Tech Start-up Ecosystem in Beirut FINDINGS AND RECOMMENDATIONS Content Authors and Acknowledgements 1 Executive Summary 2 Measuring and Analyzing the Tech Start-up Ecosystem in Beirut 5 Measuring the Tech Start-up Ecosystem 5 Analyzing the Tech Start-up Ecosystem 6 The Tech Start-up Ecosystem in Beirut 9 Skills 11 Supporting Infrastructure for Entrepreneurship 13 Investment 15 Community 16 Start-up Success Factors 19 Gap Analysis and Policy Recommendations 20 Summary of Gap Analysis and Stage of Ecosystem 20 Policy Recommendations 21 Program in Focus: Circular 331 24 References 26 Survey Methodology and Analysis 27 LIST OF TABLES Table 1.1 Networking Assets (Ecosystem’s Support Infrastructure) 7 Table 1.2 Categories of Ecosystems 8 Table 3.1 Development Stage of Ecosystem (Per Area and Overall) 20 Table 3.2 Policy Recommendations 21 LIST OF FIGURES Figure 2.1 Start-up Growth in Beirut, 2009-2015 9 Figure 2.2 Time to Complete Procedural Tasks in Life Cycle of a Start-up Across Regions 10 Figure 2.3 Gender Distribution and Job Functions of Founders 10 Figure 2.4 Average Age of Founders 11 Figure 2.5 Educational Level of Founders in Beirut 11 Figure 2.6 Highest Degree Earned by Founders at Time of Founding 12 Figure 2.7 Educational Experience of Founders in Beirut 13 Figure 2.8 Previous Function and Role Type of Founders in Beirut 12 Figure 2.9 Accelerator Programs Attended by Start-ups in Beirut 14 Figure 2.10 Accelerator’s Quality/Quantity Ratio (Selected cities) 14 Figure 2.11 Median Investment Amount by Year of Existence 15 Figure 2.12 Growth of Founders and Connections in Beirut 16 Figure 2.13 Connections in Beirut’s Tech Ecosystem 17 Figure 2.14 Progression of Ecosystem Connectivity 18 Figure 2.15 Short-Term Success 19 Figure 2.16 Circular 331 Effectiveness in Various Areas 24 Figure2.17 Effect of Circular 331 on Raising Funding and Developing the Support System 25 List of Boxes Box 1.1 GERN Ecosystems Connection Project 6 Box 2.1 Accelerators and Incubators 13 Box 3.1 What is Circular 331? 22 Box 3.2 What Are Coding Bootcamps? 23 Terms Used Start-up A newly established business venture that is in its first stages of operation. This report focuses on tech start-ups, which are start-ups that have a technological component. These start-ups are typically designed to scale up quickly. Start-up Ecosystem The combination of people, start-ups at various stages and other stakeholders and organizations supporting or connecting to these start-ups, interacting in multiple dimensions to create and scale new start-up ventures. Scale-up (Firm) OECD defines a scale-up as a firm that has an average annualized return of at least 20 percent in the past three years with at least 10 employees at the beginning of the period (OECD 2007). Venture Capital (VC) An institutional investor that provides financing to start-ups and small early stage firms. Usually VCs look for high growth potential firms to exit the investment in the short term. Angel Investor An investor who invest in ventures (primarily at an early stage) in their personal capacity (that is, investing their personal money) and may or may not have an active advisory or guidance role for the founders in the venture. Mentor An experienced person who can provide advice, knowledge, or connections to a start-up founder. Mentors usually have strong business acumen and practical experience through former entrepreneurship experience or industry knowledge. Business Acumen Theoretical or practical knowledge of how to develop and manage a business, including proclivity and speed in understanding and dealing with risks and opportunities in the business environment. Exit Refers to the point at which a founder or early stage investors sell their stakes (start-up exit) in the venture, generally either in a private acquisition or public offering. This report refers more widely to “start-up exit” as the point at which a start-up is sustainable or it has received sufficient funding to grow in the medium-term (that is, over the next five years). Authors and Acknowledgements The authors of this report are Victor Mulas, Kathy Qian, and Scott Henry. Matt Lerner, Kwame Robinson, Mireille Raad, Anastasia Nedayvoda, Ainsley Lloyd, Nga Nguyen, Cecilia Paradi-Guilford, and Elene Allende Letona contributed to this report through data, economic analysis and inputs. The survey was conducted by Endeavor Lebanon, and Berytech complemented the data collection. The Ministry of Economy of Lebanon also provided support through data access and outreach. The report was edited by Colin Blackman (Camford Associates) and designed by Wenceslao Almazan. The peer reviewers were Randa Akeel, Wissam Harake, Abdalwahab Khatib, and Zeina El Khoury. This report was partially funded by the World Bank and the Korean Partnership Facility. We are grateful to Endeavor Lebanon, and particularly its Managing Director, Tarek Sadi, for the valuable in-kind contribution to the survey collection process, guidance and advice through this process.  1 Executive Summary Technology is one of the main drivers of description and measurement of the city’s tech productivity and economic growth (Isaksson, start-up ecosystem; and (ii) a comparison and gap Ng, and Robyn 2005). Developing countries have analysis of four key components of the tech start-up traditionally had difficulties in both developing ecosystem (skills, finance, entrepreneurial supporting technology and absorbing foreign technology. infrastructure, and community). Seventy to eighty percent of the productivity gap between developed and developing countries is The objective of this report is to provide a better estimated to result from the lag in the adoption of understanding of the status of Beirut’s start-up technologies in these countries (Comin and Marti ecosystem and provide policy recommendations 2013; and Comin and Hobijn 2010). for policymakers and other stakeholders who are interested in supporting the growth and sustainability Tech start-ups are an effective mechanism to of the ecosystem. both create local technology and absorb foreign technology. In recent years, there has been a Analysis Limitations surge in tech start-ups across the world. Fueled by global technology-led cost reductions and Measuring the tech start-up ecosystem is difficult. increased access to resources, tech entrepreneurs Relevant databases of start-ups are not readily have emerged in both developed and available, and the fast-paced and multidimensional developing countries. However, there is very little dynamics of start-up ecosystems, with new ventures understanding of how these tech entrepreneurs constantly being created, failing, being closed, being form ecosystems, their internal dynamics of these bought or transformed (changing names and/or ecosystems, how they work, what makes them purpose), makes accurate measurement over time grow and achieve sustainability, how they connect inherently difficult. with the local economy to drive productivity and employment, and why some ecosystems are more For this analysis, 218 entrepreneurs were surveyed effective than others. in Lebanon between February and August 2016 and relevant data was collected for 142 start-ups This report is part of a broader research initiative and 196 start-up founders.1 The survey was based (See Box 1.1) aiming to provide answers to these on the standard questionnaire from the Global questions. It provides new data and analysis of Entrepreneurship Research Network (GERN) Ecosystem the tech start-up ecosystem in Beirut. The analysis Connection project (see Box 1.1). The findings and comprises: (i) an attempt to provide an accurate conclusions of this analysis are based on this survey 2 TECH START-UP ECOSYSTEM IN BEIRUT and so there are some limitations to this analysis. The early-stage start-ups (a unique characteristic for this stage dataset is not exhaustive and only represents a subset of development of an ecosystem). of the ecosystem’s start-ups. Moreover, it is subject to survivorship bias and does not include start-ups that The supporting infrastructure (for example, accelerators, were no longer in business when data was collected. mentors) and the community is still maturing. Accelerator Historical data about start-ups was collected through programs do not seem to provide sufficient quality training existing founders who were available at the time of for entrepreneurs to be sustainable and quality mentors the survey. are scarce, with few instances of angel investors. Events and community building appear to have been catalyzed Despite these limitations, the subset of start- by the programs of the Banque Du Liban (BDL), the Central ups surveyed provides unique insights into the Bank of Lebanon (BDL Accelerate and Circular 331), and ecosystem. Data was collected by snowballing from donor and government action still drive many activities in public data contained in existing databases, networks the ecosystem. Since the entrepreneurial community is still from key stakeholders (for example, accelerators, events, in its early stages, like-minded individuals do not connect and so on), as well as recommendations from founders. easily with one another to form clusters, suggesting a silo Although there are start-ups that were not captured in approach among ventures from different networks. this survey, nevertheless it provides one of the richest samples of data collected to date on the most influential Investment is currently abundant but there are signs founders, start-ups, intermediaries, and other ecosystem that a start-up bubble may be forming, distorting the stakeholders in Beirut. ability of the ecosystem to select the best start-ups and generate success stories that are sustainable over time. Findings Compared to other early and middle stage ecosystems (such as in Tanzania and West Bank & Gaza), processes in The tech start-up ecosystem in Beirut is an early-to- Beirut for start-ups are lengthier for credit and funding middle stage ecosystem that has passed its nascent (with West Bank & Gaza), suggesting more difficulties growth phase but is still far from maturity (see table (linked to procedures or access to finance requirements) below). The key strengths of the ecosystem are the in obtaining funding for start-ups. Finally, sustainability presence of talented people, with highly educated of the current level of access to funding, which is fueling founders (90 percent of them having a university degree the ecosystem, is not ensured without current policy and 50 percent with graduate degrees) with sufficient intervention (Circular 331), which may be crowding out relevant experience, and the availability of funding for sources of more sustainable private investment.  