Innovation and Firms’ Performance in the Rwandese Manufacturing Industry.

A firm Level Empirical Analysis

By:
Etienne Ndemezo1, Charles Kayitana2, Angelique Dukunde3 and Jean Bosco Ndikubwimana4

Abstract

The main objectives of this paper is (i) to determine factors that support engagement in innovative
activities by Rwandese manufacturing firms; (ii) to assess effects of innovation decision on
innovation performance of Rwandese manufacturing firms; and (iii) to evaluate impacts of
innovation performance on financial performance of Rwandese manufacturing firms. To address
these two objectives, we used a structural multistage framework as proposed by Crepon et al.
(1998). Data used are from the 2006 enterprise survey conducted by the World Bank in Rwanda
and referenced as ‘ID RWA_2006_ES_v01_M_WB’. This study resulted in three main outcomes:
(i) product innovation is linked directly to the process innovation, meaning that firms which engage
in process innovation introduced new or improved products on the market; (ii) innovation output,
here the ‘international quality-recognition’, is not linked to the firm engagement in innovation. It
is linked to the use of technology licensed from foreign firms; (iii) the ‘international quality-
recognition’ is the main determinant of firm’s financial performance. Consequently, in order to
boost manufacturing firms’ financial performance a public assistance in Research and
Development to Rwandese manufacturing is recommendable.

Keywords: Innovation, Manufacturing industry, International quality-recognition, Firm’s

performance

JEL Classification Codes: L60 – O14 – O31 – O32

1
Etienne Ndemezo is lecturer in the University of Rwanda, College of Business and Economics, School of
Economics. His is the corresponding author of this paper. e-mail: etienne.ndemezo@gmail.com

2
Charles Kayitana is lecturer in the University of Rwanda, College of Business and Economics, School of Business.
3
Angelique Dukunde is Lecturer in the University of Kigali, Faculty of Business Management and Economics.
4
Jean Bosco Ndikubwimana is Lecturer in the University of Rwanda, College of Business and Economics, School of
Economics, Applied Statistics Department.

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
1. Introduction growth and competitiveness among business,
specific regions and in the economy at large.
For a rapid economic growth in developing
economies, entrepreneurship and industrial Based on Schumpeter’s (1940) theory of
development are among valuable targets. creative destruction that proposes that non-
While entrepreneurship is the mainstay of innovative firms and products are replaced
significant economic growth, it can only with innovative ones; Rwanda, similar to
expect tangible results through innovation. many regional and African countries has
Innovation refers to all scientific, identified innovation to be a sustainable way
technological, organizational, financial and to ensure high economic growth and
commercial steps which actually lead to, or enterprise performance (MINECOFIN,
are intended to lead to, the implementation of 2013). Although innovation is generally
technologically new or improved products or regarded as a means of improving the
services (OECD, 2005). The OECD (2005) competitiveness of firms and their
inventories four types of innovation: product performance locally or regionally, it hasn’t
innovations, process innovations, been supported unambiguously by empirical
organizational innovations and marketing work in Rwanda.
innovations.
Different empirical evidence shows the
A number of factors can affect innovation dynamic role of entrepreneurial activity in
activities. These include economic factors promoting innovation and technology,
such as production costs and demand, factors economic growth and employment
specific to an enterprise such as skilled (Audretsch et al., 2006; Van Stel, 2006;
personnel or knowledge, and legal factors Fritsch and Mueller, 2004; 2008). On the
such as competition regulations and tax rules other hand entrepreneurship development has
(Crepon et al., 1998). Marques at al. (2011) been based on innovative ideas and use of
stressed the fact that encouraging firms to new technologies to support enterprise
innovate will lead to a better economic performance (Balachandran and Sakthivelan,
performance of firms in terms of market and 2013; Tuan, Nhan, Giang and Ngoc, 2016).
financial performance. Thus, policy that According to Tuan, Nhan, Giang and Ngoc
promotes innovation may help fostering (2016), enterprise performance can be
identified as a multidimensional concept that

