cost of education
andR&Dareflows in thismodel. The cloud represents
a stock outside the model boundary. It serves as the
acknowledgement that the observed system is con-
nected in some way to another system.
3 Results and discussion
The goal of the model presented in Fig. 1 is to
enable company owners and decision managers to
see the impact that various policies and changes
may have on labour productivity in the company.
The above model has been tested on the German
economy, using secondary data from the Eurostat
database for the year 2016. The German economy
has been chosen, as it is one of the most developed
economies in Europe and is often used as a role
model for many other European countries. In order
to simplify the visualisation and calculations, some
basic assumptions for the model are defined: the
hypothesized company has 100 unskilled workers,
each month the company employs two new un-
skilled workers, one newly employed unskilled
worker leaves the company each month (e.g. he or
she did not fit in the company, did not like the job,
etc.) and every two months one skilled worker
leaves the company (e.g. for a better paid or more
challenging job position in another company, mov-
ing to another place for family matters, etc.). Ac-
cording to Eurostat data (Eurostat, 2019), in 2016
business expenditure on R&D was at 1.99% of the
German GDP, while business expenditure on edu-
cation and training was at 7% of the German GDP
and adult participation rate in education and
training was 56.4% (employed persons). The simu-
lation is set for a period of ten years.
Fig. 2 shows a simulation of labour productivity
for the hypothesized company based on the above
mentioned data and model assumptions.
The model shows that labour productivity in-
creases by over 60% after ten years. It firstly grows
rapidly and then slows down as the time passes.
Fig. 3 reveals that it starts to slow down after 43
months, i.e. when the number of unskilled and
skilled workers meet in the company. After 43
months, there are approximately 122 employed
workers left in the company, of which 61 are un-
skilled and 61 skilled workers.
System dynamics is a useful tool for long-term,
strategic modelling and simulation. Figs. 2 and 3
show a simulation of labour productivity, the num-
ber of unskilled and skilled workers as well as the
number of total workers for a period of 120 months
Fig. 2. Simulation of labour productivity for hypothesized company for 120 months. Source: author's simulation.
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 57
in a hypothesized German company. Company
owners and decision makers are often interested in
improving labour productivity. Changes in pro-
cesses can be expensive and it is therefore better for
decision makers to simulate the effects of new pol-
icies before applying them in reality in order to
avoid unintended consequences.
The presented model can easily be used by com-
pany managers to analyse various investment sce-
narios in similar real-world companies and serve as
a decision support tool for long-term strategic
planning. Many companies find themselves in a
situation where they do not know whether it is more
convenient to increase the investment in R&D or
education and training activities of their workers in
order to improve labour productivity and company
profit. The model presented in this paper could help
companies to find the right answer through building
“what-if” scenarios. Assuming that the company
decides to increase its labour productivity by
increasing the expenditure on R&D, e.g. decides to
invest two times more (4% instead of 1.99% of the
German GDP), labour productivity would not in-
crease significantly.
Fig. 4 shows that increasing the investment in R&D
did not contribute to the significant growth in labour
productivity (62.56% vs. 61.96% after ten years). Ac-
cording to this simulation model, the investment in
R&D does not have significant influence on labour
productivity. Such a result differs from themajority of
researches done in the past and presented in the
literature overview, but is in line with the results ob-
tained by Benavente (2006) and Erdil et al. (2013).
Adoption of new technologies and introduction of
new production processes require time and effort that
usually lead to the insignificant labour productivity
growth, especially in the short run (Ahn, 1999). In-
vestment in R&D does not directly increase workers'
knowledge and skills. Highly educated and skilled
workers usually already work in R&D, but other
workers in the company need time to acquire new
knowledge and improve their performance. It is
important to invest in workers’ education to teach
them how to use new developments (e.g. develop-
ment of new software does not increase labour pro-
ductivity, if workers do not learn how to use it).
In the new era of rapid information exchange,
learning by doing has become obsolete, as it is time-
consuming. Therefore, companies should improve
workers’ knowledge and skills by investing in edu-
cation and training activities. Fig. 5 shows a simulation
model of labour productivity when the company de-
cides to invest in education and training of its workers
(in this case, it decides to invest twice asmuch, i.e. 14%
instead of 7% of the German GDP).