3 DEVELOPMENT STAGE OF BEIRUT’S START-UP ECOSYSTEM Stage Ecosystem Area Nascent Advancing Mature 1. Community 2. Skills 3. Supporting Infrastructure 4. Investment 5. Constraints OVERALL Policy Recommendations Based on this analysis, a set of policy recommendations were developed (see table below), which are further expanded in the report. Ecosystem Area Policy Recommendation Objective Strengthen coordination mechanism Expand cluster connectivity, coordinate private and ecosystem support program with and public action and increase density of stakeholders connections among all stakeholders Increase absorption by ecosystem of Community Expand ecosystem community to traditional international talent connectivity, connect industries and absorb international connectivity with domestic corporate nontech sectors Expand practical education in university and Address practical business acumen gaps, through rapid skills training programs and provide pipeline of talent for start-up scale-up, include public education programs and include lower income/educated population Skills Increase capacity building of mentors in accelerators and attraction of international Address lack of sufficient number of quality talent (as mentors, entrepreneurs or capacity mentors and strengthen support services builders) to the ecosystem Support Infrastructure Assess quality of start-up funding and prepare Address potential bubble of start-ups and phase out of Circular 331 introducing private distortion of investment market, and ensure sector funding sustainability of funding Investment Address processes constraints (for example, Reduce constraints for start-ups incorporation access to loans and funding) and operationalization Constraints 4 TECH START-UP ECOSYSTEM IN BEIRUT Measuring and Analyzing the Tech Start-up Ecosystem in Beirut This report provides an analysis of the tech start-up ecosystem (countrywide). Databases are both open and proprietary. in Beirut. The analysis comprises: (i) an attempt to provide an Access to proprietary databases is limited and in some cases accurate description and measurement of the city’s tech start- restricted (not being open to wider research). The most relevant up ecosystem and (ii) a comparison and gap analysis of four open databases of start-ups are provided by CrunchBase and key components of the tech start-up ecosystem (skills, finance, AngelList. Neither of these databases, however, necessarily entrepreneurial supporting infrastructure, and community). provide accurate or complete information. CrunchBase is a self- reported database that is not curated by an official entity, and For the purposes of this report, tech start-ups are defined as as such it can have inaccurate information such as closed or for-profit business ventures that a) have a financial model transformed ventures still being posted with the original data, targeting high growth, and b) employ an innovative and founders omitted, and so on. AngelList generally contains more technology-enabled approach to the product or service that accurate information since start-ups listed there have received they provide to ensure scalability. These ventures may not be or are actively soliciting investment from angel investors or profitable at their current stage. venture capital (VC) firms. Other global start-up repositories, such as Startup Genome,2 build on these databases with In order to capture the whole tech start-up ecosystem, this additional self-reporting data from start-ups, and therefore report expands the concept of start-ups to include newly are subject to similar limitations. Finally, while LinkedIn can emerging start-ups as well as small and medium enterprises provide more accurate data of start-ups through funders and (SMEs) that were once start-ups and have reached the scaling employers, data access and use restrictions make its use for phase. This definition allows us to collect data to describe the independent research purposes difficult. evolution of the tech start-up ecosystem over time as start-ups grow and succeed. Since the aim of this report is to analyze the These global databases, however, are of little use in building start-up ecosystem, all relevant ventures and actors that have an overall picture of the tech start-up ecosystem in many an influence on the ecosystem are included, even though they developing countries, as they contain little information may no longer be start-ups. on these regions. At the time of conducting this analysis, CrunchBase contained only 122 start-ups from Lebanon, and AngelList only 126. Measuring the Tech Start-up Ecosystem Regional and local start-up databases can be richer in data and more accurate, since they are often the result of an active Measuring the tech start-up ecosystem is a difficult task. effort to track the activity and lifecycle of start-ups. Examples Relevant databases of start-ups are not readily available, and of these databases are Digital NYC3 in New York or Tech Map4 the fast-paced and multidimensional dynamics of start-up in London. However, these databases are not present in many ecosystems, with new ventures constantly being created, failing ecosystems, particularly in developing country ecosystems, and being closed, and being bought or transformed (changing and given their localized methodologies, their data is difficult names and/or purpose), makes accurate measurement over to utilize for comparative analytics. Other datasets, such as that time inherently difficult. of the Global Accelerator Learning Initiative,5 only have enough power to report data at the regional level. Some databases have tried to collect information on start- ups. These databases are global, local (mostly at the level of While official government or NGO-managed databases in metropolitan areas’ ecosystems) and, in some cases, domestic developing countries can provide richer and more accurate Measuring and Analyzing the Tech Start-up Ecosystem in Beirut 5 data from SMEs and larger companies, they also lack both data was collected. Historical data about start-ups was collected breadth and depth when it comes to start-ups. In the case of from existing founders available at the time of the survey. Lebanon, the last official census was conducted in 1931. These limitations are discussed in detail in the Methodology section of Appendix A. Despite these limitations, the subset of To combat the poor availability of data, a survey was designed start-ups surveyed is likely to be representative sample since and deployed by extending the standard questionnaire the start-up data was collected by snowballing from public from the Global Entrepreneurship Research Network (GERN) data contained in existing databases, networks from key Ecosystem Connection project (see Box 1.1). For a broader stakeholders (for example, accelerators, events, and so on), and description and technical details of this survey, see Appendix recommendations from founders. Although there are start-ups A. We surveyed 218 entrepreneurs in Lebanon from February that were not captured in this survey, nevertheless it provides to August 2016 using an online interactive questionnaire one of the richest samples of data collected to date on the distributed through local partners by email, phone, and in most influential founders, start-ups, intermediaries, and other person.6 From these interviews, data for 142 relevant start-ups ecosystem stakeholders in Beirut. and 196 start-up founders were collected. This sample provides unique insights into the characteristics of founders, start-ups, This report assumes that, owing to the fast-moving nature investors, and supporting infrastructure in Beirut, as well as the of start-up ecosystems, any attempt to accurately measure relationships between them.7 the tech start-up ecosystem is inherently flawed – any measurement will be obsolete immediately after collection. The resulting dataset is not exhaustive, however, and only The findings and recommendations provided in this analysis represents a subset of the Beirut ecosystem’s start-ups. should be taken with this limitation in mind. Less emphasis Moreover, it is subject to survivorship bias and does not should be placed on exact numbers, which are subject to include start-ups that were no longer in business when data change with the addition of more start-ups and sensitive to minor tweaks in methodology. Rather, the data collected provides insights into general trends and the dynamics of the BOX 1.1 GERN ECOSYSTEMS ecosystem that can inform specific policies. This analysis should CONNECTION PROJECT not be considered in isolation, however, and policy makers are encouraged to confirm these findings through other available resources (for example, perspectives from practitioners and anecdotal evidence). The Ecosystem Connections Mapping Project aims to map start-up ecosystems across the world by collecting data on start-up founders Analyzing the Tech (for example, education, work experience, Start-up Ecosystem serial entrepreneurship, and so on) and the connections between themselves and other key The following analysis of the tech start-up ecosystem in Beirut stakeholders in their ecosystem (for example, is based on the data collected through the survey described mentors, investors, accelerators, universities, above. This report analyzes four key elements: i) skills, ii) and so on) to better understand and support finance, iii) supporting infrastructure for entrepreneurship, entrepreneurs in local start-up ecosystems. This and iv) community. The analysis first describes the status of data aims to identify gaps in ecosystems and each of these elements based on the data collected, and then provide a base for policy action to address these compares the results with those reported by both average and gaps and support growth and sustainability of successful start-up founders. start-up ecosystems. Constraints for start-up founders and the ecosystem are The project has mapped over ten start-up described in as much as they are referred by founders. We ecosystems in cities across the world (including, acknowledge this report does not cover all existing constraints, Bogota, Cairo, London, New York, and Singapore). such as broadband or physical infrastructure in the city, which The survey conducted for this report is also part may well help expand the ecosystem.8 of this project, adding Beirut to the ecosystems mapped. The leading partners of this project are Endeavor Insight and the World Bank. For the purposes of this analysis, successful start-ups are defined as those that have been funded and those who employ Source: http://gern.co/gern/ecosystem-connections- people. “Short-term success” is defined as obtaining funding mapping. from at least one investor; “long-term success” is defined as continuously hiring employees (as a proxy for growth). Comparing average founders and start-ups with successful ones highlights which characteristics (in terms of education, 6 TECH START-UP ECOSYSTEM IN BEIRUT experience, connections, and so on) are more predominant in available. Given limitations of start-up data across start-up successful start-ups in Beirut and whether they are consistent ecosystems, the analysis of this report will use available with those in other ecosystems or with global trends (where data as proxies for comparable start-up ecosystems at data is available). Where comparable data is available on other different level of maturity (regardless of region). For ecosystems surveyed under the GERN Ecosystems Connections instance, Dar es Salaam’s (Tanzania) ecosystem is used as Project, local results are benchmarked with other ecosystems a nascent ecosystem and West Bank & Gaza’s ecosystem to understand if there are gaps that could be addressed. as an advancing one. We understand these proxies may not be a perfect comparison but we believe they can still The four elements this report analyzes represent the key provide a valuable reference point. ingredients needed for tech start-up ecosystems to grow and be sustainable. Skills aims at understanding the educational This analysis is conducted under the premise that start- and work experience that founders have and those that are up ecosystems are communities of stakeholders and more common for successful founders. Finance looks at the that the success of such ecosystems is linked to the funding obtained by start-ups during their lifecycle and the maturity, health, and sustainability of the community. general availability of such funds. Supporting infrastructure Previous research of the World Bank (Mulas, Minges, and for entrepreneurship seeks to understand the quantity and Applebaum 2015) shows that tech start-up ecosystems act quality of support programs and resources for start-ups to as communities and that centrality (that is, the number of succeed. Supporting infrastructure encompasses accelerators ecosystem’ stakeholders a founder of start-up is connected and incubators, mentors, events, and other ecosystem and/ to directly or indirectly) is critical for start-up success. This or skills-building resources. Finally, community looks at the finding is also consistent with research from Endeavor maturity of the ecosystem as a network of stakeholders that Insight showing that access to mentors increases the support each other (directly or indirectly) for the successful probability of start-up success.9 In this environment, the outcome of start-ups. supporting infrastructure acts both as a skills and network provider and is critical for ecosystem sustainability. The This report uses comparable data from other ecosystems to ecosystem’s supportive infrastructure is mainly comprised produce benchmarking analysis when homogenous data is of networking assets (see Table 1.1). TABLE 1.1 NETWORKING ASSETS (ECOSYSTEM’S SUPPORT INFRASTRUCTURE) Collaboration Community-Building Skills Training Collaboration Spaces/Networks Network of Events Events Spaces of Mentors Mentors Collaboration