Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, BUSINESS AND MANAGEMENT (business.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

.

 

date: 18 September 2018

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and Impact

Summary and Keywords

In the late 1990s, there was considerable interest in national differences in entrepreneurial activity. The Global Entrepreneurship Monitor (GEM) research program was developed to provide harmonized, cross-national measures of participation in business creation; business creation was considered a critical aspect of entrepreneurship. This information was considered important for understanding the national characteristics associated with business creation and its subsequent impact on economic growth. The initial effort involved 10 countries in 1999. By 2014 Adult Population Surveys (APS) had been completed 705 times in 104 countries and with six special samples; this involved 2.3 million individual interviews. While there have been changes in the administrative structure and the focus of the annual global reports, the most significant data collection procedures have been stable since 2002. The GEM APS data sets are currently the only harmonized cross-national comparisons of business creation and business ownership. Designed to provide estimates of the prevalence of both business creation and existing firms, they also allow estimates of the total number of business ventures. GEM data sets are publically available three years after completion, providing a unique resource for assessing factors affecting business creation and its subsequent role in economic growth. Systematic assessments by national experts in participating countries provide measures of the national entrepreneurial framework conditions, complementing a variety of established measures of national economic and political characteristics.

There are three distinct features that characterize the GEM initiative: the unique organizational structure, the global reports summarizing annual assessments of entrepreneurial activity, and data sets assembled and made available for public use. The initial organizational structure, a collaborative arrangement among national teams, was replaced by membership in the Global Entrepreneurship Research Association (GERA) in 2004. The annual global reports emphasize comparisons among member countries, the annual national reports the country-specific situations. Both are designed to facilitate reality-based public policy.

Data collection for the APS provides harmonized comparisons of business creation across countries and within-country time series. The APS data has made clear the substantial variation among countries, by a factor of 10; that national levels of participation are very stable over time; that business creation is much more prevalent in poorer countries; that all segments of society are active in business creation; and that business creation is an important catalyst for the processes that lead to economic growth. The National Expert Survey (NES) questionnaire data provides information about the nature of the entrepreneurial framework in the GEN countries.

There is much to be learned about the relationships between national context, entrepreneurship, and economic growth. The unique information in the GEM data sets should continue to facilitate improved understanding of this important phenomenon.

Keywords: GEM, global entrepreneurship monitor, nascent entrepreneurship, total entrepreneurship activity, TEA

Introduction

Policy development and academic interest in understanding entrepreneurship and differences among countries were growing in the 1990s. The United States was widely considered the global leader in entrepreneurship. There was also a strong interest among national leaders in how to facilitate economic growth, and the Global Competitiveness Reports were providing assessments of countries as suitable contexts for private businesses (Schwab & Sachs, 1987, 1998). Like many business schools, the London Business School was developing an emphasis on entrepreneurship, led by Michael Hay. In a meeting with Dean George Bain about this new initiative in 1998, the potential for cross-national comparison similar to those in the competitiveness reports came up. This led to a conversation between Hay and Bill Bygrave, a colleague visiting from Babson College.

After some discussion, they approached Professor Paul Reynolds, also visiting from Babson College, and managing the first U.S. Panel Study of Entrepreneurial Dynamics [PSED I]. A great deal of time and money had been invested in the PSED initiative to design a procedure that would efficiently develop representative samples of those active in business creation. Furthermore, the PSED screening protocol had been successfully implemented in several other countries (Canada, Netherlands, Norway, and Sweden). This suggested that if the same screening protocol were simultaneously implemented in a set of countries, it could provide reliable cross-national comparisons of participation in business creation. And business creation was considered a critical component of entrepreneurial activity.

Following several brainstorming sessions with experienced entrepreneurial scholars, it was decided to mount a small pretest. It was already clear from PSED pretests that it was cost-effective to add a short screening module to commercial omnibus surveys. The design team developed a 10-item interview of simple “yes-no” questions. Colleagues who might be interested in being involved in Canada, Finland, and Germany were approached about the project. Along with the United Kingdom and the United States, this would provide a five-country pretest. It was agreed that each national team would sponsor a representative sample of 1,000 adults using the 10-item interview. Within two months, the data was assembled and the results provided confidence in the cross-national comparisons. More colleagues were contacted to create a 10-country effort for 1999 that included National Expert Surveys [NES] to obtain more country-specific details. Based on the success of and interest in the 1999 effort, more teams joined the project, additional modules were added to the interview schedule, the minimum sample was increased to 2,000, and the APS and NES procedures were improved. The basic APS protocol and NES questionnaires were stabilized by 2003.1 As of 2014, more than 100 countries had become involved in the GEM project for one or more annual assessments.

The initial organizational structure was considered a collaboration among participating national teams coordinated by staff affiliated with Babson College and the London Business School. It was assumed that the host institutions would find external sponsors to support the coordination team and the national teams would cover the costs of all local data collection. While the Kauffman Foundation was a major sponsor of the Babson College component through 2003, no comparable sponsor was located to cover the London Business School financial responsibilities. As a consequence, the national teams were asked to provide supplemental funds to support the coordination team. The coordination team was responsible for managing all national surveys, providing consolidated data sets for national team analysis, producing the annual global reports, and organizing an annual meeting to review progress and enhance the project design. National teams were responsible for funding their national surveys, completing national expert interviews, and producing annual national reports.

This structure changed in 2004, with the formation of the Global Entrepreneurship Research Association (GERA) as a not-for-profit (charitable) association chartered in the United Kingdom to oversee the GEM project. GERA is supervised by a board composed of members elected by the national teams and representatives of the two founding institutions. After this reorganization, the design of the annual population and expert surveys remained the responsibility of the coordination staff, but individual national teams were expected to locate, supervise, and reimburse the national survey organizations that did the fieldwork, complete the national expert surveys, and produce national reports.

There have been adjustments in the sponsoring institutions, as reflected in the presentation in Table 1. Babson College has been an official sponsor throughout the duration of the project, but London Business School disengaged after 2007, although it is still the legal host of the GERA and manages the membership funds provided by the national teams. The coordination staff consists of full-time researchers and database professionals. Various GEM team members have become involved in coordination initiatives as needed for special assessments.

Table 1. GEM Development: Number of Participating Teams and Host Institutions

Year

Number of Countries

Host Institutions, Major Sponsors

1998

5

Babson College, London Business School

1999

10

Babson College, Ewing Marion Kauffman Foundation, London Business School

2000

21

Babson College, Ernst & Young, Ewing Marion Kauffman Foundation, London Business School

2001

29

Babson College, Ewing Marion Kauffman Foundation, IBM, London Business School

2002

37

Babson College, Ewing Marion Kauffman Foundation, London Business School

2003

31

Babson College, Ewing Marion Kauffman Foundation, London Business School

2004

34

Babson College, London Business School

2005

35

Babson College, London Business School

2006

42

Babson College, London Business School

2007

42

Babson College, London Business School

2008

43

Babson College, Universidad del Desarrollo (Chile)

2009

53

Babson College, Reykjavik University (Iceland), Universidad del Desarrollo (Chile)

2010

59

Babson College, Universidad del Desarrollo (Chile)

2011

53

Babson College, Universidad del Desarrollo (Chile), Universiti Tun Abdul Razak (Malaysia)

2012

69

Babson College, Universidad del Desarrollo (Chile), Universiti Tun Abdul Razak (Malaysia)

2013

70

Babson College, Universidad del Desarrollo (Chile), Universiti Tun Abdul Razak (Malaysia)

2014

73

Babson College, Tecnológico de Monterrey (Mexico), Universidad del Desarrollo (Chile), Universiti Tun Abdul Razak (Malaysia)

2015

62

Babson College, Tecnológico de Monterrey (Mexico), Universidad del Desarrollo (Chile), Universiti Tun Abdul Razak (Malaysia)

Note: (*) Based on cover pages of annual reports, available at www.gemconsortium.org.

The following presents the objectives and conceptual framework that are the basis for the GEM research program, followed by a discussion of the two distinctive data collection activities, the Adult Population Surveys (APS) and the National Expert Surveys (NES). A review of some basic findings illustrates the scope and potential of the initiative. The chapter finishes with a discussion of the scope of impact and a discussion of research opportunities. Information on access to the data is discussed in Appendix C.

Rationale, Objectives, and Conceptualization

At the beginning of the GEM program, entrepreneurship was defined as (Reynolds et al., 1999):

Any attempt at new business or new venture creation, such as self-employment, a new business organization, or the expansion of an existing business, by an individual, a team of individuals, or an existing business.

Considered a core feature of all conceptions of entrepreneurship, this has continued to be the basic focus of the GEM data collection. The initial objectives were to determine:

  • Does the level of entrepreneurial activity vary between countries, and, if so, to what extent?

  • Does the level of entrepreneurial activity affect a country’s rate of economic growth and prosperity?

  • What makes a country entrepreneurial?

Over the past 15 years, these have been adjusted slightly, but the focus on business creation and its role in national economic well-being has remained. The conceptual initial framework is presented in Figure 1. The model was developed to capture a broad range of features and processes considered to affect business creation, as one aspect of business dynamics, and ultimately national economic growth. There are several important features of this conceptualization. Most significant, the dependent variable is national economic growth. This positioned the GEM initiative as an effort designed to advance general economic well-being, not simply an effort to promote entrepreneurship.