The increase of investment in education and
training leads to the growth in labour productivity
(67% after ten years). This simulation model proves
that there is a positive impact of business investment
in education and training on labour productivity and
such results are in linewith the researches done in the
past. However, by using this simulation model the
company can design better operating policies and
guide effective changes, e.g. it can plan investments in
education and training activities. At the beginning of
the investment labour productivity growth is faster, as
Fig. 3. Simulation of unskilled workers, skilled workers and total number of workers in the hypothesized company for 120 months. Source: author's
simulation.
58 ECONOMIC AND BUSINESS REVIEW 2022;24:52e63
workers acquire new knowledge and skills through
education and training activities. Therefore, the com-
pany can increase its investment in education and
training in the first years and decrease it when there
will bemore skilledworkers thanunskilledworkers in
the company. Fig. 6 reveals that with doubling the
investment in education and training, the company
will need less time to reach the moment when it has
the same number of unskilled and skilled workers (32
months instead of 43).
Companies should not only increase their ex-
penditures on education and training, but also
motivate their workers to participate in such activ-
ities. Without participation in education and
training activities, workers will need more time to
acquire new knowledge and skills. In the case of the
hypothesized company presented in this paper, if
the adult participation rate in education and
training activities was halved (28.2% instead of
56.4% employed persons), it would take the com-
pany more than 10 years to equalize the number of
unskilled and skilled workers.
System dynamics modelling is not an intervention,
but rather a useful tool for analysing the structure and
Fig. 4. Simulation of labour productivity for the hypothesized company for 120 months when it doubles its expenditure on R&D (4% instead of 1.99%
of the German GDP). Source: author's simulation.
Fig. 5. Simulation of labour productivity for the hypothesized company for 120 months when it doubles its expenditure on education and training
(14% instead of 7% of the German GDP). Source: author's simulation.
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 59
behaviour of a complex system with nonlinear links
and for designing efficient policies. The model pre-
sented in this paper highlights the importance of
learning and investing in education and training ac-
tivities in order to improve labour productivity.
Learning by doing is time-consuming and expensive
nowadays, especially in high tech and other industries
that require rapid technological changes (Ahn, 1999).
Although learning by doing is applied in many com-
panies, active research and learning is a dominant
driver of technological change (Jamasb, 2006). Adop-
tion of new technologies requires time and effort and
therefore tends to restrict labour productivity growth
temporarily. Innovation and R&D can be seen as a
prolonged learning process on experience and prob-
lem solving, as learning advances are embodied in
new technological developments that could improve
labour productivity in the long run (Baffoe-Bonnie,
2016). Learning interacts with technology to improve
labour productivity, reduces company's costs of pro-
duction and increases profit. Thus, companies should
increase the investment in professional training and
encourage their workers to actively participate in ed-
ucation and training activities in order to keep pace
with new developments and to apply new knowledge
and skills in their everyday work tasks. By increasing
the levels of engagement and creating a favourable
working environment, companies will prevent high
turnover rates that harm labour productivity and
affect their development plans. Companies are less
likely tofire trainedworkers, as educationand training
costs can be high. In addition, company-specific
knowledge acquired through worker training cannot
be applicable in other companies and could positively
influenceworkerwage and their decision to stay in the
same company (Konings & Vanormelingen, 2015).
4 Conclusion
Labour productivity is a fundamental factor in
determining how fast the economy and the average
standard of living grow. Existing studies on labour
productivity have largely focused on the production
function estimation. This paper advances those
studies by presenting a system dynamic model. The
use of system dynamics represents an untraditional
approach in modelling labour productivity. The main
advantage of the presented model is that it enables
viewing problems and actions as interconnected
wholes and takes into account lagged feedback loops.
As such, it can easily be applied by company owners
and decision makers in long-term strategic planning
and labour productivity improvement.