and Bootcamps and community-building technology training Accelerators (network Angel investors Meetups spaces (e.g., coworking linked to community value) (network value) spaces, makerspaces, building fab labs) Rapid technical and Tech community Incubators (network Venture capital entrepreneurial skills events/conferences value) (network value) programs Networks of mentors and start-up “alumni” networks (if different from accelerators, incubators, angel investors, and venture capital) Source: Mulas, Minges, and Applebaum 2015. Measuring and Analyzing the Tech Start-up Ecosystem in Beirut 7 Based on these four elements, the analysis categorizes pipeline is starting to form with increasing private sector ecosystems into three broad categories: a) nascent ecosystems, investment in early stage star-ups, but there are still gaps in b) advancing ecosystems, and c) mature ecosystems (see the path to scale up and exit. Table 1.2). c) Mature Ecosystem. These ecosystems have a large a) Nascent Ecosystem. There are a limited number of start- number of start-ups in all stages (for example, growing, ups, most of which are in the very early or early stages. scaling up, exits) and the ecosystem is highly interconnected. The community of entrepreneurs is forming, and has low The majority of entrepreneurs have business acumen and density of connections with few clusters, if any. In these previous relevant business experience, there are several ecosystems founders lack business experience, mentors are generations of start-ups with multiple serial entrepreneurs scarce and inexperienced and there are very few generations and successful exits. Mentors are abundant, they have of entrepreneurs (most entrepreneurs are in their first or solid practical experience, and there is a solid base of angel second venture), there are few or no start-up exits or if there investors. The ecosystem is an international hub itself and are any, they are outliers. There are few or no international attracts international talent. The finance pipeline has no connections. The finance pipeline has multiple gaps and gaps and early stage funding is provided by sustainable private early stage finance is very rare (if it exists). private funds. b) Advancing Ecosystem. These ecosystems have an These categories are broad and serve to provide a sense increased number of start-ups, with most in the early stages of where ecosystems are in their life cycle. As this research but with increasing numbers of scale-ups. The community continues and we can access data from a larger sample of of entrepreneurs has several clusters and a high density, ecosystems from the GERN Ecosystems Connection project, we and there are a handful of successful start-up exits. There will be able to provide more concrete metrics on these stages. is still a lack of business acumen among entrepreneurs but there are a growing number of serial entrepreneurs and the This report focuses on supply-side factors of start-up ecosystem has more than three generations of start-ups. ecosystems and does not analyze extensively the demand side. There are an increasing number of international connections Further analysis would need to be conducted to understand and mentors with local start-up experience. The finance deeply demand-side factors. TABLE 1.2 CATEGORIES OF ECOSYSTEMS Stage Ecosystem Area Nascent Advancing Mature Limited number of start-ups, most Increased number of start-ups, most in early Large number of start-ups in all in early stage stage with increasing number of scale-ups stages Low density of connections Several clusters Highly dense, hyperconnected Low number of clusters High density in clusters clusters Community Lack of business acumen and experience Limited business acumen Business acumen and practical experience Very few serial entrepreneurs Increasing number of serial entrepreneurs and limited generations of and more than 3 generations Several generations of start-ups entrepreneurs Exits start to appear Successful exits Skills No substantial exits Mentors are available and they have local Plenty of mentors with sound Mentors are scarce and practical experience practical experience inexperienced Increasing number of international International hub attracting Supporting No international connections connections international talent Infrastructure Gaps in finance pipeline No gaps in finance pipeline Finance pipeline with some gaps Very few private sector funding Private early stage finance Private early stage investment exists Investment opportunities sustainable 8 TECH START-UP ECOSYSTEM IN BEIRUT The Tech Start-up Ecosystem in Beirut The tech start-up ecosystem in Beirut is an early to middle but incorporation, renting an office, hiring qualified personnel, stage ecosystem, that has passed its nascent growth phase and obtaining a credit line from a bank typically takes about but is still far from maturity. On average, each year, 12 more a month. Obtaining funding is a much lengthier process start-ups are created than in the previous year, resulting in a 24 taking about four months to close (for example, individual or percent compound growth rate in start-up creation since 2009 institutional investor). (see Figure 2.1 ).10 Compared to other early and middle stage ecosystems (such as About two-thirds of the start-ups surveyed reported hiring at Dar es Salaam, Tanzania, and West Bank & Gaza), processes in least one employee, with a median of three jobs per start-up, Beirut are lengthier for receiving credit and funding (with West consistent with the evolving stage of an ecosystem, with start- Bank & Gaza), indicating greater difficulty (linked to procedures ups primarily in their early stage. or access to finance requirements) in obtaining funding for start-ups. The contrast with West Bank & Gaza’s ecosystem is Start-ups in Beirut face some bureaucratic hurdles in their life particularly striking: in Beirut, it takes twice as long for a start-up cycle (see Figure 2.2). Ventures can rapidly open a bank account to obtain credit and almost four times as long to receive funding.11 FIGURE 2.1 START-UP GROWTH IN BEIRUT, 2009-2015 100 90 Compound Annual Regional Growth Rate: 24% 80 70 # of Startups 60 50 40 30 20 10 0 2009 2010 2011 2012 2013 2014 2015 YEAR Note: Data shows tech start-up ventures as reported by founders of active start-ups during date of survey. Data of start-ups founded in year 2016 was not included in this figure because the survey was carried out in mid-2016, which was not comparable with all previous data from complete years. The Tech Start-up Ecosystem in Beirut 9 On average, start-ups in Beirut have 1.8 founders and Founders are predominantly male, and founders are twice as each founder has launched 1.3 start-ups. This is consistent likely to focus on business than technical functions, with 40 with other early to middle stage ecosystems, when serial percent doing both simultaneously (see Figure 2.3). entrepreneurs begin to feature. FIGURE 2.2 TIME TO COMPLETE PROCEDURAL TASKS IN LIFE CYCLE OF A START-UP ACROSS REGIONS COUNTRY Set up Bank Dar es Salaam 4 COUNTRY Account West Bank & Gaza 3 Dar es Salaam West Bank & Gaza Beirut Beirut 5 Become Dar es Salaam 22 Incorporated West Bank & Gaza 30 Beirut 28 Rent an Dar es Salaam 30 o ce West Bank & Gaza 15 Beirut 30 Hire an Dar es Salaam 30 Employee West Bank & Gaza 15 Beirut 30 Obtain Dar es Salaam 90 Credit West Bank & Gaza 15 Beirut 30 Obtain Dar es Salaam 125 Funding West Bank & Gaza 33 Beirut 120 MEDIAN DAYS 0 10 20 30 40 50 60 70 80 90 100 110 120 130 Note: Measurements refer to Dar es Salaam (Tanzania), Beirut (Lebanon) and West Bank & Gaza start-up ecosystems FIGURE 2.3 GENDER DISTRIBUTION AND JOB FUNCTIONS OF FOUNDERS Gender Distribution of Founders Job Functions of Founders GENDER JOB FUNCTION Male Business Female Technical Both (Business & Technical) 17.72% Other 18.82% 7.87% 40.16% 81.18% 34.25% 10 TECH START-UP ECOSYSTEM IN BEIRUT Founders are 29.8 years old on average at the time of founding Skills (see Figure 2.4).12 This puts Beirut on par with mid-stage ecosystems, such as Bogota, and mature ecosystems, such as Founders in Beirut are typically highly educated and have New York, which has an average age of founders in the late had some professional experience before they start their 20s and early 30s and confirms the maturing stage of the ventures. However, most of this professional experience ecosystem (see Figure 2.4). is in entry-level positions, which does not provide all the needed business acumen for start-up management. FIGURE 2.4 AVERAGE AGE OF FOUNDERS Educational levels are especially high among founders in the Average Age of Founders Across Various Regions context of Lebanon,13 with over 90 percent having a university degree and over 50 percent having a postgraduate degree (that is, masters, professional,14 or doctorate). Somewhat surprisingly, there is little difference in the educational REGION background of funded founders compared with the general 34 Dar es Salaam population of founders (see Figure 2.5), but perhaps this is 33 West Bank & Gaza to be expected given that the general cohort of founders in 32 Beirut 32.0 Beirut is highly educated and the percentage of founders with New York City graduate degrees is comparable with more mature ecosystems 31 Bogotá 30.6 such as Singapore or New York City (see Figure 2.6). 30 29.8 29 FIGURE 2.5 EDUCATIONAL LEVEL OF FOUNDERS Avg. Age 28 27.7 IN BEIRUT 27 26 Highest Degree Earned by Founders at Time of Founding 25 24 23.7 Doctorate Professional Dar es Salaam New York City BEIRUT 23 Master West Bank 22 Bachelor Bogotá 5.34% 3.82% Associate 21 High school 20 Age Distribution of Founders in Lebanon 18.32% AGE (BIN) 35 41.22% Below 20 32 33 20-24 30 25-29 30.53% 30-34 35-39 # of Fouders 25 40-44 45 and above 21 Highest Degree Earned by Funded Founders at Time of Founding 20 15 4.65% 15 18.60% 10 6 5 41.86% 2 1 0 Below 20 20-24 25-29 30-34 35-39 40-44 55-59 34.88% Note: Based on 110 responses (out of 255, 43 percent) that specified the age of founders. The Tech Start-up Ecosystem inBeirut 11 The majority of founders (51 percent) have a degree in ecosystems. In a much more mature ecosystem, such as New science, technology, engineering or mathematics (STEM), 14 York City, the experience of founders is higher, with 9.9 years percent had a degree in business, and 13 percent had both in 3.25 companies. STEM and business degrees.15 Fifty-two percent of STEM degrees were master’s or higher. Indeed, founders that have The largest category of previous function is analyst, with 29 successfully obtained funding in Beirut were slightly more percent of founders with such experience.16 While only 22 likely to have specialized in business (see Figure 2.7). percent of founders have experience in a managerial role, more than half of founders have had business experience, with The average founder in Beirut has 7.8 years of work slightly more than a quarter of founders experienced in both experience with no more than 2.7 companies. This work business and technical roles (see Figure 2.8). experience is similar to other emerging and middle stage FIGURE 2.7 EDUCATIONAL EXPERIENCE OF FIGURE 2.8 PREVIOUS FUNCTION AND ROLE FOUNDERS IN BEIRUT TYPE OF FOUNDERS IN BEIRUT Educational Experience of Founders at Time of Founding Previous Functions of Founders at Time of Founding 30% 12.90% 30% 25% 13.71% 22% Precentage of Founding Instances 20% 50.81% 22.58% 15% MAJOR Business 10% Stem 10% 8% Both (Stem & Business) Other 5% 3% 4% Educational Experience of Funded Founders at Time of Founding 1% 6.82% 0% Intern Freelancing Analyst Manager Director C-Level Advisor Previous Role-Type of Founders 20.45% Business Technical Both Other 50.00% 22.73% 27.50% 33.00% 13.50% 26.00% 12 TECH START-UP ECOSYSTEM IN BEIRUT FIGURE 2.6 HIGHEST DEGREE EARNED BY FOUNDERS AT TIME OF FOUNDING Region Doctorate Beirut New York City Singapore West Bank & Gaza Professional Beirut New York City Singapore West Bank & Gaza Master Beirut New York City Singapore West Bank & Gaza Bachelor Beirut New York City Singapore West Bank & Gaza Associate Beirut West Bank & Gaza Highschool Beirut Singapore West Bank & Gaza % of Founders 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% Supporting Infrastructure BOX 2.1 ACCELERATORS AND INCUBATORS for Entrepreneurship Consistent