The Lebanese context
Lebanon is a small-sized economy, with an estimated population of 4.1 million inhabitants, with a relatively high-level GDP of 41.7 billion USD and a per capita GDP of 15,900 USD in 2012, at the time of the survey. It is usually assumed that the economy is mainly driven by services such as banking and tourism representing 75.8% of GDP. Industry represents 19.7% of GDP, thus making it a rather non-negligible economic sector, in contrast with the usually belief that Lebanon has no industry. On the contrary, agriculture represents only 4.6% of GDP. The main problem of the country is the political instability both internally and externally due to the war in Syria that translates in a very high number of refugees and political insecurity concerning the future. Prior to the so-called “Arab spring” (2011), Lebanon experienced a rather strong economic growth reaching 8.5% in 2009, 7% in 2010. After the war in Syria, economic growth has substantially dropped reaching a mere 1.5% in 2011 and 2% in 2012.
Lebanon has a small but diverse and fragmented S&T community embedded in 41 universities and higher education institutions (12 of them with science and/or technology faculties) and 5 rather small public research centers (Hanafi, Arvanitis, 2016, chapter 4). All indicators (publication output, research budget, number of active researchers, etc.), show that most of the research is mainly carried out in four universities: the Lebanese University (UL), the American University of Beirut (AUB), the Université St-Joseph (USJ), The Lebanese-American University (LAU) and Balamand University (UB) and within one of the four specialized research centers of the National Council for Scientific Research (CNRS) as well as in the Lebanese Agricultural Research Institute (LARI). Moreover, research is quite active, with strong and increasing collaborations, mainly with European Union research collaborations: up to half of the scientific publications registered in the Web of Science are co-authored with foreign authors.
Although the size of most manufacturing companies in Lebanon is rather small, the country benefits form a surprisingly active private sector in R&D (Atweh, Arvanitis, 2014, CNRS, internal survey report). The National innovation survey showed that the main difficulty concerning R&D is rather its lack of formal contacts between the academic scientific community and the business R&D units. Interestingly, anecdotal evidence shows an intense personnel interaction between the business sector and universities, both with the public Lebanese University (UL) and the private universities (AUB, USJ, LAU, etc.) (Hanafi, Arvanitis, 2016). Hopefully, initiatives such as the creation of the Lebanese Industrial Research Association (LIRA) in 1997, which has been re-instituted in 2015 after some five years of inactivity, the success of the business incubator Berytech, and the promotion of joint industry-university research projects has permitted an increase of private sector contributions and participation. There is also an increasing number of small private research institutes, often NGOs, that carry out studies, mainly socio-economic studies and policy-oriented analysis, opinion polls, market studies, and studies for international organizations. They very frequently use the services of university staff (Hanafi, 2011).
Lebanon is a very active entrepreneurial country to all experts interviewed, and, although this entrepreneurship is usually not automatically considered innovative, all field interviews show a definite orientation toward rather audacious entrepreneurial activities, with strong foreign experiences. Innovation has been a central debate concerning the private sector after 2000, triggering many studies on the subject.
Stakeholders in various studies have identified considerable and persistent obstacles to innovation, most notably weakly enforced intellectual property; a limited market; meager training on innovation; little research and development; and poor infrastructure (Doumit, Chaaban 2012). Nonetheless, the innovation survey has been providing surprising results concerning these obstacles, and to a great extent do not confirm the view of a passive entrepreneurs, limited to well-known markets (Chakour, 2011; Ahmed, Julian, 2012; Mezher et al., 2008). These studies tend to defend the idea that the business practices, mainly because most firms are SMEs, family owned and managed by the family members are opposed to innovation. Nonetheless, the few empirical studies of entrepreneurship on this particular topic point rather to a quite active attitude toward innovation even in SMEs (for a review, see Stel, 2012). What is absolutely certain is a lack of institutional innovation (banks and other financing institutions are especially seen to be conservative) (Chakour, 2001), lack of institutional support and infrastructures by the state (Ahmed, Julian, 2012), and a very unstable and rapidly changing market to which firms need to adapt (Mezher et al., 2008). These obstacles will be addressed in part 3 of this article.