Thepurposeof thispaper is toexamine the influence
of investments in R&D and education on labour pro-
ductivity using system dynamics modelling. There-
fore, thispaper represents a contribution to the studies
dealing with these relationships and to the system
dynamics application. The existing literature mostly
shows the importance of innovation and R&D in
raising labour productivity and emphasizes that
companies should adopt innovations as quickly as
possible. However, this study reveals that business
expenditure on education and training aswell as adult
participation in such activities have stronger influence
on the labour productivity growth. The findings pro-
vide further evidence on the importance of absorbing
Fig. 6. Simulation of unskilled workers, skilled workers and total number of workers in the hypothesized company for 120 months when it invests two
times more in education and training (14% instead of 7% of the German GDP). Source: author's simulation.
60 ECONOMIC AND BUSINESS REVIEW 2022;24:52e63
new knowledge generated by R&D. The impact of
innovations on the labour productivity growth de-
pends largely on their diffusion and adoption which
can mostly be done by education and training
activities.
In the era of Industry 4.0, the flexibility of workers
and their ability to acquire new knowledge have
become the most desirable characteristics. Large
and rapid flow of every day information requires
constant updating in order to keep pace with the
competition. New technologies have become more
complex and require specific training for their users.
Hence, learning by doing is not sufficient anymore,
as it is time consuming.
According to this study, adult participation in edu-
cation and training activities has become one of the
maindrivers that boost labourproductivity. Education
should be universally available, as the higher the
average level of education in an economy, the higher
the accumulated human capital and the higher the
labour productivity. It is important that companies
and policymakers encourage and support adult
participation in education and training activities,
because low productivity growth is correlated not just
with low wages and low competitiveness but also low
government revenues. Reduced government reve-
nues, among other things, have as a consequence
fewer funds for productivity enhancing expenditures
in the areas such as innovation, R&D, education, and
therefore decrease labour productivity and living
standard over time. Policy interventions can include
supporting educational institutions, creating and
supporting training programs, supporting specialised
online portals for teaching, etc.
The presented model can easily be adjusted to
each company and extended by adding variables,
stocks and flows to analyse additional problems and
actions. However, there is no model that will
perfectly represent the reality and some limitations
should be taken into consideration. Labour turnover
is not always easy to predict and it affects produc-
tivity. Workers that leave or join the company will
differently affect productivity, depending on their
level of education and performance.
The simulation model developed in this study is
limited to just one country and uses the average
data for one year. Further research on a larger
sample using bigger datasets should be done in the
future to confirm the obtained results. The data
regarding business expenditures on R&D and edu-
cation and training are aggregated data for com-
panies of all sizes. Thus, it would be interesting to
see in further studies, whether the same results
would be achieved in case that the developed model
used data only for micro or large companies.
References
Afrooz, A., Rahim, K. B. A., Noor, Z. B. M., & Chin, L. (2010).
Human capital and labor productivity in food industries of
Iran. International Journal of Economics and Finance, 2(4), 47e51.
https://doi.org/10.5539/ijef.v2n4p47
Ahn, S. (1999). Technology upgrading with learning cost: A solution for
two productivity puzzles. OECD Economics Department Work-
ing Paper No. 220. https://ssrn.com/abstract=193173
Almeida, R., & Carneiro, P. (2009). The return to firm investments
in human capital. Labour Economics, 16(1), 97e106. https://
doi.org/10.1016/j.labeco.2008.06.002
Arrow, K. J. (1962). The economic implications of learning by
doing. The Review of Economic Studies, 29(3), 155e173. https://
doi.org/10.2307/2295952
Atkinson, R. D. (2018). How ICT can restore lagging European pro-
ductivity growth. Washington DC: Information Technology &
Innovation Foundation. https://ssrn.com/abstract¼3324656
Baffoe-Bonnie, J. (2016). Productivity growth and input demand:
The effect of learning by doing in a gold mining firm in a
developing economy. International Economic Journal, 30(4),
550e570. https://doi.org/10.1080/10168737.2016.1204341
Bahk, B.-H., & Gort, M. (1993). Decomposing learning by doing in
new plants. Journal of Political Economy, 101(4), 561e583.
https://doi.org/10.1086/261888
Bartel, A. P. (1995). Training, wage growth and job performance:
Evidence from a company database. Journal of Labor Economics,
13(3), 401e425. https://doi.org/10.1086/298380
Baumann, J., & Kritikos, A. S. (2016). The link between R&D,
innovation and productivity: Are micro firms different?