with the ecosystem’s maturity, the supporting infrastructure for start-ups in Beirut’s Although the dataset and analysis does not differentiate ecosystem is passed its initial growth phase but still between incubators and accelerators, there is a difference maturing. The two main elements of the supporting in their definition: infrastructure analyzed are accelerators and incubators, terms that are used interchangeably in this report (see Accelerators support entrepreneurs and start-ups in the Box 2.1), and mentors. early stages of development and they are often comprised of the following features: (i) a highly competitive and open application process for entrepreneurs, (ii) provision of small Accelerators amounts of seed investment, (iii) a focus on small teams Accelerators support start-up growth by providing rather than individual founders, (iv) intensive support for skills and networks of connections. Previous research a limited period of time (usually 3-6 months), with active mentorship and networking, and (v) collaborative work from the World Bank shows that accelerators have a among start-ups through cohort or classes of start-ups. key role in supporting the community of start-ups that generate the ecosystem, providing the necessary social Incubators are spaces that support start-ups by providing connectivity among entrepreneurs and other ecosystem office space and administrative support services. The stakeholders (Mulas, Minges, and Applebaum 2015). most typical services are legal, recruiting, IT, accounting, Research from the Aspen Institute also suggests that public relations and pooled buying programs. In addition, accelerators have a positive impact in supporting early incubators may also provide coaching, mentorship, and stage ventures by providing access to a network of help with access to funding on an informal basis. Start- mentors and capacity building, particularly regarding ups pay rent (which is usually below the market rate) for business skills and acumen (I-DEV International 2014; the office space and normally a time limit is not set for Baird, Bowles, and Lall 2013; and Roberts, Lall, and start-ups staying in the incubator (the average length of Baird. 2016). stay ranges widely from 18 months to five years). Some incubator providers may ask for a share of future profits or There are 10 unique accelerator programs that have require a minority stake in the start-up as a prerequisite for been attended by start-ups in Beirut’s ecosystem (see access to the incubator. Figure 2.9).17 These accelerators have accelerated a Source: Mulas, Minges, and Applebaum 2015. total of 28 start-ups. Three of these accelerators are international (i.e., not based in Beirut), showing the The Tech Start-up Ecosystem inBeirut 13 connection of Beirut’s ecosystem to external resources in other suggests that accelerators in Beirut are not having a meaningful ecosystems (see Community section). Accelerator programs impact in improving the quality of start-ups. In fact, of the were very limited in Beirut until 2015, when Circular 331 funding 55 start-ups that were funded, more (39) were unaccelerated resulted in a substantial increase of programs for Beirut’s start- than accelerated (16). Half of the accelerated start-ups ups. Three out of four accelerator programs located in Beirut received funding in the same year that they participated in the (UK-Lebanon Tech Hub, Altcity Bootcamp and Speed at BDD) acceleration program received funding from Circular 331. Using these numbers, the effectiveness of accelerators in Beirut’s ecosystem may be estimated by calculating a ratio of quality, as Two-thirds of accelerated start-ups were funded, while only represented by the acceleration multiplier of investment (0.95) one-third of unaccelerated start-ups were funded, that is, start- to quantity, as represented by the acceleration multiplier for ups supported by accelerators are twice as likely to receive funding probability (2), which results in a ratio of 0.48. This is funding as those that do not. This means that the acceleration consistent with the quality to quantity ratio from other middle- multiplier for funding probability is 2.18 This is normal stage ecosystems, such as Cairo, which also has a ratio of less for nascent ecosystems, where accelerators are crucial in than 1, while mature ecosystems such as New York City will increasing the quantity of funded start-ups in the ecosystem’s have a ratio of more than 1 (see Figure 2.10). pipeline. Preliminary analysis of similar measurements in more mature ecosystems, such as New York City and Santiago, These results support the conclusion that it is difficult to access shows that this ratio reduces to become closer to 1 as the venture funding without attending an accelerator program, funding ecosystem matures and accelerators are no longer the even though these programs do not meaningfully improve gatekeepers to investment networks. the quality of start-ups they host compared to unaccelerated start-ups. Nevertheless, half of accelerated start-ups received Of start-ups that were funded, accelerated start-ups received additional investment in the years after graduating, indicating an average of 1.38 investments, while unaccelerated start-ups that there is plenty of capital available in the ecosystem received an average of 1.45 investments,19 an acceleration regardless of the quality of start-ups in the pipeline. This is multiplier of investment 0.95.20 If we interpret the number of unsurprising given Circular 331’s role in catalyzing investment investments a start-up receives as a signal of its quality, this availability (see Program in Focus). FIGURE 2.9 ACCELERATOR PROGRAMS FIGURE 2.10 ACCELERATOR’S QUALITY/ ATTENDED BY START-UPS IN BEIRUT QUANTITY RATIO (SELECTED CITIES) 1.40% 16 ACCELERATOR NAME 1.28 Altcity Bootcamp 1.26 14 Bader 1.20% Berytech 4 Blackbox (USA) Cap Omega (France) 12 1.04 CME O shore 1.00% Hax (China/USA) 0.95 Seeqnce 10 Speed at BDD UK-Lebanon Teach Hub 3 0.80% # of Startups 1 8 1 0.60 0.60% 1 4 0.55 6 0.47 0.40% 4 4 0.36 West Bank & Gaza New York City Dar es Salaam 0.20% 2 4 Singapore BEIRUT Santiago Medellin 1 2 Cairo 1 1 1 0 0.00% 2011 2012 2013 2015 2016 Note: no acceleration occurrences were reported in 2014. 14 TECH START-UP ECOSYSTEM IN BEIRUT Mentors Berytech, and the IFP (International Fairs and Promotions) Group. The top groups that founders in Beirut claimed to Mentorship is a knowledge transfer mechanism for be members of were Endeavour, Tripoli Entrepreneurs Club, entrepreneurs to acquire business acumen, understand the and Bootcamp Sal. unspoken rules of start-up challenges, and access networks of talent, knowledge, and resources. Mentors need to be knowledgeable and experienced. A study for the U.K. Investment government found that the most important characteristic of a mentor is proven business success in the area of work and The value of investors extends beyond the money they network of contacts (BMG Research and Galli 2013). Mentorship provide. Early stage investors are often valued both for their relationships were found to often develop informally through networks and for their experience and subject area expertise, a preexisting relationship. For example, acceleration and hence the phrase “smart money.” This report considered all incubation programs typically assign start-ups formal mentors organizations that invest in high-growth start-ups, venture to assist them for a designated period of time. Research from capital firms, angel investors, and other individuals. Endeavor Insight shows that top performing start-ups have much higher support from mentors.21 Forty-three investors were recorded in Beirut. About half of these were venture capital firms and the other half angel In Beirut, 20 percent of founders received mentorship from investors. They made a total of 58 investments in 35 start- a total of 53 mentors. This mentorship tended to occur on a ups, and all investments were identified as equity financing. one-to-one basis and generally lasted about one year. Unlike The median number of start-ups invested in per investor was other ecosystems, such as Tanzania’s, Beirut does not currently 1, but with several notable outliers. Speed@BDD invested in possess “supermentors” associated with accelerators to mentor nine different start-ups, Berrytech in four, and Middle East batches of entrepreneurs, indicating an opportunity to catalyze Venture Partners (MEVP) in three. the mentorship ecosystem by creating more formalized programs. Although there is a significant fall in the number of start- ups funded in their first year compared to subsequent Groups and Events years, a substantial number of start-ups receive funding in the second year and beyond. This indicates that the In addition to accelerators and mentors, groups and events are ecosystem has developed beyond the nascent stage and also a mechanism for support and knowledge dissemination. is maturing into a middle stage ecosystem. Furthermore, BDL Accelerate is a key event in the ecosystem, bringing the median investment amounts in later years of existence together both foreign and domestic players. Other important is encouraging in that they are exceeding $100,000 (see events attended by founders were hosted by The World Bank,22 Figure 2.11). FIGURE 2.11 MEDIAN INVESTMENT AMOUNT 40 30 #of startups receiving funding 25 20 15 10 5 0 $40k $30k $400k $150k $350k $850k $1,300k $2,000k $1,000k $3,100k $6,500k $500k Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 9 Year 10 Year 12 Year 13 Year 14 Startup year of existence when receiving funding Notes: Only three start-ups received investment more than once; No funding was reported by startups in year 11 of existence. The Tech Start-up Ecosystem inBeirut 15 Community Start-up ecosystems operate as communities, where many of whom are well-connected internationally and often knowledge spillovers and access to resources flows through have dual or triple nationalities with other countries. This a network of embedded connections (see Figure 2.12). The is a unique feature of the Beirut ecosystem and of Lebanese tighter and more connected an ecosystem, the more efficient society in general, which has strong international connections is the flow of knowledge and access to resources. The less through its diaspora and transient populations. connected it is, the less effective is the ecosystem in spotting talent and nurturing potential ventures into successful The community of Beirut’s ecosystem is consistent with its start-ups. nascent stage with low density and a limited number of clusters.23 Higher density and more clusters allow entrepreneurs Connectivity matters because the success of start-ups is to connect to knowledge and resources through other impacted by their network and access to other ecosystem actors in the ecosystem. The less dense the ecosystem is, the stakeholders and their networks (Mulas, Minges, and more difficult it is for a founder to find their way to mentors, Applebaum 2015). In Beirut, the main cluster of start-up investors or other relevant knowledge or resources required activity revolves around BDL Accelerate. The Beirut ecosystem for their venture. Clusters serve as multipliers of density, has also begun to pull in more global networks, such as that helping founders leapfrog orders of connection (for example, from Startup Weekend or the University of Pennsylvania (see connections that in other case are a 5th or 7th order connections, Figure 2.13)). that is, the founder is connected through 5 or 7 connections to the target person, become a 2nd or 3rd order connection, where The most prominent network in Beirut is BDL Accelerate, the founder is connected through two or three people to the which is consistent with the role the Bank du Liban has played target connection). in the ecosystem through Circular 331 and complementary initiatives. Through first and second-order connections, the When comparing the density and clusters of the community Beirut ecosystem connects with international clusters such as with those of more advanced ecosystems (see Figure 2.14), Startup Weekend and the University of Pennsylvania, which Beirut features low density with one cluster in formation and presents the potential to access financial and intellectual some potential sub-clusters. Cairo