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
can be measured by three indicators: to deal with both problems of selection bias
production, finance and marketing. and simultaneity. Data used are from the
“Rwanda 2006 Enterprise Survey” of the
Having accepted the importance of
World Bank and referenced as ‘ID
innovation and technology, it is
RWA_2006_ES_v01_M_WB’.
disappointing to note that in Rwandan
context there is no prior specific research This study resulted in three main outcomes:
done to measure the impact of innovation and (i) product innovation is linked to the process
technology on enterprise performance and innovation, meaning that firms which have
even the most direct and quantifiable engaged in process innovation have also
outcomes of innovation and technology on introduced new or improved products on the
entrepreneurship. market; (ii) innovation output, here the
‘international quality-recognition’, is not
Thus, with the merit of fulfill this gap, this
linked to the firm engagement in innovation.
paper has as main objective of investigating
It is linked to the use of technology licensed
the impact of innovation on entrepreneurship
from foreign firms; (iii) the ‘international
development in Rwanda. Therefore, two
quality-recognition’ is the main determinant
specific objectives are assigned to this study:
of firm’s financial performance.
(i) determine factors that support engagement
in innovative activities by Rwandese The paper is organized as follows: after this
manufacturing firms; (ii) assess the effects of introduction, we present a brief literature
innovation decision on innovation review. It is followed by a methodological
performance of Rwandese manufacturing section and empirical results. The paper ends
firms; and (iii) evaluate impacts of by a conclusive section which summarizes
innovation performance on financial the main findings and gives
performance of Rwandese manufacturing recommendations.
firms.
2. Literature review
To address these three specific objectives, we
As stressed by Fagerberg (2004), innovation
had recourse to the structural multistage
is not a new phenomenon, but in spite of its
approach as suggested by Crepon et al.
importance it has not received enough
(1998). The Crepon (1998) modeling permits
attention of scholars. However, research on

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
innovation and economic and social change some inefficient firms. Thus, he suggested to
has proliferated in recent years, particularly direct attention to the extent to which entry
in social sciences. Especially, researches on and exit regulation impacts the
the relationship between innovation and rationalization of industry structure in
productivity or performance of firms has response to innovative activity.
been synthesized by Mairesse and Mohnen
As Hall (2011), Mohnen and Hall (2013)
(2010), Hall (2011) and Mohnen and Hall
analyzed the effects of technological and
(2013) particularly. Mairesse and Mohnen
non-technological innovations on the
(2010) analyzed innovation surveys’
productivity of firms by reviewing the
characteristics and econometric problems
existing evidence from literature. They
raised by such data collected. While Hall
updated the survey by Hall (2011) and
(2011) study concerned the synthesis of
complemented the Mairesse and Mohnen
researches about the relationship between
(2010) survey on the use of innovation
innovation and productivity at the firm level,
surveys to better understand innovation.
Mohnen and Hall (2013) updated the
From this survey of empirical literature, they
literature review in both previous studies.
concluded that innovation leads to a better
With the main target of determining the productivity performance. Also, they
relationship between innovation and observed that all types of innovation
productivity of European firms, Hall (2011) influence the productivity, but isolating
reviewed the ways in which economists have individual effect remains difficult because of
analyzed the relationship between simultaneity of different types of innovation.
productivity and innovation. He concluded Further, they observed that the effect of
that there are substantial positive impacts of innovation is divided into two parts; one
product innovation on revenue productivity, going to the real output, and another
but that the impact of process innovation is pertaining to the price at which the output is
more ambiguous. Also, he observed that at sold. However, they concluded that it is very
the individual firm level, process innovation difficult to dissociate them because of
can increase real output while leaving measurement issues.
revenue mostly unchanged. Further, one of
Individual studies give further insights about
consequences of innovation is likely to be the
the relationship between innovation and
entry of new innovating firms and the exit of
4