Research Policy, 45(6), 1263e1274. https://doi.org/10.1016/
j.respol.2016.03.008
Baumgartner, J. P. (2009). The corporate innovation machine. https://
www.creativejeffrey.com/creative/innovationMachine.pdf
Benavente, J. M. (2006). The role of research and innovation in
promoting productivity in Chile. Economics Of Innovation And
New Technology, 15(4e5), 301e315. https://doi.org/10.1080/
10438590500512794
Benos, N., & Karagiannis, S. (2016). Do education quality and
spillovers matter? Evidence on human capital and productiv-
ity in Greece. Economic Modelling, 54, 563e573. https://doi.org/
10.1016/j.econmod.2016.01.015
Chiang, H. (2004). Learning by doing, worker turnover and produc-
tivity dynamics. Econometric Society 2004 Far Eastern Meetings
593. https://ideas.repec.org/p/ecm/feam04/593.html
Coccia, M. (2011). The interaction between public and private
R&D expenditure and national productivity. Prometheus, 29(2),
121e130. https://doi.org/10.1080/08109028.2011.601079
Crepon, B., Duguet, E., & Mairesse, J. (1998). Research, innovation
and productivity: An econometric analysis at the firm level.
Economics of Innovation and New Technology, 7(2), 115e158.
https://doi.org/10.1080/10438599800000031
Crespi, G., & Zuniga, P. (2012). Innovation and productivity: Ev-
idence from six Latin American countries. World Development,
40(2), 273e290. https://doi.org/10.1016/j.worlddev.2011.07.010
Da Rocha, L. P., Pero, V. L., & Corseuil, C. H. (2019). Turnover,
learning by doing, and the dynamics of productivity in Brazil.
EconomiA, 20(3), 191e210. https://doi.org/10.1016/j.econ.
2019.11.001
Desjardins, R. (2015). Participation in adult education opportu-
nities: Evidence from PIAAC and policy trends in selected
countries. Background paper for the Education for All Global
Monitoring Report. https://unesdoc.unesco.org/ark:/48223/pf00
00232396
Duggan, J. (2016). An introduction to system dynamics. In System
dynamics modeling with R, lecture notes in social networks (pp.
1e24). Springer International Publishing. https://doi.org/
10.1007/978-3-319-34043-2_1
Duran, J., & Rillaers, A. (2002). Physical and human capital invest-
ment: Relative substitutes in the endogenous growth process. Insti-
tuto Valenciano de Investigaciones Economicas, S.A.,
Working papers, serie AD 2002-18.
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 61
Engelbrecht, H.-J. (1997). International R&D spillovers, human
capital and productivity in OECD economies: An empirical
investigation. European Economic Review, 41(8), 1479e1488.
https://doi.org/10.1016/S0014-2921(96)00046-3
Erdil, E., Cilasun, S. M., & Eruygur, A. (2013). Do R&D expendi-
tures matter for labor productivity in OECD countries? An
unresolved question. Hacettepe University Journal of Economics
and Administrative Sciences, 31(1), 71e82. https://doi.org/
10.17065/huniibf.103656
Eurostat. (2019). Eurostat database. https://ec.europa.eu/eurostat/
data/database
Fang, Y., Lim, K. H., Qian, Y., & Feng, B. (2018). System
dynamics modeling for information systems research: Theory
development and practical application. MIS Quarterly, 42(4),
1303e1329.
Forrester, J. W. (1958). Industrial dynamics: A major
breakthrough for decision makers. Harvard Business Review,
36(4), 37e66.
Galor, O., & Moav, O. (2004). From physical to human capital
accumulation: Inequality and the process of development. The
Review of Economic Studies, 71(4), 1001e1026. https://doi.org/
10.1111/0034-6527.00312
Griffith, R., Huergo, E., Mairesse, J., & Peters, B. (2006). Innova-
tion and productivity across four European countries. Oxford
Review of Economic Policy, 22(4), 483e498.
Groff, J. S. (2013). Dynamic systemsmodeling in educational system
design& policy. Journal ofNewApproaches in EducationalResearch,
2(2), 72e81. https://doi.org/10.7821/naer.2.2.72-81
Hashi, I., & Stoj
cic, N. (2013). The impact of innovation activities
on firm performance using a multi-stage model: Evidence
from the Community Innovation Survey 4. Research Policy,
42(2), 353e366. https://doi.org/10.1016/j.respol.2012.09.011
Hendel, I., & Spiegel, Y. (2013). Small steps for workers, a giant leap
for productivity. CSIO Working Paper, No. 0117.
