The Global Entrepreneurship and Development Index Dr Habil

The Global Entrepreneurship and Development Index Dr. Habil László Szerb Associate professor University of Pécs, Faculty of Business and Economics Acknowledgements: OTKA Research Foundation (K 81527) provided financial support for the project, thanks for it.

Economic development and entrepreneurship • Theoretical setup – Joseph Schumpeter (1911) - Innovation – Paul Douglas (1934) – (K, L) – Robert Solow (1957) - TFP – W. W. Rostow (1960) – Stages of Growth – Paul Romer (1990) - Knowledge – Porter and Sachs (2002) Stages of Development – Samuelson (2009) Acs, Audretsch Strom 2009 • TFP is what remains unexplained after (capital and labor) accounted for. – Knowledge (Romer) – Institutions (North) – Entrepreneurship (Schumpeter) • How to combine inputs is key to development. • Need entrepreneurs.

The connection between entrepreneurship and economic development

Entrepreneurship and Prosperity • Factor driven stage most people are involved in underproductive, unproductive or destructive entrepreneurship. • Efficiency driven stage people shift out of destructive entrepreneurship • Innovation most people shift out of unproductive and most destructive entrepreneurship.

Explaining development – the role of indices • • • Global Competitiveness Index – World Economic Forum The Index of Economic Freedom – Heritage Foundation Ease of Doing Business - World Bank Global Creativity Index – Richard Florida (Prosperity Index – Legatum) – None of these focus on entrepreneurship!

Entrepreneurship definitions, concepts, measure • Entrepreneurship definitions – different in terms of aims, can be societal, academic or teaching phenomenon (Shane and Venkatamaran 2000) – One-dimensional definitions • Factor of production • Assemble resources • New business creation • Innovation • Opportunity recognition and exploitation – Multidimensional definitions – includes more than one from the followings: • Innovation • Opportunity exploitation • New venture creation • Risk taking • (judgemental) decision making • Specific behavior – entrepreneurial attitudes • (positive) result orientation – wealth, growth, value creation

Entrepreneurship definitions, concepts, measure • Entrepreneurship concepts: – Process perspective • examining business gestation, life-cycle issues – Context perspective • The environmental (outside) factors of entrepreneurship • What is not entrepreneurship? – Small business, – Self employment – Ordinary, routine type managerial tasks, – Activities with no or very low risk – Change of ownership – Mergers, buyouts

Entrepreneurship definitions, concepts, measure • Measures of entrepreneurship – One dimensional measures dominantly • Self employment, ownership rate – attitude towards SE • Total Early-phase Entrepreneurial Activity (TEA) Index • Business density • Entry/exit rate • Innovation rate (e. g. share of innovative businesses) • High growth businesses – Multidimensional measures – can be interpreted as quasi entrepreneurship measures • Doing business • Index of Economic Freedom • Global Competitiveness Index • Creativity Index

Entrepreneurship definitions, concepts, measure • Basic problems – multidimensional definition and concept – one-dimensional measure – Environmental, institutional factors are missing – Measures mainly quantity and not quality – Correlation between economic development and the measure of entrepreneurship is negative – contradicts to mainstream dominant theories

Spearman Correlations between TEA and the Business Indexes GCI IEF -. 533** -. 323* -. 430** TEA GCI = Global Competitiveness Index IEF = Index of Economic Freedom EDB = Ease of Doing Business Index TEA = Early-stage entrepreneurial activity n=57 EDB

Spearman Correlations between COMPENDIA and the Business Indexes GCI IEF -. 760** -. 073 EDB -. 209 COMP GCI = Global Competitiveness Index IEF = Index of Economic Freedom EDB = Ease of Doing Business Index COMP = COMPENDIA, Business Ownership rate (the number of business owners divided by total labor force) n=22

Index building – the building blocks • The GEDI is a complex measure of entrepreneurship • The Building Blocks – Entrepreneurial attitudes is defined as a population’s general attitudes about entrepreneurship including opportunity recognition, networking, start-up skills taking risk, acceptance of entrepreneurs with high status – Entrepreneurial activity is defined as the quality of startup including the motivation of start-up, the level of education of the entrepreneur, the sector (high or medium), and the potential not to be overly competitive. – Entrepreneurial aspiration is defined as the early-stage entrepreneur’s effort to introduce new products and/or services, develop new production processes, penetrate foreign markets, substantially increase their company’s number of employees, and finance the business with formal and/or informal venture capital.