Entrepreneurial processes are considered to complement other mechanisms, associated with both the large firm and small and medium firm sectors that may also facilitate national economic growth. This emphasis facilitates attracting financial support, particularly from public agencies, and shifts the focus to the role of entrepreneurial processes in the broader economy.

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 1. GEM Conceptual Model: 2000.

A second feature is related to the emphasis on developing empirical data related to the model. There was and is considerable variation in the availability of reliable cross-national measures of the various elements in the model. There have been, both in 1999 and now, considerable work on developing harmonized cross-national measures of the social, cultural, and political context; a range of general national framework conditions; and national economic growth. The World Economic Outlook has continued to provide reliable, harmonized measures of Gross Domestic Product, adjusted for purchasing power parity and computed on a per capita basis (International Monetary Fund, 2016). Standardized measures of a wide range of national characteristics are assembled by a number of cross-national comparisons, such as the Global Competitiveness Reports (Schwab & Sala-i-Martín, 2015), the World Competitiveness Reports (IMD, 2015), the Index of Economic Freedom (Heritage Foundation, 2015), and others. The details in the appendices of these reports provide useful data on specific national characteristics.

There were not, however, reliable cross-national measures of either the entrepreneurial framework conditions or the three features considered important to affecting or measuring entrepreneurial activity—the perception of opportunity, capacity to pursue business creation, and measures of business dynamics, particularly measures of new firm creation.

“Business dynamics” is considered to include the creation, deactivation, expansion, and contraction of business firms. Reliable cross-national measures of these changes do not exist. While it may appear that national business registries—and almost all countries maintain some type of list of private businesses—might be a suitable source of business creation measures, this is not the case. The diversity in criteria for adding new listings and the different points in the firm life course when new ventures are listed, then and now, prevent precise cross-national comparisons. In addition, most registries are developed to facilitate some form of tax collection and are considered highly confidential. Gaining access to confidential tax records is a major challenge. Even now, national registries are a very poor source of cross-national comparisons of the presence of business firms or tracking dynamics as firms are created and discontinued.

The solution that was adopted was to create two data collection activities. The first, based on the successful fieldwork of the Panel Study of Entrepreneurial Dynamics (PSED) projects, was to implement Adult Population Surveys (APS) to measure business creation activity, perceptions of the entrepreneurial climate, and self-assessments regarding entrepreneurial potential. The second, to capture measures of the entrepreneurial framework conditions, was a series of personal interviews with well-informed individuals, the National Expert Surveys (NES). The NES involved both an unstructured discussion and a fixed-choice questionnaire that would allow national experts to provide comments and judgments on their country as a context for business creation.

Data Collection: Business Dynamics

Adult Population Surveys

Adult population surveys were developed to provide harmonized measures of the perception of opportunities, measures of the capacity to implement new firms, and actual behavior associated with creating or managing business ventures. The design reflected the impact of several constraints.

  • It was important to minimize costs, as each national team was confronted with the challenge of raising financial support each year to support data collection, and many teams are located in smaller countries or countries with limited resources.

  • The interview schedule was to be administered to all segments of society, many with limited education or business exposure, in a wide range of languages.

  • There was considerable diversity in the organizations that would administer the interview schedules.

This led to utilizing a short interview schedule with direct, straightforward questions and simple response alternatives. Many items required a “yes” or “no” response, reflecting an assumption that this is similar in all languages.

A further issue was the use of the interviews to develop measures from two groups of respondents. The first are those representing the general adult population, a source of information on the entrepreneurial climate, perceptions of opportunities, and general capacity to implement new firms. The second are those identified as active in creating new firms or managing existing ventures. The sampling design was created to minimize the sampling errors of measurement related to estimating the prevalence of participation in business creation, which averages less than 10 per 100 respondents. It was assumed that samples of 2,000 per country would provide relatively precise measures of the prevalence of nascent entrepreneurs. As a result, population sample sizes were large enough to provide relatively precise measures of the attitudes and perceptions of the general population.

The initial protocol was developed as part of a longitudinal research design, where the individuals active in business creation were to be contacted at periodic intervals to determine progress toward implementation of a new firm (Reynolds, 2000; Gartner, Shaver, Carter, & Reynolds, 2004; Reynolds & Curtin, 2009, 2010). The initial GEM pilot study, completed in 1998, utilized only 10 “yes” or “no” questions. This was elaborated in 1999 by adding some additional questions for those considered to be active in the business creation process, nascent entrepreneurs. Additional modules were added in 2000 for those considered to be business owners and informal investors. Both the screening items and special modules were modified and enhanced in 2001. Since 2002 there has been little change in the core items, but the availability of an ongoing survey operation has led to the introduction of a number of special topics, some for a single year and others for multiple years. Most notable has been the introduction of items to identify and describe social entrepreneurial activity (SEA) in 2009 (Bosma & Levie, 2010) and an expansion of modules focusing on entrepreneurial employee activity (EEA) or business-sponsored business creation where the founding team may not share ownership, in 2011 (Bosma, Wennekers, & Amorós, 2012).

An overview of the procedure for 2003 is provided in Figure 2. As can be seen, there are four components. The screening items, asked of all adult population respondents, are represented in the boxes to the left. Based on responses to the screening items, those considered candidates for nascent entrepreneur status then received the items in the second module, those considered candidates for business owner-manager status (new and established firms) received the items in the third module, and those reporting they had made informal investments in new firms received the fourth module. The majority of respondents, from 70% to 95%, did not qualify for any module (nascent, new firm, or informal investor). Information from all respondents about basic socio-demographic features and responses to all basic screening items are included in the data sets. Copies of all questionnaires used between 1998 and 2012 are available in the consolidated GEM APS codebook (Reynolds, 2016).

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 2. Structure of GEM Adult Population Interview Schedule.

In terms of resource allocation, the most expensive part of the interview was locating and interviewing a representative sample of adults with the screening module. Only a minority are asked details about start-up ventures, operating businesses, or informal investments. Efforts to contain costs led to a decision to restrict the number of screening items to 10. After 2002 there was an interest in including a total of 14 items in the screening module. To minimize costs, a special procedure was used. Any respondent who qualified for any special topic module was asked all 14 items. All those who did not qualify for any special topic module were asked the first six and randomly selected to receive one of two sets of four additional items, identified as A or B in Figure 1.2 Over time, some national teams have added a range of modules related to special topics that has expanded the interview schedule. But all national teams cover the basic GEM topics agreed to at the national design meeting.

Identifying Participation in Business Activity

The final form of the screening items used to determine participation in business creation are:

  • You are, alone or with others, currently trying to start a new business, including any self-employment or selling any goods or services to others?

  • You are, alone or with others, currently trying to start a new business or a new venture for your employer as part of your normal work?

Those responding yes to either or both items are then asked additional questions to identify those with the following characteristics:

  • Had been active in start-up behavior in the past 12 months

  • Expected to own all or part of the new firm

  • Had not paid owners’ salaries and wages for more than three months3

Those that were active, expected to own the venture, and had not reached initial profits were considered nascent entrepreneurs.

A third screening item was used to identify those that might be considered, at the time of the interview, business owner-managers:

  • You are, alone or with others, currently the owner of a company you help manage, self-employed, or selling any goods or services to others?

It turns out that “trying to start” and “owning and managing a business” are not precise concepts in any language. Some of those who claimed they were “trying to start” a business reported substantial periods of profits. Some of those who claimed they were owning and managing a business had yet to achieve profitability. As a result, during the processing of the data these two types of special cases are reallocated. Those reporting they are starting a business that was profitable are reclassified as owner-managers, and those reporting they are owner-managers but had not reached profitability are reclassified as nascent entrepreneurs. As exactly the same questions are asked about the initiatives in the two modules, the full descriptions of these efforts are reclassified.

Those identified as owner-managers are separated into two groups:

New Firm Owner-Managers are those that:

  • Were active in the management of a firm

  • Owned all or part of the business

  • Had paid owners’ salaries and wages for more than three months but for less than 42 months (3.5 years).4

Established Firm Owner-Managers are those that:

  • Were active in the management of a firm

  • Owned all or part of the business

  • Had paid owners’ salaries and wages for more than 42 months (3.5 years).

While the unit of analysis in the surveys was individuals, it can be converted to a sample of business ventures by adjusting individual case weights through division by the size of owning-managing team.5

Total Entrepreneurial Activity/Early Stage Index

Business creation can be considered a two-stage process, the start-up or pre-profit stage followed by a period of initial profits as a new firm becomes an established business. The combination of low prevalence rates and relatively small samples—the GEM minimum is 2,000—leads to rather wide margins of error.6 The margin of error can be reduced if more activity is included in the measure of activity, which does not require an increase in the sample size. This led to the creation of the Total Entrepreneurial Activity (TEA) or Early Stage Entrepreneurial Activity index. Those that qualified as either a nascent entrepreneur or a new firm owner-manager were assigned a value of 1, all others a value of 0. Capturing these two early stages of the firm life course increased the prevalence rate by an average of 67%.

As shown in Table 2, the global prevalence of nascent entrepreneurs is 8.1 per 100, the global prevalence of new firm owner-managers is 5.9 per 100, and the global prevalence in these two early stages, the TEA index, is 13.6 per 100. The TEA index is less than the sum of the two stages (14.0 per 100) because a small proportion were involved in projects in both stages but were counted once. This provides an accurate measure of human participation but a slight undercount of total business creation activity. The bottom row of Table 2 provides the prevalence of those considered active an owner-manager of established firms, those with profits for over 3.5 years. For all measures there is a substantial variation across the 104 countries, as shown in the two right columns.