Hirsch, W. Z. (1952). Manufacturing progress functions. The Re-
view of Economics and Statistics, 34(2), 143e155. https://doi.org/
10.2307/1928465
International Labour Organization. (2019). Statistics on labour
productivity. https://ilostat.ilo.org/topics/labour-productivity/
Jamasb, T. (2006). Technical change theory and learning curves:
Patterns of progress in energy technologies. Energy Journal, 28,
1e24. https://doi.org/10.5547/ISSN0195-6574-EJ-Vol28-No3-4
Kemp, R. G.M., Folkeringa, M., de Jong, J. P. J., &Wubben, E. F. M.
(2003). Innovation and firm performance. (Scales; No. H 200207).
EIM.
Klomp, L., & Van Leeuwen, G. (2001). Linking innovation and
firm performance: A new approach. International Journal of the
Economics of Business, 8(3), 343e364. https://doi.org/10.1080/
13571510110079612
Konings, J., & Vanormelingen, S. (2015). The impact of training on
productivity and wages: Firm-level evidence. The Review of
Economics and Statistics, 97(2), 485e497. https://doi.org/10.1162/
REST_a_00460
Lane, D. C., & Sterman, J. D. (2011). Jay Wright Forrester. In
A. A. Assad, & S. I. Gass (Eds.), Profiles in operations research:
Pioneers and innovators (pp. 363e386). Springer. https://doi.org/
10.1007/978-1-4419-6281-2_20
L€o€of, H., & Heshmati, A. (2006). On the relationship between
innovation and performance: A sensitivity analysis. Economics
of Innovation and New Technology, 15(4/5), 317e344. https://
doi.org/10.1080/10438590500512810
Maciulyte-Sniukiene, A., & Matuzeviciute, K. (2018). Impact of
human capital development on productivity growth in EU
member states. Business, Management and Education, 16(1),
1e12. https://doi.org/10.3846/bme.2018.66
Mairesse, J., & Robin, S. (2009). Innovation and productivity: A firm-
level analysis for French manufacturing and services using CIS3 and
CIS4 data (1998e2000 and 2002e2004). CREST-ENSAE.
Insight Maker. (2019). Start building insights and models for free.
https://insightmaker.com
Mohnen, P., & Hall, B. H. (2013). Innovation and productivity: An
update. Eurasian Business Review, 3(1), 47e65.
Montgomery, F. J. (1943). Increased productivity in the
construction of liberty vessels. Monthly Labor Review, 57(5),
861e864.
Pietrzak, M. B., & Balcerzak, A. P. (2016). Quality of human capital
and total factor productivity in new EU member states. In
T. Loster, & T. Pavelka (Eds.), Conference proceedings of the 10th
international days of Statistics and economics (pp. 1492e1501).
Libuse Macakova, Melandrium https://msed.vse.cz/msed_
2016/sbornik/toc.html
Rapping, L. (1965). Learning and world war II production func-
tions. The Review of Economics and Statistics, 47(1), 81e86.
https://doi.org/10.2307/1924126
Reese, H. W. (2011). The learning-by-doing principle. Behavioral
Development Bulletin, 17(1), 1e19. https://doi.org/10.1037/
h0100597
Sheshinski, E. (1967). Tests of the “learning by doing hypothesis”.
The Review of Economics and Statistics, 49(4), 568e578. https://
doi.org/10.2307/1928342
Solomon, E., Ugur, M., Guidi, F., & Trushin, E. (2015). Variations
in the effect of R&D investment on firm productivity: UK ev-
idence. Greenwich Papers in Political Economy, 14081, 1e25.