or Medellin have already resources not available in the domestic market. This is evolved into highly connected clusters driving connectivity in consistent with the background of Beirut startup founders, the ecosystem. FIGURE 2.12 GROWTH OF FOUNDERS AND CONNECTIONS IN BEIRUT YEAR 550 500 EDGES LABEL acquired 450 accelerated/incubated mentored invested in 400 founded 350 Count Cumulative 300 250 200 150 100 50 0 2009 2010 2011 2012 2013 2014 2015 2016 16 TECH START-UP ECOSYSTEM IN BEIRUT FIGURE 2.13 CONNECTIONS IN BEIRUT’S TECH ECOSYSTEM BDL ACCELERATE UNIVERSITY OF PENNSYLVANIA (International connection STARTUP WEEKEND to outside cluster) (International connection to outside cluster) Note: This graph highlights BDL Accelerate (purple), Startup Weekend (yellow), and the University of Pennsylvania (red). The Tech Start-up Ecosystem inBeirut 17 FIGURE 2.14 PROGRESSION OF ECOSYSTEM CONNECTIVITY MEDELLIN Advancing CAIRO BEIRUT Nascent Note: Network graphs were created by taking all the people in the ecosystem and creating edges to all other people they were directly or indirectly connected to in order to exaggerate the effects of clusters for illustration purposes. Because this treatment visualizes the influence of edges seen in Figure 2.13 by counting them more than once, the density of the clusters and the overall graph will appear different from Figure 2.13, where each edge is represented only once, even though the underlying data is the same. 18 TECH START-UP ECOSYSTEM IN BEIRUT Start-up Success Factors Start-up success is difficult to determine as tech ventures by a comment from a survey respondent who stated that, operate in a fast-paced environment under continuous change. “I wish it wasn’t centralized in the limited number of VCs and To analyze factors that have determined start-up success, this banks that are all in one bundle and interconnected. Some analysis identifies two moments in the growth of start-ups. We competition between the VCs and the programs would push consider “short-term success” as when a venture obtains funding them to provide better services.” from an outside investor and “long-term success” as when a venture hires employees consistently (this assumes continuous Factors of Long-term Success growth as the talent-knowledge assets of the start-up grow). The most significant factors for long-term success (that is, Factors of Short-term Success hiring employees over time and continuing to do so) are: a) obtaining an investment, and b) having additional work The most significant factor for raising funding in Beirut is to experience. Participation in acceleration programs and be highly connected to other stakeholders. Given the early mentorship relationships have no significant effect on long- stage of the ecosystems community (see Community section), term hiring probability. Moreover, serial entrepreneurs in there are very few clusters available for start-ups and very few Beirut are also less likely to hire consistently, suggesting that start-ups inside these clusters. Those start-ups that are inside start-ups have difficulties in scaling up, staying in a continuous these clusters have access to investors and can leverage those early stage cycle. There is no significant impact of educational connections to obtain their first funding. Once start-ups have factors on job creation. become part of one of these clusters and have obtained funding, they increase substantially their network of investors, reinforcing These results suggest that the support infrastructure of the their positioning in relation to future investments. Older start- ecosystem (for example, accelerators, mentorship) is not ups (those that have survived over time) are also more likely to supporting scale-up of start-ups adequately, either because raise funding. quality is low or because their impacts are not yet visible. However, the low probability that serial entrepreneurs will These results suggest that there appears to be a main cluster mature their businesses suggests that there may be a lack of in the investment ecosystem, and the majority of activity is general skills available in the ecosystem to transform ventures occurring in this maturing cluster. This hypothesis is supported into sustainable businesses. Hiring employees over time FIGURE 2.15 SHORT-TERM SUCCESS SHORT TERM LONG TERM SUCCESS SUCCESS BE HIGHLY CONNECTED OBTAINING AN INVESTMENT TO OTHER STAKEHOLDERS HAVING ADDITIONAL WORK EXPERIENCE RAISING FUNDING HIRING EMPLOYEES OVER TIME The Tech Start-up Ecosystem inBeirut 19 Gap Analysis and Policy Recommendations Summary of Gap Analysis and Stage of Ecosystem The evidence from our analysis points to the conclusion that the start-up ecosystem in Beirut is an early advancing (middle stage) ecosystem (see Table 3.1). TABLE 3.1 DEVELOPMENT STAGE OF ECOSYSTEM (PER AREA AND OVERALL) Stage Ecosystem Area Nascent Advancing Mature Community Skills Supporting Infrastructure Investment Constraints OVERALL There are multiple gaps that can be addressed through policy ventures. There is a lack of experienced entrepreneurs and, action and support. The community is immature with only one combined with the limited availability of quality mentorship, functioning cluster, and as such start-ups are constrained in the these factors reduce the availability of young talent to complete diversity of resources they can access. Events and community the cycle of successful ventures. There is a potential bubble of building appear to have been catalyzed by programs of the start-up ventures being funded by the abundance of available Banque Du Liban (BDL), the Central Bank of Lebanon (BDL seed funding, which may be distorting the ability of the Accelerate and Circular 331), and donor and government action ecosystem to select the best start-ups and generate success still drives a large number of activities in the ecosystem. There stories that are sustainable over time.24 Finally, the sustainability is an incipient number of support programs for entrepreneurs of the current level of access to funding, which is fueling the (for example, accelerators) but they do not yet seem to be ecosystem, is not ensured if Circular 331 is phased out (see providing services of sufficient quality to support sustainable Box 3.1). 20 TECH START-UP ECOSYSTEM IN BEIRUT Policy Recommendations Table 3.2 summarizes the policy recommendations for the Beirut ecosystem, based on the key gaps and constraints identified. TABLE 3.2 POLICY RECOMMENDATIONS Ecosystem Area Policy Recommendation Objective Expand cluster connectivity, coordinate private Strengthen coordination mechanism and and public action and increase density of ecosystem support program with stakeholders connections among all stakeholders Increase absorption by ecosystem of international Expand ecosystem community to traditional talent connectivity, connect with domestic industries and absorb international connectivity Community corporate nontech sectors Expand practical education in university and Address practical business acumen gaps, through rapid skills training programs and include provide pipeline of talent for start-up scale-up public education programs and include lower income/educated population Skills Increase capacity building of mentors in accelerators and attraction of international talent Address lack of sufficient number of quality (as mentors, entrepreneurs or capacity builders) to mentors and strengthen support services Support the ecosystem Infrastructure Assess quality of start-up funding and prepare Address potential bubble of start-ups and phase out of Circular 331 introducing private distortion of investment market, and ensure sector funding sustainability of funding Investment Address processes constraints (for example, access Reduce constraints for start-ups incorporation to loans and funding) and operationalization Constraints These recommendations only address short-term actions to continuous consultation processes and support programs for support the Beirut ecosystem. Policy makers should constantly public-private partnerships (PPPs) (ITIF 2017). monitor the ecosystem (which can be done through the coordination mechanism once in place) and iterate the policy Beirut has multiple intermediaries and stakeholders with active approaches as needed and address new upcoming gaps or programs (including Endeavor Lebanon, Berytech, Wamda, growth hurdles as they arise. As the ecosystem grows and Lebanon UK Tech Hub, Alt City, and Beirut Digital District’s evolves into more maturity, new needs will emerge and other (BDD) Speed@BDD, among others). However, these activities specific policies may be more applicable. are not connecting different clusters of start-up founders as well as they could, suggesting a lack of in-depth connection. Better and wider coordination of stakeholders beyond BDL Accelerate and codevelopment of day-to-day events from 1. Community multiple stakeholders (which can be catalyzed through public Strengthen coordination mechanisms and ecosystem support) could help in developing more in-depth connectivity support programs for stakeholders. Ecosystem coordination between clusters, programs, and mentors. increases the effectiveness of community-building events and connectivity among clusters. Successful ecosystems, such as Increase absorption by the ecosystem of international in Buenos Aires, or New York, have created different types of talent, and improve connectivity with domestic corporate coordination mechanisms among stakeholders led by public nontech sectors. BDL Accelerate has been successful in policy actors (for example, municipal or government innovation promoting Beirut’s ecosystem and elevating Lebanon’s brand agencies). These vary from stakeholders’ roundtables to as an incipient start-up hub in the region. It has also attracted Gap Analysis and Policy Recommendations 21 international talent to Beirut, providing a window to connect BOX 3.1 WHAT IS CIRCULAR 331? the ecosystem to international talent beyond the already well-developed diaspora network. If they can be absorbed by the ecosystem, international talent and know-how is key to developing the ecosystem beyond its growing stage. Diaspora In August 2013, BDL announced Circular 331, networks may provide an initial connection with international injecting the potential of $400 million into talent, but if diaspora talent is not retained, that connection the Lebanese enterprise market. Circular 331 will not provide a solid foundation for the ecosystem. guarantees 75 percent of the banks’ investments in the knowledge economy through direct Israel’s ecosystem developed specific programs to attract start-up equity investment or indirect start-up international talent (through multinationals and start- support entities. up support programs) and develop the absorption of the ecosystem (through PPP programs connecting these talent Local banks receive a seven-year interest- and multinationals with domestic start-ups). This resulted free credit from BDL, which can be invested in in a close link between the Tel Aviv ecosystem and New York treasury bonds with an interest rate of 7 percent. and Boston, fueling investment from VCs and talent exchange In return, the bank commits to investing in the between these locations (indeed, many successful Israeli start- knowledge economy. Local banks can invest up ups have active operational offices in Tel Aviv and the other to 3 percent of their capital in start-up support US ecosystems) (ITIF 2017; and Getz, Daphne, and Goldberg entities, funds, or directly into start-ups. BDL 2016). Beirut could leverage BDL Accelerate’s connections guarantees 75 percent of the investment, and its diaspora networks to further codevelop programs to derisking it by mitigating the potential losses and support the ecosystem with international talent (for example, reducing them to a mere 25 percent. Circular 331 accelerator or skills training programs). This could be catalyzed is designed to diminish risk for the conservative with public support programs, similar to the way in which local banks and does so by dictating the banks’ Israel and other ecosystems, such as Santiago in Chile or portfolio diversification. A bank can invest up to Seoul in Korea, have done (see policy recommendations on 10 percent (of its 3 percent) in any one start-up, Skills