Table 2. Global Measures of Participation in Business Creation: 2000–2014

Number per 100 adults 18 to 64 years old

Global Average

National Low

National High

Nascent entrepreneur

8.1

2.0

34.3

New firm owner-manager

5.9

0.6

27.7

TEA Index

13.6

2.8

54.7

Established firm owner-manager

9.3

1.6

36.5

Note: Based on 2,251,894 respondents from 687 samples representing 104 countries averaged across years 2001 to 2014. Equal weights for all countries, Yemen excluded.

Several other features have been systematically measured for all ventures. Beginning in 2001, all those participating in the firm creation process were asked about the reason for their participation: to take advantage of a “business opportunity” or because there were “no better choices for work.” This was originally labeled as “opportunity versus necessity,” reflecting an assumption that “no better choice” was unemployment. Subsequent research, however, has found a large proportion of established owner-managers’ report they have “no better choices,” suggesting that managing a profitable business is more rewarding that available employment. The word “necessity” does not reflect the situation of successful business owners with routine employment options.

The interviews also provided measures of the size of the start-up or management team, judgments about the extent to which the venture is providing innovation in the markets, relative emphasis on new technology, extent of out-of-country exports, the sector in which the venture will or does operate,7 and employment expected in the first year or at the time of the interview and expected in the next five years.

Sampling and Weights

The selection of countries to be involved reflected the professional networks of the coordination team and the ability of the national teams to develop financial support to cover the costs of data collection. While the sample of countries cannot be considered a random selection of any known population of nations, countries representing most regions of the world and most stages of economic development are represented. Global coverage improved after 2010 as more developing countries (particularly in Sub-Saharan Africa) were included in the program. Table 3 presents the countries included between 1998 and 2014 and the number of annual APS surveys. They are organized by global region. In addition, APS surveys have been completed in five distinctive sub-samples. Details of the national sample sizes by year are provided in Appendix A.

Table 3. National Participation in GEM: 1999–2014

Asia: Developed

Central, Eastern Europe

Latin America, Caribbean

Hong Kong

5

Bosnia & Herzegovina

7

Argentina

15

Japan

16

Croatia

13

Barbados

4

Korea, South

9

Czech Republic

3

Belize

1

Singapore

11

Estonia

3

Bolivia

3

Taiwan

6

Georgia

1

Brazil

15

Hungary

13

Chile

12

Western Europe, Israel

Kosovo

1

Columbia

9

Austria

4

Latvia

9

Costa Rica

3

Belgium

15

Lithuania

4

Dominican Republic

3

Denmark

15

Macedonia

4

Ecuador

7

Finland

16

Montenegro

1

El Salvador

2

France

16

Poland

7

Guatemala

5

Germany

15

Romania

8

Jamaica

9

Greece

12

Russia

11

Mexico

10

Iceland

9

Serbia

3

Panama

5

Ireland

14

Slovak Republic

4

Peru

10

Israel

11

Slovenia

13

Puerto Rico (US)

3

Italy

15

Suriname

2

Luxembourg

2

Middle East, North Africa

Trinidad & Tobago

5

Netherlands

14

Algeria

4

Uruguay

7

Norway

15

Egypt

3

Venezuela

5

Portugal

8

Iran

7

Spain

15

Jordan

2

Asia: Developing

Sweden

13

Lebanon

1

Bangladesh

1

Switzerland

10

Libya

1

China

11

United Kingdom

16

Morocco

1

India

9

Qatar

1

Indonesia

3

North America, Oceania

Saudi Arabia

2

Kazakhstan

2

Australia

10

Syria

1

Malaysia

7

Canada

10

Tunisia

3

Pakistan

3

New Zealand

5

Turkey

7

Philippines

3

United States

16

United Arab Emirates

4

Thailand

8

West Bank/Gaza

3

Tonga

1

Yemen

1

Vanuatu

1

Vietnam

2

Sub-Saharan Africa

Angola

5

Regions/Subgroups

Botswana

3

Azores (Portugal)

1

Burkina Faso

1

Maori Subsample (NZ)

1

Cameroon

1

Scotland (UK)

2

Numbers represent annual surveys between 1999 and 2014.

Ethiopia

1

Shenzhen (China)

2

Ghana

3

Wales (UK)

2

Malawi

2

Namibia

2

Nigeria

3

South Africa

13

Zambia

3

Developing representative samples of adults was a two-stage process. The first step involved selecting a representative selection of households, leading to a contact with an adult member. In countries where a high proportion of households have land-line telephones, this was done by creating a random set of numbers considered to be household phone numbers. In countries with a high proportion of cell-phone-only adults, this was supplemented with random samples of active cell phone numbers. Numbers were then called, generally up to three times, until an adult respondent answered the phone. In countries with a low proportion of households with phones, geographic areas were selected at random for personal contacts by interviewers, who then approached households for a face-to-face interview. In some developing countries, phone interviews are conducted in the major urban areas and supplemented with face-to-face interviews in rural regions.

Adults from each household were selected for interviews in one of two ways. In some cases it was the first adult contacted, and in others a person would be randomly selected from those adults living in the household. In many surveys there was a deliberate attempt (quota sampling) to complete half of all interviews with men and half with women. In most cases, the women’s quota was filled before the men’s.

As response rates at the household and individual level varied widely, all survey operations used post-stratification weighting procedures to develop case weights. This involved comparing the survey sample with the most reliable national statistics on the characteristics of the adult population (gender, age, household income, region, etc.). Case weights were then assigned to the sample population so the characteristics of the sample would match the national population on these characteristics.

In the GEM APS data set, three types of weights are included:

WEIGHT: Original weights provided by the survey research vendor, recentered (adjusted) so that the average value for the sample for each year equals 1,000. (The sum of the weights equals the sum of the cases.)

WEIGHT_L: Original weights adjusted so they are only available for those 18–64 years of age, an estimate of the age at which individuals are assumed to be active in the labor force and the only age range included in all national samples by survey venders.

WEIGHT_A: Original weights adjusted so they are only available for those 18 years of age and older, considered an appropriate range for assessments involving informal investors, many of whom are older and retired from the labor force.

The number of cases with derived weights is reduced from the total sample because of the restriction on age. Most important are the omissions of those less than 18 years of age. In many countries people under 18 are assumed eligible for the labor force and are included in population surveys; to facilitate cross-national comparisons, the age range is standardized to 18 to 64. There are, in addition, a small proportion of cases, about 0.4% or 1 in 250, where the age of the respondent is not available; these cases are excluded from all derived weights.8 This increases confidence that the representativeness of the samples has been harmonized for all countries.

Data Collection: Expert Interviews, Questionnaires

The conceptualization of national characteristics in Figure 1 included both general features assumed to affect all economic activity and those specific to business creation, the entrepreneurial framework conditions. While many standardized cross-national measures of the general characteristics were available, there was limited information about those specific to entrepreneurship. One solution was to identify national experts on entrepreneurship for a unstructured, face-to-face discussion, followed by a brief self-administered questionnaire. The spontaneous responses to the unstructured sections of the discussion provided very useful detailed material for completion of national reports. The responses to the fixed-choice items on the questionnaire have provided a substantial resource for characterizing multiple aspects of the national entrepreneurial context.

The initial design of the questionnaire emphasized nine entrepreneurial framework conditions.

  1. (A) Financial support for new and small firms

  2. (B) Government policies to support new and small firms

  3. (C) Government programs to support new and small firms

  4. (D) Entrepreneurial education and training

  5. (E) Research and development transfer to new and small firms

  6. (F) Commercial, legal infrastructure relevant to new and small firms

  7. (G) Internal [within the country] market openness

  8. (H) Access to physical infrastructure for new and small firms

  9. (I) Cultural, social norms that encouraged new and small firms

Multi-item modules representing these aspects were included in all NES starting with 1999. These nine dimensions affected the data collection in two ways, the selection of the experts and the construction of the questionnaire items.

Assessment of the responses in the early years led to the separation of three of the nine measures into two subscales:

  1. (B.1) Government policies

  2. (B.2) Government regulations and taxes

  3. (D.1) Entrepreneurial emphasis in primary and secondary education

  4. (D.2) Entrepreneurial emphasis in post-secondary education

  5. (G.1) Internal market dynamics

  6. (G.2) Internal market burdens

The result was the creation of twelve entrepreneurial framework conditions for the assessments.

Over the years, modules related to additional national features have been added to these annual assessments, including:

  • Presence of entrepreneurial opportunities in the country

  • Support for social entrepreneurship

  • Availability of skills and capacity to implement new firms in the general population

  • The presence of motivations to create new firms in the general population

  • The presence of conditions that would provide protection for intellectual property rights

  • Support for the participation of women in entrepreneurial initiatives

  • Positive context for high-growth new firms

  • Customers open to innovation and new products and services

  • National context supports immigrant entrepreneurship

  • Emphasis on collaboration among businesses

  • Public support for business collaboration

  • Motivation for youth participation in business creation

  • Public support for youth participation in entrepreneurship

  • Openness to market innovations and new goods and services

These additional topics affected the content of the questionnaire, but not the identification or selection of experts.