Sprague, S. (2014). What can labor productivity tell us about the
U. S. economy? Beyond the Numbers: Productivity, 3(12), 1e9.
https://www.bls.gov/opub/btn/volume-3/what-can-labor-
productivity-tell-us-about-the-us-economy.htm
Stauvermann, P. J., & Kumar, R. R. (2018). Adult learning, eco-
nomic growth and the distribution of income. Economies, 6(1).
https://doi.org/10.3390/economies6010011
Sterman, J. D. (2002). Business dynamics: System thinking and
modeling for a complex world. Working paper series ESD-WP-
2003-01.03. Massachusetts Institute of Technology. https://
dspace.mit.edu/handle/1721.1/102741
Sterman, J. D., Fiddaman, T., Franck, T., Jones, A., McCauley, S.,
Rice, P., Sawin, E., & Siegel, L. (2013). Management flight
simulators to support climate negotiations. Environmental
Modelling & Software, 44, 122e135. https://doi.org/10.1016/
j.envsoft.2012.06.004
Stewart, T. A. (2010). Intellectual Capital: The new wealth of orga-
nization. Crown Publishing Group.
Stoj
cic, N., & Hashi, I. (2014). Firm productivity and type of
innovation: Evidence from the community innovation survey
6. Croation Economic Survey, 16(2), 121e146. https://doi.org/
10.15179/ces.16.2.5
Tabari, N. A. Y., & Reza, M. (2012). Technology and education
effects on labour productivity in the agricultural sector in Iran.
European Journal of Experimental Biology, 2(4), 1265e1272.
The AnyLogic Company. (2019).What is system dynamics modeling?
https://www.anylogic.com/use-of-simulation/system-
dynamics/
Tsai, K. H., & Wang, J. C. (2004). R&D productivity and the
spillover effects of high-tech industry on the traditional
manufacturing sector: The case of Taiwan. The World Economy,
27(10), 1555e1570. https://doi.org/10.1111/j.1467-9701.2004.
00666.x
Uriona Maldonado, M., & Grobbelaar, S. S. (2017). System dy-
namics modelling in the innovation systems literature [Paper pre-
sentation]. 15th Globelics International Conference, Athens,
Greece.
Wakelin, K. (2011). Productivity growth and R&D expenditure in
UK manufacturing firms. Research Policy, 30(7), 1079e1090.
https://doi.org/10.1016/S0048-7333(00)00136-0
Wright, T. P. (1936). Factors affecting the cost of airplanes. Journal
of the Aeronautical Sciences, 3, 122e128. https://doi.org/10.2514/
8.155
Yunus, N. M., Said, R., & Hook, L. S. (2014). Do cost of training,
education level and R&D investment matter towards influ-
encing labour productivity? Jurnal Ekonomi Malaysia, 48(1),
133e142.
62 ECONOMIC AND BUSINESS REVIEW 2022;24:52e63
Appendix: Model equation list
Simulation settings  Time Start: 0
 Time Length: 120
 Time Step: 1
 Time Units: Months
Model variables
Adult participation in
learning
 Value: 0.564
BERD  Value: 0.0199
Business exp. on
education
 Value: 0.07
Labour productivity  Value: ([Worker]þ1.84*[Skilled
worker])/([Worker]þ[Skilled
worker])þ0.3*[BERD]
Total workers  Value: [Skilled worker]þ[Worker]
Model stocks
Skilled worker  Initial Value: 0
Value added  Initial Value: 100
Worker  Initial Value: 100
Model flows
Cost of education and
R&D
 Rate: [Value added]*([Business exp.
on education]þ[BERD])
 Alpha: Value added
 Omega: None
 Positive Only: Yes
Education  Rate: 0.4/12*[Adult participation in
learning]*[Worker]þ10*[Business
exp. on education]*([Adult
participation in learning]*
[Worker]/18þ(1-[Adult
participation in learning]*
[Worker])/24)
 Alpha: Worker
 Omega: Skilled worker
 Positive Only: Yes
Income  Rate: [Labour productivity]*
100-[Value added]
 Alpha: None
 Omega: Value added
 Positive Only: Yes
New workers  Rate: 2
 Alpha: None
 Omega: Worker
 Positive Only: Yes
Skilled worker leaving  Rate: 0.5
 Alpha: Skilled worker
 Omega: None
 Positive Only: Yes
Worker leaving  Rate: 1
 Alpha: Worker
 Omega: None
 Positive Only: Yes
ECONOMIC AND BUSINESS REVIEW 2022;24:52e63 63