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
performance and raise detailed econometric impact of process innovation was either not
problems according to specificities of each significant or close to zero. Also, Legros and
other. Also, they give various understandings Galia (2012), analyzing the sources of
about the probability of firms to engage in knowledge and their effects on productivity
innovative activities. in French manufacturing, found that the
market share and firm size have a positive
Crepon et al. (1998), using an econometric
impact on innovation decision and intensity
method which corrects for selectivity5 found
of R&D. Also, they concluded that size and
that in France, some factors affect positively
worker’s involvement matters to ISO 9000
the probability for a firm to engage in
certification. However, this main result is
innovation activities. They are number of
amplified by existence of competing products
employees, sales share and distribution,
and patents. So that, they suggest that firms
market demand and the technology.
must invest not only in R&D, but also in
However, they observed that a small
different sources of internal and external
proportion of firms engage in research
knowledge such as workers’ training and ISO
activities and/or apply for patents. About the
9000 certification
effects of innovation on performance, taking
into account both simultaneity and selectivity Previous results confirm conclusions of
bias, they concluded that innovation output Griffith et al. (2006) who studied the role that
rises with innovation effort (investment in innovation plays in productivity of firms in
R&D) and firms productivity correlates with four European countries using a structural
innovation output, represented by patents multistage model (Crepon et al., 1998). These
number or innovative sales. countries are France, Germany, Spain and
UK. They used data from the third wave of
Considering different types of innovation,
the internationally harmonized community
Mairesse and Robin (2009) found that
innovation surveys. They found that firms
product innovation appears to be the main
that operate mainly in international markets
driver of labour productivity in the French
and larger firms are more engaged in formal
manufacturing and service industries. The
innovative activities (here R&D). Also, they

5
This method is known over the acronym of CDM,
initials of authors’ names: Crepon, Duguet and
Mairesse.

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
found that the process innovation is more concluded on the unwillingness of the private
positively influenced by suppliers’ sector to invest in R&D and the low
information, while the product innovation is productivity of innovation. Thus, they
more influenced by the demand information. suggest leveraging private investment in
Further, conclusions of Griffith et al. (2006) innovation through public investment.
state that the process innovation is associated
Bronwyn et al. (2009), using the structural
with productivity only in France, while
multistage model which incorporates
product innovation is associated with
information on innovation success, analyses
productivity in three countries, namely
impacts of innovation on productivity of
France, Spain and UK.
SMEs in Italy. They found that the
Also, previous findings are supported by the international competition fosters R&D
conclusion of Zemplinerova and intensity, especially within high-tech firms.
Hramadkova (2012) in Czechoslovakia. Determinants of engagement in both product
These authors found that probability to and process innovations were firm size,
engage in innovation for a firm is positively investment in R&D and in equipment. Also,
influenced by its size. Also, analyzing the they found that both product and process
relationship between innovation and innovation influenced the productivity of
productivity in the Mexican manufacturing SMEs firms; especially process innovation.
industry, Brown and Guzman (2014) However, they observed that productivity of
concluded that firms that have more larger and older firm among SMEs was less
propensity to innovate are the largest, with influenced by their innovativeness.
high technological intensity and market
In Belgium, Van Beveren and
share. Also, other outcomes of their study are
Vandenbussche (2009), using data from the
that advertising; knowledge appropriability,
Community Innovation Survey for Belgium
foreign direct investment, information
attempt to explore the relationship between
technologies and access to credit have a
firm-level innovation activities and the
positive effect on innovation efforts.
propensity to start exporting. Their study
In Greece, Beneki et al. (2012) attempted to resulted in significant positive effects of
investigate the relationship between combination of product and process
innovativeness and firm performance and innovation on probability to enter the export