The Lebanese research and innovation has been described with a lot of detail (Hanafi, Arvanitis, 2016). Lebanon has been the country of numerous experiences to support an ‘entrepreneurship ecosystem’, in particular because of post-civil war and persistent political difficulties that translate in high emigration and instability in the region. Among the various initiatives the most remarkable has been the foundation of Berytech in 2001, which is probably the most successful business incubator in the MENA region. To date, Berytech has housed more than 170 entities, assisted more than 2,000 entrepreneurs in several outreach programs, disbursed more than US$ 350,000 in grants to start-ups, and invested more than US$ 5 million in Lebanese technology companies. It was among the first of such institutions in the region to receive accreditation from the EU as a Business Innovation Center, opening access to international networks for its companies and affiliates. In 2012, and with the support of the EU, Berytech launched the Beirut Creative Cluster, grouping more than 30 leading companies in the multimedia industry, and was the European Bronze Label for Cluster Management Excellence. Other business incubators of smaller size and less successful have been created in the North of Lebanon (BIAT in Tripoli), in the South (Southbic in Saida). Finally, Berytech has been housing the first venture capital fund of Lebanon for an amount of US$6 million, for Lebanese technology start-ups.
Another important initiative has been Kafalat, which provides funding, subsidized and guaranteed loans and guarantees for SMEs. Kafalat has maintained its action steadily and created also an innovative business fund. Its blend of both risky business funding and mainstream support to entrepreneurial activities has been instrumental in periods of low investment from the banking sector. It set-up a collaboration with the World Bank addressed to support small and micro-enterprises for more than US$ 30 million.
Public policy has not been particularly efficient in supporting business investments. In order to provide some support, and also trying to attract some foreign investment, the Lebanese government created an Investment Development Authority Lebanon (IDAL) seeking to enhance entrepreneurship through tax incentives, administrative reforms and the support of the incubators. A variety of international organizations are very present in Lebanon such as the World Bank, many United Nations organizations (ESCWA, UNDP, UNIDO, UNRWA) as well as a variety of foreign foundations. The European Union has had a quite permanent presence in the country, although the policy is limited by the absence of a bilateral agreement in science and technology contrary to many other Mediterranean Partner countries of the EU (Arvanitis et al., 2013).
It’s interesting to note that all the initiatives in favor of the business sector have been supported by private institutions and mostly a variety of academic non-for profit institutions, which is a Lebanese asset. Such is the case of the Université Saint-Joseph (USJ) that housed and supported the creation of Berytech, the American University of Beirut (AUB), that more recently created an entrepreneurship center. Some smaller initiatives are known from a wide range of private sector and civil society organizations supporting entrepreneurs (e.g. entrepreneurship education, funding for mature entrepreneurs and microfinance institutions). The public sector in favor of research and innovation, apart from the decisive action of the National research council, has been very timidly developed; one can mention the existence of the Industrial Research Institute (IRI) which is an organ of the Ministry of industry, acting for mainly certification and testing activities; the activities of the Lebanese industrial association (ALI) that supports industrial activities and has been involved in supporting technology in industry quite actively; LIRA, a scheme to enhance industry-research activities, launched in 1997 by ALI in collaboration with the Lebanon’s National Council for Scientific Research (CNRS) and the Ministry of Industry.
Nonetheless, although experts praise the survival and expansion of businesses in Lebanon after the hard times of the civil war (Ahmed, Julian, 2012) they also signal the absence of a coherent industrial policy, and the difficulty in defining a strong policy. In 2012, the government launched a series of policy measures in support of the country’s industrial sector, proposing fiscal measures, subsidies, import protection, and support to specific sub-sectors such as the aluminum industry and other energy intensive industries.
Finally, it is necessary to remind that the National Council for Scientific Research (CNRS) has drafted a comprehensive science, technology and innovation policy (CNRS, 2006) after a consultation of 30 experts. An action plan had been approved by the government in 2002, but its implementation was interrupted by the assassination of Prime Minister Rafic Hariri in February 2005. Since then there has been no explicit innovation or science and technology policy, but at the same the CNRS has been promoting actively research, international collaborations, and supported research-production activities (See Hanafi and Arvanitis, 2016, p. 191).
Sample and Data description
This study uses the data gathered from the first national innovation survey in Lebanon, which was conducted in the fall of 2012 by the Lebanese National Council for Scientific Research (CNRS) with the support of the World Bank. The questionnaire was designed by an author of this article on behalf of the CNRS and the survey was conducted by INFOPRO, a private enterprise in Lebanon under CNRS guidance. Survey results were handed over to CNRS early in 2013.