Interview and Questionnaire Design

The NES interview was designed as an open discussion of entrepreneurial issues in the country followed by completion of the questionnaire. The initial discussion might last from 30 to 60 minutes and provided a great deal of useful information for the teams in the completion of their national reports. But given that the information was collected in a wide variety of languages, there has been less effort to create harmonized cross-national summaries of the extemporaneous observations of the national experts.

The questionnaire was developed to provide reliable multi-item indices of each entrepreneurial framework condition. Responses to each item are gathered on a five-point scale: “completely true,” “somewhat true,” “neither true nor false,” “somewhat false,” and “completely false.” As an example, the items used to develop the index for physical infrastructure included the following:

  1. 1. In my country the physical infrastructure (roads, utilities, communications, waste disposal) provides good support for new and growing firms.

  2. 2. In my country it is not too expensive for new and growing firms to get good access to communications (phone, internet, etc.).

  3. 3. In my country, a new or growing firm can get good access to communications (phone, internet, etc.) in about a week.

  4. 4. In my country, new and growing firms can afford the cost or basic utilities (gas, water, electricity, sewer).

  5. 5. In my country, new and growing firms can get good access to utilities (gas, water, electricity, sewer) in about a month.

Over the years, the item wording and number of items has been adjusted so that the Cronbach’s measures of reliability are generally between 0.60 and 0.95.9

To provide a basis for comparing the extent to which the national experts were personally involved in business creation, relevant items from the adult population interview were incorporated into the expert self-administered questionnaire. They are generally much more active than those in the representative sample. All national experts were also asked to provide basic socio-demographic information; age, gender, years of work, educational background, etc. Copies of the self-administered questionnaire used from 1999 to 2003 are available in the consolidated codebook for this period (Reynolds, Autio, & Hechavarria, 2008).

Selection of the National Experts

It is, of course, impossible to develop a representative sample of experts in each country. Requiring diversity in this convenience sample, however, helps minimize bias related to the respondent’s experience and background. For these convenience samples, the national teams were asked to identify, using their personal and professional social networks, individuals who could be considered experts in each of the nine entrepreneurial framework domains. These could be individuals from government agencies, universities, consulting firms, the financial sector, mass media, and even experienced, established entrepreneurs. They were encouraged to select individuals from all geographic regions as well as from different industry sectors. They were asked to try to include at least one established entrepreneur or small business person in each of the nine domains. The initial quota consisted of four experts for each of the nine entrepreneurial framework conditions, or a total of 36. In some years, some countries interviewed many more national experts. Germany reported 173 interviews in 2005, and Brazil interviewed 105 in 2014. As no estimates of the structure of the total population of national experts exist, each case is assigned a weight of one in the national analysis.

This is a procedure subject to considerable judgment and reflects the extent to which the national team members had extensive and sophisticated social networks in their home countries. In most countries this was the case, and the initial round of interviews was usually completed successfully. Smaller countries, such as Finland, Singapore, Belgium, or New Zealand, however, found that after the first year there were problems filling the quotas with new respondents. In some cases the quota might be reduced to two for each of the nine entrepreneurial domains, a total of 18 respondents. As some countries were involved in the project for a number of years, they found that the only way to achieve their quota of interviews was to return to the same individuals on subsequent administrations. There is no information available on the extent of these repeated interviews in the data sets.

While the initial strategy was to complete face-to-face interviews with the respondents, this involves some travel expenses in larger countries, such as India, Brazil, or the United States. As a result, a wider range of strategies have been utilized for collecting completed questionnaire data. By 2012 18% of 2,783 questionnaires were completed as part of a face-to-face interview, 36% by e-mail, 24% online, 19% by fax or surface mail, and 2% over the phone.

Following accepted procedures, the identity and contact details of the national experts are not included in the data sets. But there is considerable variation in identifying the panel of experts among countries. Some keep their identity strictly confidential, while others provided lists, even photographs, of the experts as part of their national reports. The general nature of the personal data would suggest that identifying the responses of specific experts would be virtually impossible.

Countries and Number of Respondents

The number of countries and expert questionnaire responses in each annual NES data file are provided in Table 4.10 It should be noted that this does not always refer to the year in which the data was collected, as in some cases the interviews and questionnaires were completed in prior years. In addition, there was not always an independent sample for each year. In later years, in some countries experts responding in early years were contacted and agreed to complete a new interview for the new data cycle year. While this provided a current assessment of the national situation, they will not be completely independent from assessments in prior years.

Table 4. National Expert Questionnaire Sample Sizes

Year

No of Countries

No of Expert Respondents

1999

9

281

2000

31

788

2001

27

991

2002

34

1,300

2003

33

1,325

2004

30

2005

35

1,323

2006

37

2007

31

1,114

2008

30

1,253

2009

45

1,636

2010

54

2,008

2011

49

1,852

2012

69

2,783

2013

69

2014

73

2,823

2015

62

Note: (*) Number of cases in individual respondent data file.

Issues in Analysis

This strategy of using national experts to provide information related to national characteristics has been used in a variety of other cross-national comparisons such as the Global Competitiveness Report (Schwab & Sala-i-Martín, 2015), the World Competitiveness Report (IMD, 2015), Report on Economic Freedom (Heritage Foundation, 2015), and Transparency International Perception of Corruption (Transparency International, 2015). In all cases, standardized interview items are translated into other languages or answered by individuals for whom English is not their first language. There is no way to measure whether or not individuals with different first languages are responding to the items or the response alternatives in the same way. Those comfortable with Brazilian Portuguese may respondent differently than those comfortable with Japanese, Finnish, Russian, Quebec French, or one of the many official languages in India. Nonetheless, major differences in the values of reliable indices provide some evidence of national variation on characteristics. Any empirical evidence is often better than no empirical evidence. Within-country differences over time may be considered with some confidence.

A second issue is the incomplete documentation available on the NES questionnaire data. Extensive details are available on the first five years, from 1999 to 2003 (Reynolds, Autio, & Hechavarria, 2008). Both individual-based and national summary databases are available for most years on the GEM website, but details on the questionnaires are not always available. Fortunately, there has been very little change in the items used for the basic entrepreneurial framework conditions since 2003, and serious analysts can estimate reliability and co-variance among indices with the individual respondent data sets.

Empirical Regularities

The GEM program has provided empirical evidence regarding a number of basic features of business creation, as a national characteristic.

  • There is considerable variation among countries in the prevalence of business creation

  • Within countries there is high year-to-year stability in the prevalence of business creation

  • More highly developed countries have much lower levels of business creation

  • There is a modest, statistically significant relationship between business creation and subsequent economic growth

  • All those in society are involved; the largest proportion of business creation is among the poor in developing countries

  • There is substantial country-to-country variation in entrepreneurial framework conditions

The following overview illustrates these basic empirical patterns.

The diversity in prevalence of business creation is presented in Figure 3. Each bar represents the average TEA Index value in all years for which data was collected from 2000 to 2014 for 102 countries.11 Details are provided for all countries in Appendix B. The lower part of each bar represents the prevalence of nascent entrepreneurs involved in active efforts to create a new firm in the pre-profit stage. The top portion represented the prevalence of those owning and managing a new venture with profits for up to 3.5 years. The level of TEA activity varies from less than 3 per 100 adults to over 30 per 100 adults. Further, this is clearly a continuous variable, with small differences separating most adjacent countries. Explaining the diversity in this important feature of human activity is an attractive challenge for the research community.

There are substantial differences in the ratio between the prevalence of new firm owner-managers and the prevalence of nascent entrepreneurs. A small number might be an indication that a large proportion of nascent ventures do not reach profitability. An effort has been made to explore factors that may account for this country-to-country variation (Bergmann & Stephan, 2013).

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 3. Cross-National Variation in the TEA Index.

Equally significant is the remarkable stability of business creation over time. This is illustrated by comparing the level of activity in 2003 with that in 2014, 11 years later, for 27 countries, shown in Figure 4. The correlation of 0.90 reflects the high level of stability in the relative level of activity and is clearly statistically significant. Assessments of year-to-year patterns indicate that stability in the absolute level of activity is the most common pattern (Reynolds, 2015).

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 4. Stability in the TEA Prevalence Rate: 2003 compared to 2014.

During the initial years of the GEM program, the majority of participating countries were developing countries, particularly those in Western Europe. Early assessments based on several dozen countries seemed to suggest a “U-shaped” relationship with economic development. Countries with very low and very high levels of GDP per capita seemed to have more business creation activity (Wennekers, van Stel, Thurik, & Reynolds, 2005).

Two subsequent changes have indicated a much less complicated relationship. First, as shown in Table 2, GEM coverage has expanded to cover many developing countries in Latin America, the Middle East, and North Africa (MENA), Sub-Saharan Africa, and developing Asia. The GEM coverage is now representative of all global regions.

Second, it is now clear that it is more useful to consider order of magnitude differences in both the level of development as well as the level of business creation activity. This is captured by using a logarithmic transformation on both measures, which converts very skewed (lopsided) distributions into normal or bell-shaped distributions.12 This shifts emphasis from differences of 10%, 20%, or 30% to differences of five or ten times.

With such adjustments, the level of economic development, represented by the average GDP per person from 1995 to 1999, can be considered in relation to the subsequent level of business creation, averaged over the years 2000 to 2014. The resulting relationship, as shown in Figure 5 for 96 countries, appears to be a uniform downward slope, a negative correlation of −0.62. There is no “U-shape” in the pattern. Poorer countries have much more business creation than richer countries.