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
market. However, they pointed to on the innovative performance. Further, they
endogeneity of innovation activities, and so, concluded that factors which influence
observed that firms that have good prospects innovative performance had also positive
of entering the export market in the next effects on production, market and finance
period are more likely to invest in innovation performances.
activities.
About methodological issues, we can refer to
However in developing countries, findings Lööf and Heshmati (2006). Indeed, with aim
about the impact of each type of innovation of analyzing the sensitivity of the estimated
are somehow different. Waheed (2011) when relationship between innovativeness and firm
analyzing the influence of product and performance, Lööf and Heshmati (2006)
process innovations on firms’ productivity in found that the simultaneity between
Bangladesh and Pakistan found that the innovation activities and productivity or
process innovation affects more the performance is of great importance and
productivity of firms rather than the product merits much more attention rather than the
innovation. In Mexican, Brown and Guzman selectivity bias. Also, they found enough
(2014) found that firms that have the higher homogeneity of the relationship between
propensity to innovate are those which are innovation and productivity in both
larger in terms of intensity in high technology manufacturing and services firms.
and market share. Also, they found that Firms
Studies which link innovation to firm
which innovate have a level of labor
performance or that analyse determinants of
productivity 1.3 times higher than firms that
firm innovativeness are yet few in developing
do not innovate. However, their study doesn’t
countries. Especially, in Rwanda, on the best
distinguish between process and product
of our knowledge, no other similar study has
innovations.
been conducted. Thus, this paper seems
In Vietnam, Tuan et al. (2016) found that having merit of filling this gap. Also, findings
process, organization and marketing from this study could be usefulness for policy
innovation respectively have the significantly making about efficiency of innovativeness
positive impact on innovative performances. within Rwandese manufacturing industry.
However, they observed that product
innovation activities had no statistical impact

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
3. Methodological framework equations: the first which concerns the
decision to innovate and the second which
A. The model
deals with the innovation input, for example
In order to address three objectives of this investment in Research and Development.
study, we refer to the CDM framework as Two equations are linked to their
detailed in Crepon et al. (1998). The determinants in the first two stages of the
theoretical foundation of this methodology is innovation process.
the Cobb-Douglass production function:
In the third stage, the innovation output (for
Q  AL K  (1) example the number of patents) of firm is
related to its innovation intensity (spending
Where Q is the total profit before taxes, A is
in R & D). The last stage concerns the
the level of innovation, L and K labor and
relationship between firm’s performance,
capital inputs. Parameters α and β are
innovation input and innovation output.
elasticities of production with respect to labor
and capital inputs. We summarize the four equations used in the
CDM modeling.
However, the CDM models raise two
econometric problems: the selectivity bias gi*  x1i 1  u1i (2)
and simultaneity bias. The selectivity bias
arises from the fact that not all firms engage With git* a latent variable of innovation
in innovation and some innovations are not decision equals to 1 if the firm has undertaken
successful. The simultaneity bias is from innovation activity or zero if not. Variable
many factors which can influence both firms’ x1it represents vector of explanatory
decision to innovate, its level of expenditure variables, 1 the associated coefficient
on innovation as well as its final
vector. Subscript i designates the firm, while
performance.
u1it is the term error.
In order to deal with these two problems, the
ki*  x2i 2  u2i (3)
CDM model is constructed in three steps. The
first step accounts for the fact that the firm is Where ki* represents the amount invested in
engaged in innovation activity. Here, the
innovation for the firm i. Variable x2i
innovative activity is described by two

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
represents vector of explanatory variables, Where qi is the indicator of firm’s financial
2 the associated coefficient vector. performance, ki* and t i* are variables
Subscript i designates the firm, while u2i is representing innovation input and output
the term error of the equation (3). respectively; x4i is the vector of explanatory
variables, 4 the associated coefficient
With reference to the innovation literature,
these first two stages of the systemic vector and u4i the term error. All variables of

approach are estimated jointly by a equation (5) are in log form except dummy
generalized Tobit model with the maximum variables. With reference to the innovation
likelihood estimation method. literature, this last specification is estimated
with three-stage least squares (Lööf and
The second stage accounts for the impact of
Heshmati, 2006), where the inverse Mills
engagement and investment in innovation to
ratio is introduced in the equation (3) in order
the innovation output. Here, can be
to deal with the selection bias.
considered as innovation output the number
of new or improved products/services or the B. Model specification