A total of 478 firms operating in the industrial sector have been surveyed. The sample was composed of manufacturing firms (441 enterprises), micro-firms (110 enterprises) and firms in the information and communication technologies (ICT) (37 enterprises). All industrial sectors have been included but activities in agriculture, mining, pharmaceutical and services (except ICTs) have been left out of the realm of the survey. The survey sample was designed also to be as close as possible to the distribution of the industry in Lebanon. An effort was made to cover more appropriately the very small firms (1 to 5 employees) and also a bias toward medium to large firms that are known to be more inclined to R&D and technological development.
Employment represents around 20,000 persons in the Innovation Survey, when the 2007 Industrial Survey of the Ministry of Industry counted 75000 employees in manufacturing industry. The Innovation survey thus represents around 10% of enterprises of the industrial survey and 25% of employment.
The ICT sector comprised of 37 firms in the sample represents about 2700 persons, while 110 micro companies represent around 1000 persons. The bulk of the sample (which is called the “industrial segment” in the survey report) contains 441 companies (85.5% of the sample). Most firms are rather young: 59.8% in the survey were created in the last twenty years. The survey shows no statistical relation between the age of the firms and their innovation capabilities, which is first indication that that their accumulated experience along the lifetime though substantial is not the only determinant of the introduction of innovation. As we will see, it rather depends upon strategic decisions.
60% of the firms do not have engineers among their employees and 25% of firms do not have technicians among their employees. 42% of firm owners have a B.S. or License degree. Only a total of 23% of firm owners have higher levels of education such as a Master’s or engineering degree. Less than 7% of firms with market shares between 50% and 80% have more than 20% of employees with science oriented majors.
The most common legal types of firms found in the survey were Anonymous Capital firms (S.A.R.L) or Family Owned, and they rely mainly on private and national capital. Foreign capital represents a negligible source of investment. Firms mostly target the domestic market, whereas 24.9% export more than half their sales and 12.7% of firms are strongly export oriented. The sampled firms are estimated to represent around 10% of industrial exports of the country. Only 56.9% of the survey sample is exporting.
The survey was asking firms to declare what was their share of the market. 167 firms (34.9% of the whole sample and 34.5% of the industry segment of the sample) did not answer to this question; indicating in the survey, that they did not know these figures. Nonetheless, 81% of firms know both the number and the names of their competitors and 12% mention they know their competitors without mentioning a name. In other words, the firms believe they know quite well their market. Not surprisingly, smaller firms know less well their competitors; nonetheless, the figures statistically do not show a relation between the size of the firm and the degree of knowledge of the competition. Similarly, more firms that export seem to know better the market competition, but again there is no statistically significant relation between export-orientation and knowledge of the competition.
The survey undergoes in a detailed description of R&D, showing firms seem to prefer to engage in R&D mainly for trouble-shooting and product development. The majority of firms have quite bad capacities of identifying future technologies – with the notable exception of ICT companies. Although only 23% of firms (110 firms) report having an R&D department, some 185 more firms (38%) report developing R&D activities even without having an R&D department (“Informal R&D”). Thus, 62% of firms realize some R&D activity and 38% have no R&D activity whatsoever. Traditional sectors, with the exception of Food products, have lower intensity in R&D than sectors oriented towards industrial clients (mechanics, machinery, equipment, chemistry…). For all these sectors, internal R&D is among the main channels of innovation.
Few firms accepted to report on expenditures in R&D (18%). 89 firms responded for this figure in 2010, reporting a total of US$ 4,992,500 and 87 firms for 2011, reporting a total of US$ 5,757,000. Average figures are an average of US$ 56,000 (2010) and US$ 66,000 (2011) per firm. Despite a bad economic prospect, average R&D expenditures for those firms reporting these figures, increased by 17% between 2010 and 2011. If all firms had answered, the survey sample would have estimated overall R&D expenditures between 9.8 and 11 million USD. By extrapolating the survey data, taking into account size and sector distribution, and by excluding the very atypical case of ICT firms and micro-firms, the survey estimated that the R&D expenses of the Lebanese industry to be around US$ 120 million for 2011, which represents around 0.3% of the GDP. This is a very large and until this survey unknown figure, showing the importance of private R&D.