But the range related to both measures is considerable. GDP per capita has been adjusted for Purchasing Power Parity and varies from $300 per year for Ethiopia and Malawi to almost $100,000 per year for the United Arab Emirates and Qatar. The variation in participation in business creation varies from 3 per 100 adults in Japan and Belgium to over 30 per 100 in Uganda, Malawi, Nigeria, and Ghana. There is little systematic difference in countries where the annual GDP per capita varies by 10% or 20%; major differences in GDP per capita are associated with major differences in business creation activity. For Germany the GDP per capita was $26,359 from 1995 to 1999, and the TEA prevalence rate for 2000 to 2014 was 5.0 per 100. Argentina had about half the GDP per capita at $11,519 and a TEA prevalence rate of 14.1 per hundred, almost three times that of Germany. Thailand had a GDP per capita of $6,777, about half that of Argentina, and a TEA prevalence of 21.6 per 100, 50% higher than Argentina. And Nigeria, with a GDP per capita of $2,064, had a TEA prevalence of 37.8 per 100, 75% higher than Thailand. Clearly, order of magnitude differences make a difference. Small differences are probably confounded by errors of measurement.

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 5. Economic Development and Prevalence of Business Creation.

Estimates of a relationship between business creation and subsequent economic growth require both a lag between measures and a sample large enough to examine the relationship. Only recently has it been possible to produce such assessments with the GEM data. The average TEA prevalence over five years (2006 to 2010) and the relationship to the average annual growth in GDP per capita from 2011 to 2015 for 77 countries are presented in Figure 6. While the positive correlation of 0.32 is rather low, it does indicate that business creation may be important to economic growth. It should be noted that the low correlation reflects a number of countries that achieved economic growth despite low levels of business creation, located in the lower right corner of Figure 6—countries like the United Kingdom, Singapore, and Romania. Except for Venezuela, which experienced low growth from 2011 to 2015, reflecting a drop in oil revenue and considerable internal dislocations, there are no countries in the upper left portion of Figure 6. A combination of high levels of business creation and low economic growth is quite rare.

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 6. Business Creation and Subsequent Economic Growth.

The most obvious conclusion from the relationship in Figure 6 is that higher levels of business creation are generally associated with national economic growth, complementing other processes that may have an impact, as illustrated in the conceptual model in Figure 1. An active entrepreneurial sector may well act as a catalyst, with nascent ventures serving as an intermediary that facilitates the introduction of new productive processes delivering new goods and services.

Participation by those in different economic circumstances in business creation is a major issue. Entrepreneurship is often presented as an opportunity for the well-educated with access to financial support, such as students pursing entrepreneurship under the sponsorship of elite MBA programs. This implies that those with modest human and financial capital would not be involved. The expansion of the GEM program to cover more developing countries facilitates estimates of the total amount of participation, not just the prevalence rate, in eight world regions. Based on data collected from 2000 to 2012, it is estimated that 446 million were active as nascent entrepreneurs or new firm owner-managers in 2012. The distribution of activity based on daily income and world region are presented in Figure 7.

While in developed countries the largest proportion of TEA active individuals are in wealthier households, in the developing world those from modest circumstances dominate. In fact, over half of the total of 422 million are individuals in developing countries from poor circumstances. While most of the ventures initiated by individuals in the lowest 40% in terms of daily income are small scale, many are substantial and their ventures account for over 20% of job creation, 35% of recent technology ventures, 50% of those expecting to affect markets, and over 20% of exports (Reynolds, 2012, chapter 5). This mass of business creation is probably a major factor in facilitating economic advancement in developing countries.

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 7. Total Business Creation Activity by Global Region and Household Income.

The questionnaires completed during the National Expert Surveys provide assessments of the national entrepreneurial framework conditions. The pattern of responses among 69 countries for 2014 is illustrated in Figure 8. The 12 dimensions are ranked from left to right by the average value: five represents the most positive assessment, one the most negative, and three is a neutral response. The horizontal bars represent the average rating; the range of responses is represented by the vertical bar.

Global Entrepreneurship Monitor (GEM) Program: Development, Focus, and ImpactClick to view larger

Figure 8. National Entrepreneurial Framework Conditions: 2014 Average and Range of Values.

The average rating of the physical infrastructure relevant to new and growing firm creation is seen as most supportive. Coverage of firm creation in the primary and secondary school systems is considered the least supportive of entrepreneurship. There is considerable diversity across the countries in the ratings of all 12 conditions. This diversity will facilitate assessments of the relative impact of the national conditions on the amount and type of entrepreneurial activity. One note of caution: the countries with the most supportive entrepreneurial framework conditions are those with the highest levels of economic development. Yet these developed countries are the ones with the lowest levels of business creation activity. This suggests that the relationship may be more complicated than suggested by Figure 1.

The GEM research program has created data sets that provide a unique and detailed overview of business creation as a basic activity in all countries. Entrepreneurship is an important part of the mechanisms that lead to economic growth and absorb the attention and energy of hundreds of millions. The GEM data sets provide multiple opportunities for expanding understanding of factors affecting business creation and its consequences.

Impact, Scope of Assessment

The GEM research program has been the basis for a considerable amount of publication for two audiences. The largest volume of work has been the annual global and national reports oriented toward policymakers. Prepared each year since 1999 and available on the GEM project website, these have been widely disseminated and utilized in policy discussions. In addition, as each national team has generally produced annual country reports, there are hundreds of these documents, often in the language of the country. These are also available on the GEM project website. Further, many regional reports have been prepared, such as for Scotland and Wales in the United Kingdom and many of the individual regions of Spain. In short, no project related to the study of entrepreneurship has come close to generating such a massive body of policy relevant assessments.

The second audience has been the scholarly community. There have been several attempts to summarize this work. One assessment considered 109 peer-reviewed journal articles published up to the end of 2010 (Bergmann et al., 2014). The first articles that could be located appeared in 2003, and there was a steady flow of 15–25 articles per year appearing in a wide range of scholarly journals; 58 appeared in Social Science Citation Index (SSCI) journals and 51 in others, mostly located in developing countries. Given the expanding breadth of GEM and the benefits of longitudinal national assessments, this annual flow has probably increased. This initial assessment found work on both the micro-level, where the individual or business venture was the unit of analysis, and the macro-level, which might emphasize the prevalence of business creation or growth in national gross domestic product (GDP).

An effort to track the intellectual development of the GEM effort began by identifying 86 GEM-related assessments in the SSCI in January 2013 (Ramos-Rodríguez et al., 2015). This set of 86 articles had references to a total of 3,347 documents, 118 of which were cited five or more times in the original GEM articles. Tracking the intellectual background involved identifying which pairs of references were cited in the same GEM assessments. The 16 core documents in this network of citations are listed in Table 5. These documents have eight or more co-citations among the 86 GEM-based assessments. The number of citations among the 86 GEM documents is provided in the first column.

Three of the 16, by Baumol (1990), North (1990), and Schumpeter (1934), were well known years before the GEM project was in place. The lines between references reflect the fact that the two articles are both referenced more than seven times in the GEM articles. The article at the center (Reynolds et al., 2005) was an overview of the APS data collection procedures developed by the coordination team. This analysis provides a guide to the academic work that is highly cited in the development of GEM assessments.

Both assessments illustrate the expansion of scholarly work utilizing the GEM data. As more countries are represented by the project and the time series are extended, there will be further expansion of the use of these resources for policy and academic issues.

Table 5. Intellectual Background Documents in Published GEM Assessments

Cits

Document

46

Reynolds, P., Bosma, N., Autio, E., Hunt, S., DeBono, N., Servais, I., . . .Chin, C. (2005). Global Entrepreneurship Monitor: Data collection design and implementation: 1998–2003. Small Business Economics, 24, 205–231.

26

Wennekers, S., van Stel, A., Thurik, R., & Reynolds, P. (2005). Nascent entrepreneurship and the level of economic development. Small Business Economics, 24(3), 293–309.

19

Davidsson, P., & Honig, B. (2003). The role of social and human capital among nascent entrepreneurs. Journal of Business Venturing, 18(3), 301–331.

18

Arenius, P., & Minniti, M. (2005). Perceptual variables and nascent entrepreneurship. Small Business Economics, 24(3), 233–247.

18

Wennekers, S., & Thurik, R. (1999). Linking entrepreneurship and economic growth. Small Business Economics, 13(1), 27–55.

18

Shane, S., & Venkataraman, S. (2000). The promise of entrepreneurship as a field of research. Academy of Management Review, 25(1), 217–226.

17

North, D. (1990). Institutions, institutional change, and economic performance. Cambridge, U.K.: Cambridge University Press.

17

Reynolds, P., Bygrave, W., Autio, E., Cox, L., & Hay, M. (2002). Global Entrepreneurship Monitor: 2001 executive report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.

15

Carree, M., van Stel, A., Thurik, R., & Wennekers, S. (2002). Economic development and business ownership: An analysis using data of 23 OECD countries in the period 1976–1996. Small Business Economics, 19(3), 271–290.

15

Sternberg, R., & Wennekers, S. (2005). Determinants and effects of new business creation using Global Entrepreneurship Monitor data. Small Business Economics, 24(3), 193–203.

14

Baumol, W. (1990). Entrepreneurship: Productive, unproductive, and destructive. Journal of Political Economy, 98(5), 893–921.