number of patents. In equation (2), the dependent variable is the

engagement of the firm in product
ti*  x3i 3  u3i (4)
innovation. Here, we consider that the firm
Where t* is an innovation output; other engages in innovation if it introduced a new
variables and parameters being defined as or significatively improved product on the
above. market. Here, the process innovation is
considered as a prerequisite to reach the
The last equation accounts for the effects of
product innovation. Explanatory variables of
innovation input and output to the firm’s
equation (2) contain process innovation,
performance. In the literature, they used
domestic competition, foreign ownership and
usually the Cobb-Douglass production
added value per worker.
function augmented to innovation variables.
Because of lack of appropriate data, equation
qi  x4 i 4  k  t  u4 i
*
i
*
i (5) (3), which concerns the relationship between
the intensity of investment in innovation and
appropriate explanatory variables, is not used

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
in our modeling. Instead of equation (3), we utilization (representing the competitiveness
used equation (4) which refers to innovation of the firm), the technology intensity
output. Thus, we determined the relationship represented here by the electricity
between innovation decision and innovation expenditure, variables of market conditions
output, represented here by the fact that the (represented by direct export sales share and
firms has an international quality- national market sales share), use of
recognition. The innovation output equation information technology represented by use of
contains as explanatory variables the process e-mails and web site, company age, number
innovation as defined earlier, number of new of employees, total fixed assets (representing
competitors entering the market, use of here the physical capital) and the
website and e-mails with clients and international quality recognition (here
suppliers, use of licensed technology, representing the innovation output as
company age by 2006, total fixed assets per announced earlier).
worker, added value per worker and number
of employees in 2005. Equation (4) includes C. Data requirement
also the inverse Mill’s ratio.
Data used were collected by the World Bank
Equation (2) and equation (4) are estimated in 2006 in Rwanda and under the theme
jointly using the generalized tobit method. “Enquête sur le climat des investissements et
la productivité” and referenced as
Also, equation (4) and (5) are estimated
RWA_2006_ES_v01_M_WB. It could be
jointly. All variables of equation (5) are in log
preferable to use “the World Bank innovation
form except dummy variables. With
follow up module - enterprise survey Rwanda
reference to the innovation literature, this last
2011” which were dedicated to the specific
specification is estimated with three-stage
issue of innovation in manufacturing and
least squares (Lööf and Heshmati, 2006); but
services firms. However, this survey combine
in our modeling they are estimated using the
both manufacturing and service firms; and
two-stage least squares method because the
because of a lot of missing observations, the
equation (3) is missing.
number of manufacturing firms is not enough
to permit an appropriate analysis.
Equation (5) is the performance one and
Consequently, using the 2011-innovation
contains as explanatory variables the capacity
survey could not allow determining the true

10

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
relationship between innovation and firm’s Large
(100
performance in the manufacturing sector.
employe Small (5-
Thus, one of limitations of this study is data es and 19
more), 11 employe
which are dated and don’t give the actual es), 23
picture of innovativeness of manufacturing Medium
(20-99
firms today. employe
es), 25
4. Empirical findings

A. Overview of the sample. Figure 1: Sample size.

Source: Author’s computation from ID
RWA_2006_ES_v01_M_WB
We start this presentation by descriptive
statistics about age, size, capacity utilization
and employees of the sample used. According to the figure 1, the sample is
dominated by small and medium firms. They
A.1Employment and size of firms
are respectively 23 and 25 companies. Only
The sample used contains 59 manufacturing 11 firms are categorized as large.
companies located in Huye and Kigali cities.
With reference to number of employees,
Categorization used is borrowed from the
about 41% of firms are classified as small
World Bank and define company size as
companies using less than 19 employees.
follows:
Also, about 81% of companies studied are
– A small company is defined as using into category SMEs: they employ less than
between 5 and 19 full time employees 100 individuals.,
included.
Small, Medium
10 , 40
– A medium company has between 20
and 99 full time employees. Large,
305
– A large company is defined as
possessing 100 employees and above.