below). thus spreading the risk. BDL takes on 75 percent of the risk and only 50 percent of any profit made, making it attractive for local banks. Finally, Beirut ecosystem could also expand beyond the tech sector, to include traditional industries. This was one of the key Source: http://2015.bdlaccelerate.com/everything-you- features of New York ecosystem success, for instance, which need-to-know-about-bdlcircular-331/ resulted in specialized start-ups with in-depth knowledge of industry niches and improved competitiveness of traditional sectors in the city, such as finance, media, advertising, and fashion (Mulas and Gastelu-Iturri 2016). Policy actions can be applied to catalyze industry-start-up innovation through open innovation and service codevelopment processes. For example, Paris has supported the corporate-start-up service codevelopment process through initiatives such as Data City Paris.25 In New York, the city used PPPs to catalyze sector- specific accelerators for media and fashion, among others (Mulas and Gastelu-Iturri 2016). to address skills gaps ranging from entrepreneurship and 2. Skills business skills to specific coding and technical skills (see Box 3.2) (Mulas and Gastelu-Iturri 2016). Public policy programs Expand practical education in universities and through can also help provide community spaces, and accelerators can rapid skills training and public education programs. Public also be supported to enhance their capacity to provide higher policies can catalyze and support both practical education quality mentorship and training to start-up ventures with programs to address the skills gaps in tech start-up ecosystems. potential through their programs (in the case of accelerators) Coding bootcamp methodologies (a rapid skills training of related activities. methodology), are nascent in Beirut,26 but they have proven successful in quickly assessing market gaps and demands in Additionally, public policies can catalyze the introduction of tech start-up ecosystems. Moreover, these programs can serve practical education programs for university students and address to include the less-educated population into the ecosystem, part of the gap in business acumen. Initiatives, such as Demola by providing a basic set of skills connected to the ecosystem’s in Tampere, Finland28 and Jacobs Technion-Cornell Institute29 needs. For instance, New York City’s initiative to support rapid in New York, were created to address this specific gap in their skills training programs resulted in General Assembly, one of ecosystems. In both cases, a practical project-based education is the largest providers of bootcamps worldwide, which serves added for students to learning-by-doing with business. 22 TECH START-UP ECOSYSTEM IN BEIRUT 3. Support Infrastructure the ecosystem through mentorship (for example, through simple mentorship or angel investment) the ecosystem will Increase capability of mentors in accelerators and attraction not be sustainable over time. of international talent (as mentors, entrepreneurs, or capacity builders) to the ecosystem. Support can also be Second, Circular 331 may be crowding out private sector led provided to increase the capability of mentors (for example, investment. To phase out Circular 331 in an orderly manner, through training and capacity-building programs). However, private sources of investment should be introduced and local capacity programs have limited impact beyond nascent crowed in. BDL should prepare for sustainability of the early stages of ecosystems. This is because there is a general lack stage funding and reducing the crowd out effect in private of practical experience and knowledge among mentors and sector led investment. trainers since there have not been many cases of successful ventures that have been proven internationally (and whose founders are willing to be mentors or angel investors). 5. Constraints Address processes constraints. The transaction constraints A way to address this lack of experience and knowledge is highlighted by entrepreneurs puts Beirut on a par with very to attract international talent with such practical acumen. early stage ecosystems, such as Tanzania, making start-up There are several support programs that can achieve this creation burdensome for founders and diverting attention and goal, ranging from events that gather international talent resources at an early stage. In a similar advancing ecosystem, and connect it to the ecosystem, to more structural programs. such as in West Bank & Gaza, renting office space and hiring While BDL Accelerate has led to a broader connection of the employees can be done in half the time, credit loan processes ecosystem’s intermediaries (for example, community spaces, are six times faster, and funding processes three times faster. accelerators, and so on) with international talent and partners, Policy reforms can streamline these processes by establishing these partnerships do not seem to have resulted in long-term specific conditions for start-ups processes (for example, mentorship relationships with entrepreneurs. reducing requirements). Santiago, Chile, followed a more structured approach, creating a program (Startup Chile)30 to attract international talent to the BOX 3.2 WHAT ARE CODING BOOTCAMPS? ecosystem. The program is in essence an acceleration-funded program for high-skilled international talent to conduct their ventures’ initial stage in Chile. The program introduced specific activities to ensure knowledge spillovers between international Coding bootcamps are intensive short-term talent and domestic entrepreneurs. For instance, international programs, typically lasting three to six months, entrepreneurs were collocated in a coworking space with designed to train participants in programming domestic entrepreneurs with a requirement that they provide skills to make them immediately employable capacity building and workshops on entrepreneurship and in entry-level tech positions.27 In essence, they technical skills for domestic entrepreneurs. This mechanism has combine characteristics of traditional vocational been emulated by other ecosystems, including the K-Startup training programs with the intensity of military Grand Challenge in Korea.31 bootcamps for new recruits, intermingling soft and tech skills learning in an intense manner, in what could be referred to as “skills accelerators.” Coding 4. Investment bootcamps follow a structured process with three main characteristic features: 1) intense rapid skills Assess quality of start-up funding and prepare phase training; 2) an experiential learning approach; and out of Circular 331 introducing private sector funding. 3) curricula based on, and continuously adapting Circular 331 has provided an influx of seed investment to to, industry’s demand. the ecosystem resulting in wide availability of funding for early stage ventures. Though some procedural constraints Although the bootcamps methodology has still remain for access to finance (see Constraints policy primarily focused on coding skills, it has been also recommendations), overall Beirut seems to have sufficient adapted for business and entrepreneurial skills as funding available for start-ups. However, two concerns are well as other technical skills. Usually, bootcamp present. First, there are signs that the amount of funding programs embed “life skills” in their curriculum, availability may be larger than the capacity of the ecosystem to enabling their graduates to be competitive generate sufficient quality start-ups, resulting in the financing irrespective of the industry in which they choose of nonsustainable ventures (see Program in Focus). This to work, for example, the ability to master new may be corrected if policy programs increase support to the knowledge quickly and efficiently, effectively ecosystem for producing a higher number of quality start-ups work in a team, meet tight deadlines, and so on. (see overall policy recommendations). This type of start-up is Evidently, these “life skills” belong to the subset of future-proof soft skills (World Bank.2017). key for sustainability, as the research from Endeavor Insight in New York shows (Morris 2015). If the ecosystem cannot produce enough success stories whose founders reinvest in Gap Analysis and Policy Recommendations 23 Program in Focus: Circular 331 In August 2013, BDL announced Circular 331, injecting the risk for the conservative local banks and does so by dictating potential of $400 million into the Lebanese enterprise market. 32 the banks’ portfolio diversification. A bank can invest up to 10 Circular 331 guarantees 75 percent of the banks’ investments percent (of its 3 percent) in any one start-up, thus spreading in the knowledge economy through direct start-up equity the risk. BDL takes on 75 percent of the risk and only 50 investment or indirect start-up support entities. percent of any profit made, making it attractive for local banks. Local banks receive a seven-year interest-free credit from BDL, which can be invested in treasury bonds with an interest rate Responses to our survey suggest that Circular 331 has of 7 percent. In return, the bank commits to investing in the increased access to funding, but not necessarily smart knowledge economy. Local banks can invest up to 3 percent funding, and that the funding has not increased education of their capital in start-up support entities, funds, or directly and training opportunities. The majority of respondents, into start-ups. BDL guarantees 75 percent of the investment, however, agreed that overall, the program has been helpful derisking it by mitigating the potential losses and reducing in catalyzing the investment and support ecosystems (see them to a mere 25 percent. Circular 331 is designed to diminish Figures B.1 and B.2). FIGURE 2.16 CIRCULAR 331 EFFECTIVENESS IN VARIOUS AREAS C331 3.6 3.5 3.4 3.3 Average Score (1-5) 3.2 3.1 3.0 2.9 2.8 Boosting Startup Developing Support Changing Fostering Education Providing Ecosystem Creation Systems Entrepreneurial Culture & Training Smart Money 24 TECH START-UP ECOSYSTEM IN BEIRUT Entrepreneurs surveyed seemed to suggest that the program FIGURE2.17 EFFECT OF CIRCULAR 331 ON may benefit financial institutions more than start-ups RAISING FUNDING AND DEVELOPING THE themselves, and there is a risk in providing too much funding without a quality pipeline of start-ups. SUPPORT SYSTEM These results should be interpreted somewhat cautiously Easier to Raise Funding? given that many survey outreach partners benefitted from Circular 331, and as such they may have been biased in reaching out to respondents who were more likely to have Yes benefitted from the program. No “Regarding Circular 331, I believe it is a positive contribution to fostering the growth of the tech sector in Lebanon. On the other hand, the risk is to 43% fund start-ups without valid business prospects, especially if the incentives of the different parties are not aligned properly.” “Can’t find tech employees anymore, they are all entrepreneurs funded by 331.” “Circular 331 should be extended beyond banks to 57% any investor or any start-up founder that is willing to move its operations to Lebanon. Banks do not need or deserve guarantees as they bring no additional value to the table and have more money than they Easier to Develop Support System? know what to do with.” Yes No 26% 74% Program in Focus: Circular 331 25 References Baird, Ross, Lily Bowles, and Saurabh Lall. 2013. Bridging the Isaksson, Anders, Thiam Hee Ng, and Ghislain Robyn. “Pioneer Gap”: The Role of Accelerators in Launching High-Impact 2005. Productivity in Developing Countries: Trends and Enterprises. Aspen Network of Development Entrepreneurs Policies. Vienna: United Nations Industrial Development and Village Capital. https://assets.aspeninstitute.org/ Organization. http://www.unido.org/fileadmin/ content/uploads/files/content/docs/ande/Bridging%20 user_media/Publications/Pub_free/Productivity_in_ the%20Pioneer%20Gap%20The%20Role%20of%20 developing_countries_trends_and_policies.pdf. Accelerators%20in%20Launching%20High%20Impact%20 Enterprises%20.pdf. ITIF (Information Technology and Innovation Foundation).2017. Learning from Leading Startup BMG Research and Leandro Galli. 