14

Schumpeter, J. (1934). The theory of economic development: An inquiry into profits, capital, credit, interest, and the business cycle. New Brunswick, NJ: Transaction Books. Translated from 1911 original in German by R. Opie.

13

Bosma, N., Jones, K., Autio, E., & Levie, J. (2008) Global Entrepreneurship Monitor: 2007 executive report. Babson Park, MA: Babson College.

13

Acs, Z., Arenius, P., Hay, M., & Minniti, M. (2005). Global Entrepreneurship Monitor: 2004 executive report. Babson Park, MA: Babson College.

13

Wong, P., Ho, Y., & Autio, E. (2005). Entrepreneurship, innovation and economic growth: Evidence from GEM data. Small Business Economics, 24(3), 335–350.

12

Minniti, M., Bygrave, W., & Autio, E. (2006). Global Entrepreneurship Monitor: 2005 executive report. Babson Park, MA: Babson College.

Note: (*) Number of times cited among the 86 GEM articles.

Source: Documents co-cited in GEM assessments eight or more times, Ramos-Rodríguez et al. (2015, Fig. 6).

Future Opportunities

The major data collection objective of the GEM program, to provide a harmonized comparison of the national prevalence of business creation, appears to have been realized. No other data set now exists that can provide standardized information on over 100 countries that represent well over 90% of the global population. The benefits to the national teams appear to be substantial, suggesting that there will continue to be enough participants for the GEM program to continue for some time, perhaps decades.

As a vigorous entrepreneurial sector is an important feature that facilitates economic growth, it should remain as one of the central concerns in national policymaking. Entrepreneurship does not provide any economic contribution—jobs, new goods and services, exports—until a new venture is created. The major advantage of the GEM protocol for the APS surveys is provision of estimates of the implementation of new ventures. It is the creation of new businesses that provides social benefits of interest to policymakers and politicians.13 The expansion of scholarly and academic interest in entrepreneurship has grown substantially since the initiation of the GEM program, which should lead to greater use of this unique resource. The GEM policy of making the APS and NES survey data generally available, at no cost, three years after it is collected provides the wider research community with access to a substantial, expanding resource.

One of the major benefits of this data for scholarly research is the combination of standardized measures of perceptions, attitudes, and behaviors related to business creation across the most diverse contexts in the modern world. There is no comparable source of data that has harmonized measures across different political and social contexts. If variation in context affects business creation, it should be apparent in assessments based on the GEM data.

References

Acs, Z., Szerb, L., & Autio, E. (2015). Global entrepreneurship and development index: 2016. New York: Springer Briefs in Economics.Find this resource:

Baumol, W. (1990). Entrepreneurship: Productive, Unproductive, and Destructive. Journal of Political Economy, 98(5), 893–921.Find this resource:

Bergmann, H., & Stephan, U. (2013). Moving on from nascent entrepreneurship: Measuring cross-national differences in the transition to new business ownership. Small Business Economics, 41(4), 945–959.Find this resource:

Bergmann, H., Mueller, S., & Schrettle, T. (2014). The use of global entrepreneurship data in academic research: A critical inventory and future potentials. International Journal of Entrepreneurial Venturing, 6(3), 242–276.Find this resource:

Bosma, N., & Levie, J. (2010). Global Entrepreneurship Monitor: 2009 global report. Retrieved from http://www.gemconsortium.org/report.

Bosma, N., Wennekers, S., & Amorós, J. E. (2012). Global Entrepreneurship Monitor: 2011 extended report: Entrepreneurs and entrepreneurial employees across the globe. Retrieved from http://www.gemconsortium.org/report.

Gartner, W. B., Shaver, K. G., Carter, N. M., & Reynolds, P. D. (Eds.). (2004). Handbook of entrepreneurial dynamics: The process of business creation. Thousand Oaks, CA: SAGE.Find this resource:

Heritage Foundation. (2015). Index of economic freedom 2015. Washington, DC: Heritage Foundation. Retrieved from http://www.heritage.org/index/pdf/2015/book/index_2015.pdf.Find this resource:

IMD. (2015). IMD world competitiveness yearbook. Lausanne, Switzerland: IMD World Competitiveness Center.Find this resource:

International Monetary Fund. (2016). World economic outlook database.Find this resource:

North, D. (1990). Institutions, institutional change, and economic performance. Cambridge, U.K.: Cambridge University Press.Find this resource:

Ramos-Rodríguez, A. R., Martínez-Fierro, S., Medina-Garrido, J. A., & Ruiz-Navarro, J. (2015). Global Entrepreneurship Monitor versus panel study of entrepreneurial dynamics: Comparing their intellectual structures. International Entrepreneurship and Management Journal, 11, 571–597.Find this resource:

Reynolds, P., Bosma, N., Autio, E., Hunt, S., De Bono, N., Servais, I., . . . Chin, C. (2005). Global Entrepreneurship Monitor: Data collection design and implementation: 1998–2003. Small Business Economics, 24, 205–231.Find this resource:

Reynolds, P. D. (2000). National panel study of U.S. business start-ups: Background and methodology. In J. A. Katz (Ed.), Advances in entrepreneurship, firm emergence and growth (Vol. 4, pp. 153–228). Stamford, CT: JAI Press.Find this resource:

Reynolds, P. D. (2016). Global Entrepreneurship Monitor [GEM]: Adult population survey data sets: 1998–2012.

Reynolds, P. D., Autio, E., & Hechavarria, D. (2008). Global Entrepreneurship Monitor [GEM]: Expert questionnaire data, 1999–2003.

Reynolds, P. D., Bygrave, W. D., Autio, E., Cox, L., & Hay, M. (2002.) Global Entrepreneurship Monitor: 2002 executive report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.Find this resource:

Reynolds, P. D., Bygrave, W. D., Autio, E., & Hay, M. (2002). Global Entrepreneurship Monitor: 2002 summary report. Boston: Babson College.Find this resource:

Reynolds, P. D., Bygrave, W. D., Autio, E., Arenius, P., Fitzsimons, P., Minneti, M., & others. (2004). Global Entrepreneurship Monitor: 2003 Executive report. Babson Park, MA: Babson College.Find this resource:

Reynolds, P. D., & Curtin, R. T. (Eds.). (2010). New firm creation: An international overview. New York: Springer.Find this resource:

Reynolds, P. D., & Curtin, R. T. (Eds.). (2009). New firm creation in the U.S.: Initial explorations with the PSED II data set. New York: Springer.Find this resource:

Reynolds, P. D., Hay, M., Bygrave, W. D., Camp, S. M., & Autio, E. (2000). Global Entrepreneurship Monitor: 2000 executive report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.Find this resource:

Reynolds, P. D., Camp, S. M., Bygrave, W. D., Autio, E., & Hay, M. (2001). Global Entrepreneurship Monitor: 2001 research report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.Find this resource:

Reynolds, P. D., Camp, S. M., Bygrave, W. D., Autio, E., & Hay, M. (2001). Global Entrepreneurship Monitor: 2001 executive report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.Find this resource:

Reynolds, P. D., Hay, M., & Camp, M. (1999). Global Entrepreneurship Monitor: 1999 executive report. Kansas City, MO: Kauffman Center for Entrepreneurial Leadership.Find this resource:

Reynolds, P. D. (2012). Entrepreneurship in developing economies: The bottom billions and business creation. Foundations and Trends in Entrepreneurship, 8(3), 141–277.Find this resource:

Reynolds, P. D. (2015). Business creation stability: Why is it so hard to increase entrepreneurship? Foundations and Trends in Entrepreneurship, 10(5–6), 321–475.Find this resource:

Schumpeter, J. (1934). The Theory of Economic Development: An inquiry into profits, capital, credit, interest, and the business cycle. New Brunswick, NJ: Transaction Books. Translated from 1911 original in German by R. Opie.Find this resource:

Schwab, K., & Sachs, J. (1987). The global competitiveness report: 1987. Geneva, Switzerland: World Economic Forum.Find this resource:

Schwab, K., & Sachs, J. (1997). The global competitiveness report: 1997. Geneva, Switzerland: World Economic Forum.Find this resource:

Schwab, K., & Sachs, J. (1998). The global competitiveness report: 1998. Geneva, Switzerland: World Economic Forum.Find this resource:

Schwab, K., & Sala-i-Martín, X. (2015). The global competitiveness report: 2015–2016. Geneva, Switzerland: World Economic Forum. Retrieved from http://reports.weforum.org/global-competitiveness-report-2015-2016/.Find this resource:

Transparency International. (2015). Corruption perceptions index 2015. Berlin: Transparency International. Retrieved from https://www.transparency.org/cpi2015/results.Find this resource:

United Nations, Department of Economic and Social Affairs, Statistical Division. (2008). International standard classification of all economic activities: Revision 4.0. New York: United Nations.Find this resource:

United Nations, Statistical Office of, Department of Economic and Social Affairs. (1990). International standard classification of all economic activities: Revision 3.1. New York: United Nations.Find this resource:

United States Census Bureau. International data base (IDB). Retrieved from http://www.census.gov/population/international/data/idb.