Figure 2: Average number of employees

per company by firm size.
Source: Author’s computation from ID
RWA_2006_ES_v01_M_WB

11

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
However, individually, we observed that The capacity utilization of firms reflects also
large company use in average the highest their competitiveness. Consequently,
number of full time employees, about 305 according to their size, we observe with
individuals per company. Small firm uses in figure 4 that small firms are more competitive
average only 10 individuals. in Rwanda, they use above 76% of their
potential capacity. Medium firms are less
B.1Age of companies
competitive because they use only 55.5% of
Also, referring to the company experience, their capacity. However, large firms are also
young firms are also small while elder ones relatively more competitive rather than
are large. About 53% of firm has less than medium firms.
eight of operating years, meaning that
80
majority of firms in Rwanda are too young.
The eldest firm had 66 operating years in 60

2006. In Average large firm had 22 operating 40

years, medium firm possessed 16 operating
20
years while the small firm had only 7
0
operating years.
Small Medium Large

Figure 4. Capacity utilization of firms

small, Source: Author’s computation from ID
6.86956 RWA_2006_ES_v01_M_WB
Large, 5
21.8181
8 4.1.Innovation and financial performance
Medium
, 15.48 of firms

In order to analyse effects of innovation on

firm’s financial performance, presentation of

Figure 3: Age of firms according to their results is done in two stages. First, we analyse
size. effects of innovation engagement on
Source: Author’s computation from ID
RWA_2006_ES_v01_M_WB innovation output, here the ‘international
quality-recognition’. After, we analyse the

C.1Capacity utilization of firms

12

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
effects of ‘international quality-recognition’ recognition of firms. The last is more
on financial performance of firms. influenced by the use of technology licensed
from foreign companies. In comparison with
A. Effects of innovation decision on
other variables, use of technology licensed
innovation performance.
has the highest coefficient and is statistically
According to table 1, innovation engagement very significant. In opposite, even if the
is related to the innovation output process innovation influences significatively
(international quality recognition) via the the product innovation, it has no effect on the
inverse Mills ration. This ratio is not international quality recognition.
significatively relevant as we can observe
Also, number of full time employees is one
from its p-value. This suggests that the
of factors of ‘international quality-
selection bias is not enough, results which
recognition’. This reflects the idea that firm’s
confirms the conclusion of Lööf and
size influences the international quality
Heshmati (2006). This allows us using the
recognition of manufacturing companies.
two-stage least squares technique at the last
However, as the coefficient is too low,
stage without the inverse Mills ratio.
influence of company size on firm innovation
The product innovation, representing here the output is almost negligible. On the contrary,
innovation decision by firms is positively the use of e-mail in relations with clients or
linked to process innovation and negatively suppliers impacts negatively the international
to the value-added per worker. This attests quality recognition. This result is conflicting
that a firm engages in process innovation in and very difficult to interpret. However, we
order to have a new or improved product. think that in 2005 use of e-mails was a new
Also, the financial performance of firm is not story in the Rwandese manufacturing
the main determinant of the innovation industry, and was mainly adopted by small
decision. Firms engage in innovative firms which are not enough innovative as
activities in order to improve their finance announced above.
and not the reverse.

However, innovation engagement has no

significant effect on international quality

13

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
Table 1. Effects of innovation engagement on innovation output using the generalized to bit
regression.

International quality recognized Coefficients P-Values

Process innovation 0.238057 0.667
New competitor entered market -0.059423 0.404
Use of web site with clients or suppliers 0.055952 0.614
Use of technology licensed 0.7233031 0.000
Use of e-mail with clients or suppliers -0.212101 0.019
Company experience 0.0020135 0.483
Total fixed assets per worker 1.88E-09 0.664
Added value per worker -8.82E-10 0.901
Number of employees 0.0015207 0.000
_cons -0.183072 0.762

Product innovation
Process innovation 2.376331 0.000
Domestic competitor on production cost -0.010132 0.981
Foreign ownership 0.0061147 0.349
Added value per worker -3.34E-08 0.096
_cons -0.755568 0.059
Mills lambda 0.1361329 0.766

Rho 0.68301
Sigma 0.1993147

Number of observations 59
Uncensored observations 35
Wald chi2(9) 83.31
Prob > chi2 0000

Source: Author’s computation from ID RWA_2006_ES_v01_M_WB

B. Effects of innovation of financial efficiency of manufacturing firms is the
performance ‘international quality-recognition’. Its
coefficient is the highest and statistically
According to the table 2 below, we observe
significant. This variable is the proxy of
that the main determinant of financial
innovation output and can be interpreted as

14

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
having a positive impact on financial to be financially efficient rather than firms
performance. Firms which possess the without this recognition.
‘international quality-recognition’ are likely

Table 2. Effects of innovation output on financial performance: two-stage least squares.