Index building – the combination of the variables • The GEDI combines together individual and institutional variables – following the logic of the interaction variables applied in regression technique. Each of our 14 pillars is a result of the multiplication of an individual variable and an associated institutional variable. • A critical part of index-building is identifying the proper weights - a novel approach to determining/interpreting weight – In this case, institutional variables can be viewed as particular (country-level) weights of the individual variables. • Individual variables – source Global Entrepreneurship Monitor survey • Institutional variables – sources, other than GEM (UNESCO, WEF, OECD etc. )

Index building – Attitudes variables

Example • SKILL is defined as the percentage of the 18 -64 aged population claiming to posses the required knowledge/skills to start business • Source: GEM adult population survey • EDUCPOSTSEC is defined as the gross enrolment ratio in tertiary education • Source: UNESCO • SKILL x EDUCPOSTSEC = STARTUP SKILLS , combines together the start-up skills with high level education – Benchmarking individuals: high growth ventures are lead by educated skillful entrepreneurs – Potential support for business start-up in terms of skills – Higher probability to provide informal finance – business angel

Example – an exception • MARKETAGGLOM = MARKETDOM*URBANIZATION - The size of the market: A combined measure of the domestic market size and the urbanization that later measures the potential agglomeration effect – MARKETDOM - Domestic market size that is the sum of gross domestic product plus value of imports of goods and services, minus value of exports of goods and services. (World Economic Forum) – Urbanization that is the percentage of the population living in urban areas (Population Division of the United Nations) • OPPORTUNITY - The percentage of the 18 -64 aged population recognizing good conditions to start business next 6 months in area he/she lives. • OPPORTUNITY PERCEPTION = MARKETAGGLOM* OPPORTUNITY - This pillar captures the potential “opportunity perception” of a population by considering the size of its country’s domestic market and level of urbanization.

Index building – Activity variables

Index building – Aspirations variables

Index building - pillar combination • Another novelty of our index-building is the way the pillars are combined into sub-indexes. • Combination of the pillars – connection to weighting – use the (weighted) average of the pillars; – dimension reduction methodology – principal component-, factor analysis – Regression technique – (Grey) relational analysis – not really applied • Problems: – Does not really take into account the interdependencies of the pillars. – Relies on the statistical properties – no real theoretical foundation. – Sensitive to sample size - smaller sample may not provide acceptable solution

Index building – the Penalty for Bottleneck • Assumptions, theoretical setup – Configurational theory: not the individual elements but the integration of the elements (pillars) to the whole system is important (Miller 1986). – Theory of Weakest Link: The elements of the system can only be partially substitutable with each other • A bad performance in one pillar cannot be compensated fully with a good performance in another pillar – Theory of Constraints: The performance of the system is determined by the worst performing element (pillar) • The Penalty for Bottleneck provides a solution – Optimal configuration: the pillars should be about the same level

The general methodology of the PFB Now, let’s select unit 1 and rearrange the features from the lowest to the highest: we define the variability with the range, called bottleneck, which is the difference between the value of actual feature and the value of the worst feature. The bottleneck vector (Ri) can be defined as

The general methodology of the PFB Now, we apply a penalty function in general form such as: § The index value representing the overall performance of the unit over the k features is calculated as the arithmetic mean (hereafter mean) after applying the PFB methodology: § The value of the index is mainly determined by the variable with the worst value that can be considered as the weakest link amongst all the variables (features). § The size of the penalty depends on the difference between the value of the worst variables and the value of the particular variable: The higher the difference the higher the penalty is.