Wennekers, S., van Stel, A., Thurik, R., & Reynolds, P. (2005) Nascent Entrepreneurship and the Level of Economic Development. Small Business Economics, 24(3), 293–309.Find this resource:

Appendix A: Annual Sample Sizes by Country

Country

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

All yrs

Algeria

2,000

3,427

4,995

2,500

12,922

Angola

1,518

2,167

2,636

2,146

2,137

10,604

Argentina

2,000

1,992

1,999

2,004

2,003

2,008

2,007

2,018

2,031

2,008

2,001

2,000

2,018

2,200

2,500

30,789

Australia

2,089

2,072

3,378

2,210

1,991

2,465

2,518

2,000

2,000

2,177

22,900

Austria

2,197

2,002

4,583

4,586

13,368

Azores (Portugal)

1,010

1,010

Bangladesh

2,000

2,000

Barbados

2,928

2,055

2,355

2,000

9,338

Belgium

2,029

2,038

4,057

2,184

3,879

4,047

2,001

2,028

1,997

3,989

2,000

1,852

2,010

2,001

2,004

38,116

Belize

2,084

2,084

Bolivia

2,000

3,524

2,590

8,114

Bosnia & Herzegovina

2,028

2,000

2,000

2,277

2,001

2,004

2,015

14,325

Botswana

2,003

2,204

2,156

6,363

Brazil

1,999

2,000

2,000

2,000

4,000

2,000

2,000

2,000

2,000

2,000

2,000

2,000

10,000

10,000

10,000

55,999

Burkina Faso

2,850

2,850

Cameroon

2,087

2,087

Canada

1,004

1,003

2,003

1,939

3,014

1,943

2,004

6,412

2,038

3,286

2,479

27,125

Chile

2,016

1,992

1,992

2,007

4,008

2,000

5,000

7,195

7,195

2,420

6,703

6,212

48,740

China

2,054

1,607

2,109

2,399

2,666

3,608

3,677

3,690

3,684

3,634

3,647

32,775

Colombia

2,001

2,102

2,001

2,055

11,029

10,374

6,471

3,400

3,691

43,124

Costa Rica

2,003

2,041

2,057

6,101

Croatia

2,001

2,000

2,016

2,000

2,000

2,000

1,996

2,000

2,000

2,000

2,000

2,000

2,000

26,013

Czech Republic

2,001

2,005

5,009

9,015

Denmark

1,002

2,005

2,022

2,009

2,008

2,009

2,010

10,000

2,001

2,012

2,012

1,957

2,015

2,217

2,008

37,287

Dominican Republic

2,081

2,019

2,007

6,107

Ecuador

2,010

2,142

2,200

2,077

2,004

2,030

2,040

14,503

Egypt

2,636

2,769

2,501

7,906

El Salvador

2,180

2,014

4,194

Estonia

2,004

2,004

2,357

6,365

Ethiopia

3,005

3,005

Finland

1,000

1,000

2,002

2,001

2,005

2,005

2,000

2,010

2,005

2,005

2,011

2,004

2,006

2,011

2,038

2,005

2,005

32,113

France

1,000

2,012

1,991

2,029

2,018

1,953

2,005

1,909

2,005

2,018

2,019

2,012

2,009

4,003

2,002

2,005

32,990

Georgia

2,016

2,016

Germany

997

1,008

2,010

7,058

15,041

7,534

7,523

6,577

4,049

4,751

6,032

5,552

4,260

4,300

5,996

4,311

86,999

Ghana

2,447

2,222

2,100

6,769

Greece

2,000

2,008

2,000

2,000

2,000

2,000

2,000

2,000

2,000

2,000

2,000

2,000

24,008

Guatemala

2,190

2,285

2,398

2,142

2,158

11,173

Hong Kong

2,000

2,000

2,004

2,058

2,000

10,062

Hungary

2,000

2,000

2,878

2,878

2,500

1,500

2,001

2,000

2,000

2,002

2,000

2,000

2,003

27,762

Iceland

2,000

2,011

2,002

2,002

2,001

2,002

2,002

2,005

2,001

18,026

India

2,002

2,011

3,047

1,999

1,662

2,032

2,700

3,000

3,360

21,813

Indonesia

2,000

4,500

5,520

12,020

Iran

3,124

3,350

3,359

3,352

3,178

3,637

3,352

23,352

Ireland

2,021

1,971

2,000

1,971

1,978

1,945

2,008

2,007

2,001

2,000

2,002

2,000

2,002

2,000

27,906

Israel

992

2,621

2,055

2,004

1,933

2,019

2,030

2,073

2,007

2,007

2,039

21,780

Italy

1,000

1,998

1,973

2,002

1,967

2,945

1,984

1,999

2,000

3,000

3,000

3,000

2,000

2,052

2,000

32,920

Jamaica

2,166

3,669

2,407

2,012

2,298

2,047

2,003

2,246

2,637

21,485

Japan

1,383

1,249

2,000

1,999

1,977

1,917

1,990

2,000

1,860

2,001

1,600

2,006

2,004

2,010

2,000

2,006

30,002

Jordan

2,000

2,006

4,006

Kazakhstan

2,000

2,099

4,099

Korea, South

2,003

2,008

2,015

2,000

2,000

2,001

2,001

2,000

2,000

18,028

Kosovo

2,000

2,000

Latvia

1,964

1,958

2,000

2,011

2,003

2,001

2,000

2,000

2,000

17,937

Lebanon

2,000

2,000

Libya

2,246

2,246

Lithuania

2,003

2,003

2,000

2,000

8,006

Luxembourg

2,005

2,074

4,079

Macedonia

2,000

2,002

2,003

2,000

8,005

Malawi

2,006

2,094

4,100

Malaysia

2,005

2,002

2,010

2,053

2,006

2,000

2,000

14,076

Maori Subsample (NZ)

1,000

1,000

Mexico

2,014

1,002

2,009

2,015

2,605

2,605

2,511

2,516

2,801

2,587

22,665

Montenegro

2,000

2,000

Morocco

1,500

1,500

Namibia

1,959

2,000

3,959

Netherlands

2,013

3,510

3,505

3,507

3,582

3,535

3,539

3,508

3,003

3,502

3,500

3,501

3,005

2,260

45,470

New Zealand

1,948

2,000

1,969

1,933

968

8,818

Nigeria

2,080

2,651

2,604

7,335

Norway

2,041

2,874

2,036

2,040

2,883

2,014

1,999

1,996

2,049

2,029

2,002

2,001

2,000

2,000

2,000

31,964

Pakistan

2,007

2,002

2,000

6,009

Panama

2,000

2,001

1,998

2,004

2,005

10,008

Peru

2,007

1,997

2,000

2,052

2,021

2,108

2,010

2,071

2,075

2,078

20,419

Philippines

2,000

2,500

2,000

6,500

Poland

2,000

2,000

2,001

2,000

2,003

2,000

2,001

14,005

Portugal

2,000

1,000

2,023

2,002

2,011

2,001

2,003

2,005

15,045

Puerto Rico (US)

1,998

2,000

2,000

5,998

Qatar

4,272

4,272

Romania

2,046

2,206

2,093

2,235

2,028

2,004

2,021

2,001

16,634

Russia

2,012

2,190

1,894

1,939

1,660

1,695

1,736

7,500

3,541

2,029

2,001

28,197

Saudi Arabia

2,000

2,000

4,000

Scotland (UK)

2,056

2,118

4,174

Serbia

2,200

2,297

2,300

6,797

Shenzhen (China)

2,040

2,000

4,040

Singapore

2,120

2,004

2,005

1,874

3,852

3,876

4,011

2,000

2,001

2,000

2,006

27,749

Slovak Republic

2,000

2,000

2,007

2,000

8,007

Slovenia

2,030

2,012

2,003

3,016

3,008

3,020

3,019

3,030

3,012

2,009

2,010

2,002

2,004

32,175

South Africa

5,274

6,993

3,262

3,252

3,237

3,248

3,270

3,135

3,279

3,178

2,928

3,450

3,789

48,295

Spain

2,019

2,016

2,000

7,000

16,980

19,384

28,306

27,880

30,879

28,888

26,388

17,500

21,900

24,600

25,000

280,740

Suriname

2,290

2,200

4,490

Sweden

2,036

2,056

1,999

2,025

26,700

2,002

2,003

2,001

2,492

3,101

2,500

2,506

2,508

53,929

Switzerland

2,001

2,003

5,456

2,148

2,024

2,002

2,000

2,003

2,003

2,426

24,066

Syria

2,002

2,002

Taiwan

2,236

2,001

2,012

2,009

2,007

2,000

12,265

Thailand

1,043

2,000

2,000

2,000

2,000

3,000

2,362

2,059

16,464

Tonga

1,184

1,184

Trinidad & Tobago

2,016

2,008

2,029

2,036

2,004

10,093

Tunisia

2,000

2,001

2,000

6,001

Turkey

2,417

2,400

2,400

2,401

2,401

2,401

33,287

47,707

Uganda

1,015

2,005

2,095

2,267

2,343

2,513

2,112

14,350

United Arab Emirates

2,001

2,180

2,056

3,029

9,266

United Kingdom

1,013

1,014

2,032

5,398

16,002

21,666

24,006

10,894

43,033

41,829

8,000

30,003

10,403

10,573

11,191

11,017

10,750

258,824

United States

1,004

1,018

2,006

2,954

7,059

9,036

2,007

1,992

3,093

2,166

5,249

5,002

4,000

5,863

5,542

5,698

3,273

66,962

Uruguay

1,997

2,000

2,027

2,001

2,034

2,074

2,016

2,010

2,006

18,165

Vanuatu

1,182

1,182

Venezuela

2,000

2,000

1,794

1,693

2,000

9,487

Vietnam

2,000

2,000

4,000

Wales (UK)