R&D expenditures are highly concentrated and larger firms tend to spend proportionally more on R&D than middle-sized firms. The sectors where higher spending firms are located are ICTs and software, food and beverage, machinery. Sectors concentrating more R&D expenses are ICT, furniture and consumer goods, food and beverages, machinery and equipment. The share of R&D personnel in the workforce of a firm correlates loosely with the amounts dedicated to R&D in total sales, and R&D expenditures do not correlate to sales in significant levels. This is common to all innovation surveys as R&D has no direct effects on sales and profits but is the guarantee of a sustainable presence in markets. Moreover large firms do not necessarily spend more on R&D, but when they do, they spend proportionally much more than medium-sized firms. In other words, spending in R&D is related to a strategy that involves R&D and innovation.
Finally, the innovation survey indicates a close relation between the R&D activities and innovation, and all firms that innovate with product, services or process, show important R&D expenditures.
It’s necessary to remind that patenting and intellectual property management in Lebanon is very low. Less than 4% of the firms report registering a patent claim either in Lebanon or in a foreign patent system. Trademarks are deposited in one quarter of the cases. In 2010-2011 63% of firms acquired an internationally recognized quality certificate while only 28% of firms sought a quality certification. Quality is one of the main motives of technical interest of firms. Table 2 present some statistical details related to the sample.
Table 2 - Summary statistics of the variables
Variable Number of observation Mean Standard Deviation Min Max Firm Age 478 2.82636 1.16786 1 5 Firm size 478 2.97071 1.32036 1 5 Export 478 22.1970 30.1355 0 100 Skill 478 0.47933 0.31090 0 1 R&D 478 0.55439 0.49755 0 1 Competition 478 25.2854 20.4051 .001 100 Partnership 478 0.39748 0.48989 0 1 Technology Transfer 478 0.81380 0.38966 0 1
In order to study the factors influencing firm’s innovation, we use a probit econometric model to determine the basic determinants of a firm’s decision to innovate. We use variables such as the age, size, competition, etc. Probit models are regression-based models used to analyze binomial variables.
In our case, the innovation decision is measured by a binary dependent variable, which is equal to 1, if the firm is innovating; or equal to 0, if the firm does not innovate. Probit models measure the probability of a variable influencing this innovation results. These models are generally used to model binary data.
Considering the innovation response as the dependent variable, let p(innovation) be the probability of the firm to innovate, p(innovation) = p[innovation = 1].
We define a latent variable y*i given by the following relation:
y*i is observed only if the firm i innovated
Where Xi is the is a vector representing the variables that summarize the characteristics of the firm i, βi a vector of unknown parameters associated with the vectors Xi et εi the term of error, and we suppose εi follows a normal law N(0;σ2).
The observed dichotomy variable yi is related to the latent variable y*i by the relation as follows:
yi = 1 if y*i > 0 where yi = 1 if firm innovate yi = 0 if y*i ≤ 0 where yi = 0 if firm did not innovate
Taking into account the presence of a problem endogeneity, the basic model specification estimated for the innovation decision yi is as follows:
Instrumental variable models
The objective of this model is to estimate, for the totality of the sample, the impact of each explanatory variable on the probability of innovation in the Lebanon’s firm.
Model 2 & Model 3
In model 2 and 3, we used two different subpopulations. Model 2 is estimated just for firms belonging to the manufacturing sector excluding those belonging ICT sector. Model 3 is estimated just for a random sample of 85%. Thus it is estimated only 406 firms
innovation = f(Age; Size; export; competition; partnership; technology trasfer; R&D * Skill)
innovation = f(Age; Size; export; competition; partnership; technology trasfer; R&D * Skill; R&D * partnership)
innovation = f(Age; Size; export; competition; technology trasfer; ForeignCapital; R&D * Skill; R&D * partnership; R&D * ForeignCapital)
In models 4, 5 and 6, we propose to estimate the impact of R&D activities on innovation while controlling this impact using specific variables such as Skill, partnership and foreign capital share.
The unobservable random shocks affecting the firm’s decision to innovate could affect his ability to export. Indeed, export intensity variable is a dichotomous endogenous. It’s instrumented using other variables in the dataset as follows:
export = f(Age; Size; Sector; Skill; Group’Membership; Foreign Capital Share; Firms’location; R&D; competition; partnership; technology transfer)
To rule out the possibility that our results are driven by our choice of instruments, the three selected instruments (Sector, Group membership, Firms’ location) are introduced into the model: each separately, in pairs and three at the same time. After accounting for this endogeneity, our results suggest that by introducing the three instruments at the same time gives the most robust results. Results of these regressions are available from the authors upon simple request.