2013. Demand for Ecosystems. Washington, DC: ITIF. https://itif.org/ Mentoring Among SMEs. BIS Research Paper Number 158. events/2017/05/02/learning-worlds-leading-startup- London: Department for Business Skills and Innovation. ecosystems. https://www.gov.uk/government/uploads/system/uploads/ attachment_data/file/263226/demand_for_mentoring_ Morris, Rhett. 2015. “Mentors are the Secret Weapons among_SMEs.pdf. of Successful Startups.” TechCrunch Network, March 22. https://techcrunch.com/2015/03/22/mentors-are-the- Comin, Diego, and Bart Hobijn. 2010. “An Exploration secret-weapons-of-successful-startups/. of Technology Diffusion.” American Economic Review 100 (December): 2031–2059. https://www.aeaweb.org/ Mulas, Victor, and Mikel Gastelu-Iturri. 2016. New York articles?id=10.1257/aer.100.5.2031. City: Transforming a City into a Tech Innovation Leader. Washington, DC: World Bank. https://openknowledge. Comin, Diego, and Mestieri Marti. 2013. “Technology worldbank.org/handle/10986/25753. Diffusion: Measurement, Causes and Consequences.” In Handbook of Economic Growth, Volume 2, edited by Philippe Mulas, Victor, Michael, Minges, and Hallie Rocklin Aghion and Steven Durlauf, 565-622. Amsterdam: North Applebaum. 2015. Boosting Tech Innovation Ecosystems Holland. in Cities: A Framework for Growth and Sustainability of Urban Tech Innovation Ecosystems. Washington, DC: World Getz, Daphne, and Itzhak Goldberg. 2016. Best Practices Bank Group. https://openknowledge.worldbank.org/ and Lessons Learned in ICT Sector Innovation: A Case Study of Israel. Background Paper for the Word Development handle/10986/23029. Report 2016. Washington, DC: World Bank. http://pubdocs. worldbank.org/en/868791452529898941/WDR16-BP-ICT- OECD (Organization for Economic Co-operation and Sector-Innovation-Israel-Getz.pdf. Development). 2007. Eurostat-OECD Manual on Business Demography Statistics. Paris: OECD. http://www.oecd.org/ I-DEV International. 2014. Measuring Value Created by Impact std/39974460.pdf. Incubators and Accelerators. I-DEV International, in conjunction with the Aspen Network of Development Entrepreneurs World Bank. 2017. Coding Bootcamps: Building Future- (ANDE) and Agora Partnerships. https://assets.aspeninstitute. Proof Skills through Rapid Skills Training. Washington, org/content/uploads/files/content/docs/resources/ DC: World Bank. http://documents.worldbank.org/ ANDE%20I-DEV%20INCUBATOR%20REPORT%2011-21- curated/en/795011502799283894/Coding-bootcamps- 14%20FINAL%20FOR%20DISTRIBUTION.pdf. buildingfuture-proof-skills-through-rapid-skills-training. 26 TECH START-UP ECOSYSTEM IN BEIRUT Appendix A Survey Methodology and Analysis Methodology Additional questions were included regarding Circular 331: 1. On a scale of 1 to 5, how effective do you think Circular Survey Questions 331 was in developing support systems (incubators, The survey of start-up founders used the standard questionnaire accelerators, and so on)? developed under the Global Entrepreneurship Research Network 2. On a scale of 1 to 5, how effective do you think Circular (GERN) with some additional questions to understand specific 331 was in providing smart money to the ecosystem? constraints faced by start-ups. The GERN-developed standard 3. On a scale of 1 to 5, how effective do you think Circular questionnaire includes the following sections: 331 was in boosting start-up creation? 4. On a scale of 1 to 5, how effective do you think Circular 1. Educational history (including vocational, bootcamps, and 331 was in fostering education and training? certificate programs) 5. On a scale of 1 to 5, how effective do you think Circular 2. Employment history 331 was in changing the entrepreneurial culture? 3. Founding history (serial entrepreneurship) 6. Did Circular 331 make it easier for you to raise 4. Support programs (for example, acceleration, incubation, funding? and so on) history 7. Did Circular 331 make it easier for you to develop 5. Connections with mentors and mentees a support system (that is, attend an incubator or 6. Connections with investors (angel and institutional) accelerator)? In addition to this information, the survey captured the geographic Outreach Strategy location of start-ups and intermediaries (for the geographic analysis) and included the following standard questions: Entrepreneurs filled in an online survey available at http://survey.techecosystems.org/, which was disseminated 7. Incorporation: On average, how many days did it, or would through partnership with local organizations by email it, take you to incorporate a new start-up? and word of mouth. These partners played a crucial role 8. Funding: On average, how many days did it, or would it, take in not only identifying start-ups, but also connecting and you to set up a bank account for your start-up? introducing the study team and the survey to start-ups. 9. Credit: On average, how many days did it, or would it, take you to get a line of credit for your start-up? Partners included Endeavor Lebanon, ArabNet, Berytech, 10. Funding: On average, how many days did it, or would it, take USEK community, SETT (Startup Ecosystem Think Tank), you to raise a round of equity funding? UK Lebanon Tech Hub, Wamda, AltCity, Bader, Lebanon 11. Hiring: On average, how many days did it, or would it, take for Entrepreneurs (LFE), Lebanon Knowledge Network, you to hire an employee, from job posting to employee MIT Enterprise Forum of Pan Arab Region, Balamand start? University, BIAT Business Incubation Association, TEC (Tripoli 12. Office space: On average, how many days did, it or would it, Entrepreneurs Club), Lebanon Internet Society, Ministry of take you to obtain office space for your start-up? Economy and Trade, Smart ESA, South Accelerator, Startup 13. Exit: On average, how many days did it, or would it, take you Megaphone, and the American University of Beirut (AUB). to exit your start-up? Program in Focus: Circular 331 27 Data Pipeline information was recorded from multiple sources, in which case the source closest to the entity (that is, the founder) Survey data used in this report originated with this custom was privileged. In addition, respondents were expected to survey and was initially retrieved in raw JSON (JavaScript accurately select names of entities already in the database Object Notation) format. Data was converted to CSV (commas through an autofill mechanism in order to properly attribute separated values), combined with raw data from additional new information to existing entries. Although a machine sources, then cleaned of outliers and testing data. Nodes learning driven data deduplication process was employed without location data, and locations without geocodes, were before analysis, this may not have resolved every duplicated passed through the Google Maps application programming entry, and such fuzziness may affect exact numbers in the interface (API) in order to obtain standardized location data social network analysis. wherever possible. This new dataset was deduplicated using a process that marked similarities between names, email The sample is highly influenced by survivorship bias. addresses, URLs, and dates. Entities that were determined Entrepreneurs that are so successful that they leave the to be likely duplicates were then merged, maintaining all ecosystem were not captured. For example, the lack of existing data and privileging more recent data in the event strong evidence for serial entrepreneurs in Beirut may of conflict. Duplicate edges resulting from this process were suggest that, once they have had one successful venture, removed. Finally, college majors and company industries were categorized using a machine learning approach that used a set of manually categorized responses to predict TABLE A.1 START-UPS BY FOUNDING YEAR category based on word similarity. From this cleaned and augmented dataset, a panel and graph data dataset for analysis were then generated. Year Founded # of Start-ups Limitations 2016 22 It is difficult to determine whether the sample is representative. The collection of the data was heavily dependent upon the participation of individuals in existing 2015 25 networks, and as such the networks of the study’s partners are likely to be overrepresented. There is also a possibility that sub-segments of the of the ecosystem have been dropped or were never part of the outreach networks, and 2014 8 as such they would not end up in the resulting data sample. To a certain extent, this bias is not problematic, as it means that the data and analysis document those individuals and 2013 14 nodes that are actively and currently engaged in developing the ecosystem. This suggests that they are also more likely to participate in and respond to policies designed to develop the ecosystem. However, since there are few other 2012 20 databases with which to compare the data, it is difficult to assess exactly whether the sample is truly representative. 2011 13 The sample is small and influenced by outliers. To address this, where possible medians rather than means were used as descriptive statistics. Some of the analysis (for example, that of “successful” start-ups) relies on very small portions 2010 9 of the dataset, and thus give relatively weak information on unobserved, population-level characteristics. 2009 8 Data cleanliness is difficult to evaluate. Missing data could represent a true lack of connection in the ecosystem, or potentially could be the result of respondents dropping off the survey due to technical issues or a lack of interest to < 2009 23 complete the survey. Sometimes conflicting dates or meta 28 TECH START-UP ECOSYSTEM IN BEIRUT they pursue subsequent start-ups in other more FIGURE A.1 CUMULATIVE SUMMARY OF NODES developed ecosystems, rather than that they do not repeatedly start businesses. Currently there is no simple way to capture data on those individuals that leave, or a way of identifying them in the dataset. In addition, 1400 NODE TYPE the survey only captures entrepreneurs who were Event active during the survey period, and as such it does not Organization capture failed entrepreneurs who have dropped out of 1200 Person the ecosystem. In order to reduce the survey length in an attempt 1000 to boost response rates, the survey was necessarily Count Cumulative limited in scope. This means that there are some analyses that were not explored, such as limitations in 800 recruiting skilled employees (including engineers) or in internet connectivity. The analysis includes data across start-ups and SMEs 600 across a range of ages and neighborhoods. Due to limitations on sample size and accurate geographic data, it was not possible to conduct the analysis on 400 each age or neighborhood cohort separately. As such, the analysis may miss more nuanced findings related to time or geography. 200 Analysis 0 Stakeholders in the Lebanese start-up ecosystem can 2009 2010 2011 2012 2013 2014 2015 2016 be represented using a social network comprised of nodes (people, organizations, groups, and events), and FIGURE A.2 CUMULATIVE SUMMARY OF EDGES edges, which are the relationships between them. For the social network analysis, an edge was considered as EDGES LABEL a part of the ecosystem if either of its endpoints are in 1800 Inspired the region. From these edges, relevant nodes for the acquired network were extracted. Technically, the network is 1600 accelerated/incubated multipartite. However, given that individual investors hosted can function similarly to institutional investors, knows 1400 mentored founders of small start-ups are practically synonymous invested in with their companies, and events and groups are often Count Cumulative of Edges member of start-ups themselves, the network is treated as if it 1200 attended only contains one type of entity in order to simplify founded exploration. worked at 1000 studied at Nodes are never removed from the dataset, even if the represented entity no longer exists. In other words, 800 once a start-up has appeared in the dataset, it remains there for all subsequent years (see Figure A.1). There 600 were two reasons for this decision. First, accurate end- dates for start-ups were difficult to gather. Second, the 400 primary interest was in mapping the social dimensions of the start-up network, not creating a year-by-year catalogue of start-ups. Although start-ups may close, 200 socially they still function as nodes within the urban innovation ecosystem that may introduce other nodes 0 to second-order connections. 