Log added value Coefficients P-Values
Log capacity utilization 0.5609422 0.055
Log electricity expenses 0.2894754 0.000
Direct exports share 0.0170243 0.048
National sales share 0.0027956 0.700
Use mail with clients or suppliers 0.2029652 0.449
Use web site with clients or suppliers -0.8519364 0.004
Log company experience 0.1570394 0.226
Log number of employees 0.5086777 0.000
Log total fixed assets 0.1973117 0.001
International quality-recognized 0.9073776 0.003
_cons 5.633773 0.000

Number of observations 51
F( 9, 41) 33.110
Prob > F 0.000
Total (centered) SS 189.7216
Total (uncentered) SS 17600.32
Residual SS 20.4518
Centered R2 0.88922
Uncentered R2 0.9988
Root MSE 0.6333
Sargan statistic (over identification test of all instruments) 28.446
Chi-sq(8) p-val 0.0002
Source: Author’s computation from ID RWA_2006_ES_v01_M_WB
The second factor of performance is the use because according to the table 2, the firm size
of web site in relations with clients or is also a positive determinant of financial
suppliers. However, its coefficient is negative performance of manufacturing firms. Indeed,
and, as mentioned earlier, this can attest that increase in number of employees by 10 per
the use of information technology tools is still cent leads to the rise of financial performance
at the beginning and more adopted by small by 5 per cent. Consequently, small firms by
and less innovative firms. This is likely true

15

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
the number of full time employees are innovation output. In the second and last
handicapped by their size itself. stage, we established the relationship
between innovation output and firm’s
Also, the competitiveness, the technology
performance. Data used are from the World
intensity and the physical capital are other
Bank ‘2006 - enterprise survey’.
determinants of firms’ performance. The
competitiveness is here represented by the This study resulted in three main outcomes:
capacity utilization and export sales share of (i) product innovation is linked directly to the
the firm. They are both positively linked to process innovation, meaning that firms which
the financial performance. Total electricity decided to engage in process innovation have
cost and total fixed assets are respectively introduced new or improved products on the
proxies of technology intensity and physical market; (ii) innovation output, here the
capital. These two factors are correlated ‘international quality-recognition’, is not
because they reflect that the firm masters its linked to the firm engagement in innovation.
productive technology. Consequently, It is rather linked to the use of technology
previous results show that combination of licensed from foreign firms; However, (iii)
competitiveness and technology intensity are the ‘international quality-recognition’ is the
important instruments of manufacturing main determinant of firm’s financial
firm’s performance. performance.

5. Summary and conclusion With respect to empirical findings above, we

recommended a public assistance in R&D to
This study had as specific objectives of (i)
private manufacturing firms in order to boost
determining factors of innovation decision
their innovativeness. Also, it is advisable that
and (ii) effects of innovation activity on
manufacturing firms apply for an
manufacturing firms’ performance in
‘international quality-recognition’, because it
Rwanda. To address these two objectives we
is important for the international
recourse to the structural multistage
competitiveness of their business and for
modeling as suggested by Crépon et al.
their financial sustainability.
(1998). However, in order to conform to data
availability, we used a two-stage technique, However, it is important to underline that
where in the first stage we determined the empirical findings of this study must be
relationship between innovation decision and considered with cautions. Data used are

16

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
enough dated and didn’t permit to consider and ownership structure of firms using data
all variables of interest, particularly gender newly collected.
and age of owner. Consequently, we suggest
undertaking deeper analyses which
emphasize thematic studies (type of firms)
6. References