Logarithmic penalty function § For example, assume the normalized score of a particular feature is 0. 90, and the lowest value indicator is 0. 50. The difference is 0. 40. The natural logarithm of 1. 4 is equal to 0. 336. Therefore the final adjusted value of the feature is 0. 50 + 0. 336 = 0. 836 instead of 0. 90. § The largest potential difference between two indicators can be 1, when a particular country has the highest value in one indicator and the lowest values in another. In this case the natural logarithm of 2 = 0. 693, so the maximum penalty is 1 -0. 693 = 0. 307.

The logarithmic penalty function: original and after penalty values

Application of PFB to GEDI - steps • Multiplication of the variables to calculate pillar values • Normalization: We normalized the scores of all the 14 pillars, to maximum 1 and the lowest value to zero in each pillar. – Outliers are handled with truncation - The maximum value is set to reduce the difference between the first and the second countries and the second and third countries in a particular entrepreneurial feature indicator to 5 percent. • This method preserves the ranking of the countries in a particular entrepreneurial feature, but reduces the relative differences between the leading country and the other nations. • The calculation of the PFB adjusted values in each country: The penalty is calculated in each of the three sub-index level of the particular country. The normalized value of each pillar in a country is penalized by linking it to the score of the indicator with the weakest performance of the sub-index in that country. • The calculation of the PFB adjusted sub-indexes: The PFB adjusted sub-indexes can be received by averaging the PFB adjusted pillars for each sub-index. • The calculation of the PFB adjusted GEDI: The GEDI is derived by calculating the simple non-weighted averages of the three sub-indexes.

PFB properties • Assumption: all pillars should be positively correlated with each other – If not, the improvement of one pillar decreases the value of another pillar, • Sensitivity to outliers: goes back to the problem of selecting the benchmark properly • Problem of the potentially different distributions of the pillars and sub -indexes – Aspiration has the lowest and attitudes has the highest average – Question: is normalization (0 mean and 1 variance) a better solution? – over-normalization

Variables • • 71 countries 31 variables 16 from GEM 15 from other data sources 14 pillars The 3 As 1 super index Used the 2002 -2008 pooled data – it is rather a stock than a flow concept • (Moved to examine changes over time: two years moving average – in some cases that is not enough data – e. g. high growth businesses) • Institutional data: two years average 2006 -2007 – various sources

The GEDI rank of the countries

Table 7: The correlation coefficients between GEINDEX and other major indexes All coefficients are significant at a level better than 0. 001

Differences over development

Cluster analysis

Traffic Light Method • Bad • Better • Best

Italy’s position

Improvement – marginal analysis • Selection of the worst pillar: – QUALITY OF HUMAN RESOURCES = 0. 27 – RISK CAPITAL= 0. 27 – NEW PRODUCT = 0. 28 • Increase the worst pillar value by 0, 1 • Effect: – QUALITY OF HUMAN RESOURCES = 0. 37 – ACTINDEX increase from 0. 363 to 0. 393 – GEDI increase from 0. 407 to 0. 417 • Effect: – RISK CAPITAL = 0. 37 – ASPINDEX increase from 0. 362 to 0. 381 – GEDI increase from 0. 407 to 0. 413 – New bottleneck: NEW PRODUCT, – More than one bad pillar, the marginal improvement is smaller than in the previous case

Questions • Is there some flaw in it that needs to be corrected? • What about theoretical foundation of the index? – How to improve it? • Is the PFB methodology applicable? – What about the marginal analysis? – How to deal with the different substitution effects? • How do you move forward ? What are the issues? – Adding new years? – What kind of changes – conceptual, technical – are necessary to apply to the regional level? What regional level? • Are the policy suggestions acceptable?

Thank you for your attention!