2009 2010 2011 2012 2013 2014 2015 2016 Program in Focus: Circular 331 29 BOX A.1 WHAT IS CENTRALITY? Degree centrality By calculating centrality measures of stakeholders in the dataset, key players in the community can be identified quantitatively. The diagrams below, while not specific to our dataset, help illustrate the definition and interpretation of each type of centrality. Red indicates higher centrality values. Blue indicates lower centrality values. Degree centrality measures the number of other nodes within the Closeness centrality ecosystem each node is directly connected to. It does not take into account any second-order connections. Closeness centrality measures a node’s social distance to other nodes. It is expressed as the inverse of the average distance from each node to every other node in the network. A low closeness centrality indicates that the firm is on the edge of the network. Eigenvector centrality Eigenvector centrality augments degree centrality by taking into account the connectivity of the nodes a node is connected to. Highly connected nodes within highly interconnected clusters have high eigenvector centrality. Betweenness centrality measures how many times a node acts as a gateway Betweenness centrality in the network. The higher the betweenness centrality of a firm, the more paths run through that firm to connect two other firms. High betweenness centrality means that a node is a key bridge or facilitator between different clusters. Source: Diagrams are from https://en.wikipedia.org/wiki/Centrality. Calculating centrality measures on different subnetworks connected of all key players, and they are the connectors in the in the data builds understanding of which players in the ecosystem (see Table A.2). However, the ecosystem appears be ecosystem are the most important. In the case of Beirut, the one big cluster without much complexity, and key players are analysis reveals that accelerators and incubators are the most more or less evenly spaced in this cluster. TABLE A.2 CENTRALITY OF KEY PLAYERS IN NETWORK Degree Betweenness Eigenvector Closeness Accelerator 5.15385 35.2692 2.96326E-05 5.99645E-07 Founder 4.5 4.29082 3.99248E-05 5.99934E-07 Investor 4.575 10.0875 3.30187E-05 6.00092E-07 School 2.93966 2.51293 2.47913E-05 5.97778E-07 Start-up 2.90071 1.23404 1.34049E-05 5.97793E-07 Note: While in other regions we have calculated centralities for sub-projections of the overall network, the low density of the reported network in Beirut limited the usefulness of reporting these numbers. As such, we have excluded these tables from this report. 30 TECH START-UP ECOSYSTEM IN BEIRUT Short-Term Success These results suggest that there appears to be a main cluster in the investment ecosystem that has only A fixed effects logit model was used where the dependent recently developed, and the majority of activity is variable is the probability of the start-up raising funding in occurring in this immature cluster. This hypothesis is a given year of existence33 and the explanatory variables supported by a comment from a survey respondent are lagged in order to gauge the effect of centrality in the who stated, “I wish it wasn’t centralized in limited investment network on a start-up’s short-term success. Only number of VCs and banks that are all in one bundle and the degree of centrality is a direct measure of the number of interconnected. some competition between the VCs investors a start-up has. Eigenvector and closeness centrality and the programs would push them to provide better capture the effects of second-order and beyond connections services.” to investors. The logit regression allows us to capture the binary outcome Long-Term Success of raising versus not raising funding while maintaining In the long-term, we are interested in a start-up’s ability interpretability. To understand the effect of each explanatory to create jobs for the ecosystem. As such, we use a variable on the probability of raising funding, we calculate the logit model where the dependent variable is hiring odds ratio by taking e^(estimated coefficient). The odds ratio occurrence weighted by years of existence, such that is a multiplier. If the odds ratio is 1.5, for example, then a unit we capturing the percentage of years that the firm hires increase in the dependent variable increases the likelihood of employees.35 For simplicity, we can also interpret this raising funding by 50%. number as the average probability a firm will hire in a given year. The greater a start-up’s closeness centrality, the more likely it is to raise funding in the next year. In addition, older start-ups are Well-funded start-ups are more likely to consistently more likely to raise funding. However, these effects are offset hire. An additional founder increases hiring likelihood by the finding that the more direct investment connections by 17 percent, and additional work experience the start-up has, the less likely it is to get funding in the next increases hiring likelihood by 24 percent. Participation year, and start-ups with greater eigenvector centrality that in acceleration programs and mentorship relationships are well connected within a cluster of accelerators, investors, have no significant effect on long-term hiring and funded start-ups are less likely to raise funding. The probabilities. Serial entrepreneurs are less likely to hire negative effects are likely due to the fact that once a start-up consistently, suggesting an opportunity to increase the has received funding once, it is less likely to need it in future overall quality of the start-up ecosystem as a learning periods and that the main cluster of accelerators, investors, and mechanism for founders. There is no significant impact funded start-ups is immature. of educational factors on job creation. TABLE A.3 EFFECT OF CENTRALITY IN INVESTMENT NETWORK ON SHORT-TERM FUNDING SUCCESS34 Estimate Std. Error z value Pr(>|z|) startup_prev_degree_all_investment -1.279e+09 2.201e+07 -5.810e+01 < 2e-16 *** startup_prev_eigenvector_all_investment -2.667e+26 2.459e+08 -1.085e+18 < 2e-16 *** startup_prev_closeness_all_investment 1.329e+14 9.793e+11 1.358e+02 < 2e-16 *** year_existence 1.103e+07 3.158e+06 3.494e+00 0.00055 *** TABLE A.4 EFFECT OF START-UP CHARACTERISTICS ON LONG-TERM FUNDING SUCCESS Estimate Std. Error z value Pr(>|z|) sum_acceleration_occurrence 1.521e-01 3.051e-01 0.498 0.618176 sum_funding_amount 2.445e-07 7.623e-08 3.208 0.001339 ** num_distinct_investors -4.968e-02 1.770e-01 -0.281 0.778979 num_distinct_investor_regions -2.927e-01 3.317e-01 -0.882 0.377632 num_distinct_mentors 9.371e-02 1.060e-01 0.884 0.376738 num_distinct_founders 1.564e-01 6.973e-02 2.243 0.024892 * sum_founder_previous_startups_founded -4.177e-01 1.152e-01 -3.626 0.000288 *** sum_founder_previous_jobs 2.379e-01 8.163e-02 2.914 0.003565 ** sum_bachelor_master_professional_doctorate_degrees -1.410e-01 1.185e-01 -1.190 0.233998 sum_associate_bootcamp_certificate_degrees 1.248e-01 4.769e-01 0.262 0.793596 (Intercept) -1.454e-01 1.739e-01 -0.836 0.403078 Program in Focus: Circular 331 31 NOTES 1. As a reference for the sample size, the Investment Development 19. 55 distinct start-ups received funding out of a total of 142 unique Authority in Lebanon (IDAL) reported 800 companies in the IT start-ups. sector, with 200 companies focused on software development, in Lebanon in 2005. See: IDAL ICT Sector Statistics in Lebanon 20. Ratio of the number of investments received by funded start-ups that (2015). http://www.investinlebanon.gov.lb/Content/uploads/ were accelerated divided by the number of investments received by SideBlock/160307124909480~ICT%20Statistics%202015.pdf funded start-ups that were not accelerated. 2. https://startupgenome.com/. 21. https://techcrunch.com/2015/03/22/mentors-are-the-secret- weapons-of-successful-startups/. 3. http://www.digital.nyc/. 22. Events reported included: Lebanon Open Innovation Week, Africa 4. http://www.techmap.london/. Open Data Conference, Megawatt Challenge, and Women for Resilience Cities. 5. https://www.galidata.org/. 23. This is confirmed by respondents to the survey. One participant 6. See “Outreach Partners” in Appendix A for full list of partners. stated that, “Players in the ecosystem, for example, incubators, hubs, accelerators, and so on, should work together to build a stronger 7. See “Limitations” section in Appendix A for more information on the ecosystem rather than competing and dividing opportunities.” limitations to our approach. 24. As one responder to the survey stated: “Regarding Circular 331, 8. The methodology of this analysis is focused on the founder’s I believe it is a positive contribution to fostering the growth of the perspective and how that compares across regions. Founders in a tech sector in Lebanon. On the other hand, the risk is to fund start- given city may not perceive an existing constraint as such because ups without valid business prospects, especially if the incentives of they have adapted to the environment and do their business despite the different parties are not aligned properly.” See Program in Focus. of such constraints. 25. http://www.datacity.paris/. 9. https://techcrunch.com/2015/03/22/mentors-are-the-secret- weapons-of-successful-startups/. 26. There are very few active bootcamp provider operating in Lebanon, the most active one being SE Factory. 10. The data collected for this analysis suffers from an inherent survivorship bias, the precise impact of which is difficult to quantify 27. http://www.skilledup.com/articles/the-ultimate-guide-to-coding- (see “Limitations” section in Appendix A for more details). bootcamps-the-exhaustive-list. 11. We did not collect data on costs associated with these procedural 28. https://tampere.demola.net/. tasks as the numbers were less directly comparable across regions. 29. https://tech.cornell.edu/jacobs-technion-cornell-institute/overview. 12. Of 255 instances of founding, only 43 percent (110) of those instances have the age of the founders available. 30. http://www.startupchile.org/. 13. Gross enrollment in tertiary education in Lebanon for both sexes was 31. http://www.k-startupgc.org/. 38 percent in 2015 (http://data.uis.unesco.org/). 32. http://2015.bdlaccelerate.com/everything-you-need-to-know- 14. For example, postgraduate qualification in law, business, or medicine. about-bdl-circular-331/; and http://fletcher.tufts.edu/~/media/ Fletcher/Microsites/IBGC/Student%20Research/IBGC_Circular_331. 15. Of 255 instances of founding, only 48 percent (124) of those instances pdf. have major available. 33. The amount of funding raised is not accounted for, since this is heavily 16. Of 636 instances of work history at the time of founding, 52 percent influenced by the type of business and prone to outliers. (336) have job level listed. 34. For this regression, start-ups were restricted to those founded after 17. Since start-ups can relocate, this list includes some accelerators that 2008 in order to avoid confounding influences from the global are not themselves in Lebanon. financial crisis. 18. Ratio of accelerated firms that received investment over 35. The number of employees hired is not accounted for, since this is nonaccelerated firms that received investment. A ratio of 1 means heavily influenced by the type of business and prone to outliers. equal opportunities. A ratio below one means nonaccelerated firms are more likely to receive investment. A ratio of more than one means accelerated firms are more likely to receive investment. 32 TECH START-UP ECOSYSTEM IN BEIRUT This work is available under the Creative Commons Attribution Non- Commercial 3.0 IGO license (CC BY NC 3.0 IGO). Under the Creative Commons Attribution Non-Commercial license, you are free to copy, distribute, transmit, and adapt this work for non-commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Mulas, Victor; Qian, Kanty; and Henry, Scott. 2017. Tech Start-up Ecosystem in Beirut. Findings and Recommendations. License—Creative Commons Attribution Non-Commercial CC 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Responsibility for the views and opinions expressed in the adaptation rests solely with the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@ worldbank.org