Audretsch, D. B., Keilbach, M. C. and Fritsch, M. and Mueller, P. (2004). Effect of

Lehmann, E. E. (2006). new business formation on regional
Entrepreneurship and Economic development over time. Regional
Growth. Oxford University Press. Studies 38: 961-975.
Balachandran, V. and Sakthivelan, M. S. Griffith, R., Huergo, E., Mairesse, J. and
(2013). Impact of information Peters, B. (2006). Innovation and
technology on entrepreneurship (e- Productivity across four European
entrepreneurship). Journal of Business countries. Oxford Review of Economic
Management and Social Sciences Policy, 22(4): 483-98.
Research 2 (2): 51-56. Hall, B. H. (2011). Innovation and
Bronwyn, H.H., Lotti, F. and Mairesse, J. Productivity. Nordic Economic Policy
(2009). Innovation and productivity in Review (2): 167-208.
SMEs: Empirical evidence from Italy. Houghton, K.A. and Winklhofer, H. (2004).
Small Business Economics, 33(1):13-33. The Effect of Website and E-commerce
Brown, F. and Guzman, A. (2014). Adoption on the Relationship between
Innovation and Productivity across SMEs and Their Export Intermediaries.
Mexican Manufacturing firms. Journal International Small Business Journal 22
of Technology Management and (4): 369-388.
Innovation, 9(4): 36-52 Legros, D. and Galia, F. (2012). Are
Crépon, B., Duguet, E., and Mairesse, J. innovation and R&D the only sources of
(1998). Research and Development, firms’ knowledge that increase
Innovation and Econometric Analysis at productivity? An empirical investigation
the Firm Level. Economics of Innovation of French manufacturing firms. Journal
and New Technology, 7(2), 115-58 of Productivity Analysis 38 (2): 167-181.

17

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511
Lööf, H. and A. Heshmati (2006). On the Mohen, P. and Hall, B. H. (2013). Innovation
relationship between innovation and and Productivity: an update. Eurasian
performance: A sensitivity analysis. Business Review 3 (1): 47-65.
Economics of Innovation and New OECD (2005). Oslo Manual. Guidelines for
Technology, Vol. 15 (4/5): 317–344 Collecting and Interpreting
Mairesse, J. and Mohen, P. (2010). Using Technological Innovation Data.
innovation survey for econometric Schumpeter, J. (1940) Capitalism, Socialism
analysis. in Hall, B. H. and Rosenberg, and Democracy. Harper & Row, New
N. (eds). Handbook of Economics and York.
Innovation. Elsevier, Amsterdam: 1130- Tuan, N., Nhan, N., Giang, P., & Ngoc, N.
1155. (2016). The effects of innovation on firm
Mairesse, J. Mairesse, J. and Robin, S. performance of supporting industries in
(2009). Innovation and productivity: a Hanoi, Vietnam. Journal of Industrial
firm-level analysis for French Engineering and Management, 9(2),
Manufacturing and Services using CIS3 413-431.
and CIS4 data (1998-2000 and 2002- Van Stel, A. (2006) Empirical Analysis of
2004).Working paper. Entrepreneurship and Economic
Marques, C.S., Gerry, C., Covelo, S., Braga, Growth. International Studies in
A. and Braga, V. (2011). Innovation Entrepreneurship, Springer.
and the performance of Portuguese Waheed, A. (2011). Innovation and firm‐
business: a ‘SURE’ approach. level productivity: econometric evidence
International Journal of Management from Bangladesh and Pakistan. Working
and Enterprise Development 10 (2/3): paper series no. 2011-061, UNU-
114-128. MERIT.
MINECOFIN (2013). Economic Zemplinerova, A. and Hromadkova, F.
Development and Poverty Reduction (2012). Determinants of Firm’s
Strategy II: 2013-2018. Kigali, May. Innovation. Prague Economic Papers, 4:
487-503.

18

Electroniccopy
Electronic copyavailable
available at:
at: https://ssrn.com/abstract=3240511
https://ssrn.com/abstract=3240511