2,010

3,007

5,017

West Bank/Gaza

2,080

1,992

2,000

6,072

Yemen

2,065

2,065

Zambia

2,039

2,157

2,099

6,295

All Countries

5,018

10,420

46,363

76,819

116,776

102,878

145,189

118,191

171,631

155,183

134,990

183,074

184,102

171,297

207,582

244,471

210,584

2,284,568

Appendix B: National Level of Business Creation

Country

Annual Surveys

Total Cases

Nascent Entrepreneur Prevalence

New Firm Owner-Manager Prevalence

Established Firm Owner-Manager Prevalence

TEA Index Prevalence

Algeria

4

12,874

7.1

4.6

4.4

11.6

Angola

5

10,188

15.4

11.1

11.9

26.0

Argentina

15

26,316

9.0

5.9

9.6

14.4

Australia

11

20,760

6.7

5.1

9.4

11.3

Austria

4

13,327

4.5

2.0

7.6

6.4

Bangladesh

1

1,932

7.1

7.1

12.8

12.8

Barbados

4

8,652

11.4

5.4

10.3

16.7

Belgium

15

35,309

2.8

1.3

3.9

4.0

Belize

1

1,958

3.5

2.3

6.6

5.8

Bolivia

3

7,993

23.8

11.1

17.9

32.7

Bosnia & Herzegovina

7

13,882

4.8

3.2

6.3

8.0

Botswana

3

6,353

16.4

10.9

7.3

26.3

Brazil

15

55,834

6.3

9.1

11.9

15.0

Burkina Faso

1

2,850

12.8

10.8

20.7

22.9

Cameroon

1

2,087

24.0

12.7

19.0

34.4

Canada

9

21,028

6.1

3.6

6.9

9.2

Chile

12

42,922

11.4

6.7

7.5

17.6

China

11

32,774

7.1

9.7

13.4

16.4

Colombia

9

43,102

12.9

9.9

10.5

22.2

Costa Rica

3

6,101

10.3

3.5

3.9

13.5

Croatia

13

22,348

4.9

1.7

3.7

6.5

Czech Republic

3

8,642

5.7

2.3

5.9

7.5

Denmark

14

35,143

3.0

2.4

4.9

5.2

Dominican Republic

3

6,101

10.5

8.5

9.8

18.4

Ecuador

7

14,084

16.5

10.6

17.1

25.8

Egypt

1

2,602

9.0

4.2

6.7

13.1

El Salvador

2

3,939

10.0

8.0

12.7

17.6

Estonia

3

5,498

7.2

4.0

8.0

11.0

Ethiopia

1

3,005

5.6

9.8

12.5

15.2

Finland

15

27,711

3.4

2.5

8.7

5.8

France

15

25,278

3.7

1.2

2.6

4.8

Georgia

1

1,648

4.0

2.9

8.5

6.8

Germany

14

75,316

3.3

2.0

4.9

5.0

Ghana

3

6,759

13.3

21.6

36.5

33.6

Greece

12

23,970

4.8

2.7

13.8

7.4

Guatemala

5

11,164

10.8

7.0

7.3

17.0

Hong Kong

5

8,599

2.8

2.2

3.4

4.9

Hungary

13

27,682

4.5

2.6

6.0

7.0

Iceland

9

16,837

8.3

3.9

8.2

11.9

India

9

21,428

8.1

3.4

9.6

11.3

Indonesia

3

12,018

7.1

13.7

23.4

19.9

Iran

7

23,342

8.7

5.4

10.2

13.8

Ireland

14

25,331

4.7

3.2

7.6

7.7

Israel

10

18,767

4.0

2.9

4.0

6.7

Italy

14

29,745

3.2

1.6

4.6

4.7

Jamaica

9

21,084

14.0

8.3

11.2

21.7

Japan

15

27,215

2.0

1.4

6.4

3.3

Jordan

2

3,995

8.4

6.6

13.8

14.5

Kazakhstan

2

4,099

6.8

5.8

7.5

12.2

Korea, South

9

17,532

4.1

5.8

11.2

9.8

Kosovo

1

1,784

2.7

1.1

5.1

3.7

Latvia

9

17,937

5.7

3.7

6.8

9.3

Lebanon

1

2,000

7.1

8.6

18.1

15.1

Libya

1

2,246

6.6

4.4

6.9

10.8

Lithuania

4

8,006

5.6

4.5

8.8

10.0

Luxembourg

2

4,079

5.0

1.8

4.8

6.8

Macedonia

4

7,751

6.0

3.7

8.6

9.6

Malawi

2

3,941

14.4

19.2

15.0

31.5

Malaysia

7

14,075

2.9

4.2

7.4

7.0

Mexico

10

21,887

9.6

3.2

3.6

12.6

Montenegro

1

2,000

12.0

3.3

7.6

15.1

Morocco

1

1,500

7.5

9.3

16.7

16.3

Namibia

2

3,897

15.5

9.8

8.1

24.8

Netherlands

14

34,714

3.7

3.1

7.5

6.7

New Zealand

5

7,300

9.3

6.8

10.6

15.1

Nigeria

3

7,312

23.8

14.4

17.5

37.8

Norway

15

26,892

4.2

3.5

7.1

7.4

Pakistan

3

6,009

10.2

2.2

3.8

12.2

Panama

5

10,008

11.0

4.8

5.0

15.6

Peru

10

19,923

23.0

7.6

10.4

29.0

Philippines

3

6,500

7.8

10.8

13.2

18.3

Poland

7

13,400

5.1

3.0

6.4

8.0

Portugal

8

15,003

4.5

2.9

6.9

7.2

Qatar

1

4,272

11.2

5.2

4.8

16.1

Romania

8

14,219

4.8

2.9

4.4

7.5

Russia

11

27,802

2.6

2.0

2.4

4.4

Saudi Arabia

2

3,838

4.7

2.6

4.1

7.2

Serbia

3

5,345

6.4

3.3

7.1

9.7

Singapore

11

27,217

4.3

2.9

3.4

7.0

Slovak Republic

4

8,007

7.2

3.8

8.3

10.9

Slovenia

13

31,183

3.1

1.8

5.7

4.9

South Africa

13

41,454

4.6

2.5

2.3

6.9

Spain

15

279,162

3.5

2.7

7.2

6.0

Suriname

2

4,092

2.2

0.6

6.2

2.8

Sweden

13

46,693

2.9

2.1

6.6

4.9

Switzerland

10

20,873

3.8

3.1

9.0

6.7

Syria

1

2,002

3.9

4.6

7.7

8.4

Taiwan

6

12,006

3.6

4.1

9.0

7.6

Thailand

8

16,406

10.2

12.4

24.7

21.6

Tonga

1

1,046

8.1

10.7

2.8

18.6

Trinidad & Tobago

5

9,001

10.6

7.3

9.7

17.6

Tunisia

3

5,995

4.2

4.3

7.6

8.5

Turkey

7

47,365

6.0

3.7

7.9

9.6

Uganda

7

14,285

11.9

21.9

29.3

32.0

United Arab Emirates

4

9,016

5.5

3.5

3.6

8.7

United Kingdom

15

211,527

3.5

3.0

6.2

6.3

United States

15

49,792

7.9

3.9

6.9

11.3

Uruguay

9

14,677

9.6

4.4

7.0

13.8

Vanuatu

1

1,112

34.3

27.7

25.7

54.7

Venezuela

5

8,893

16.8

6.2

6.2

22.1

Vietnam

2

4,000

3.1

12.3

19.7

15.3

West Bank/Gaza

3

6,072

6.0

4.0

4.5

9.8

Zambia

3

6,265

23.0

15.8

12.7

37.8

Totals

686

2,068,906

Average values

8.1

6.0

9.3

13.6

Appendix C: Access to Data Sets

The primary source of GEM data and documentation will continue to be the project website (www.gemconsortium.org). National GEM team members have access to all their own country’s data in the year it is created. If they wish, they may share this data with other GEM teams. National GEM team members have access to all data one year after it is collected when they are participating. All APS and NES are made public three years after it was collected.

Data sets are prepared for both the Adult Population Surveys (APS) and National Expert Surveys (NES) in two forms. One provides data files for all individual respondents. The second consists of country-level summaries of major variables and transformations. These are provided for most years from 1999 to the present on the GEM website. In an effort to ensure cross-year standardization, the GEM coordination team has applied the same procedures for identifying nascent entrepreneurs, new firm owner-managers, and established firm owner-mangers or creating multi-item indices for each year of data. Users are responsible for consolidating annual files into a single cross-year aggregated data set.

An alternative strategy for cross-year harmonization is to assemble files for all years and ensure standardization of responses to each individual item across all years. Once this is completed, a standardized procedure for identifying respondents that would qualify as nascent entrepreneurs, new firm owner-managers, and established firm owner-mangers and creating multi-item indices is applied to the entire file. A single harmonized data file represents all countries and all years of data. This has been completed for 1,827,513 respondents participating in 563 national samples and 6 specialized regional samples in APS surveys representing 100 countries from 1998 to 2012. It is available as a public data set on Research Gate (Reynolds, 2016). The 500-page code book includes annual frequency distributions for all items and transformations, coding categories for industry sectors, and all APS interview schedules (in English).

A similar data set is provided for the early years of the National Expert Survey questionnaires. This cross-year harmonized data set covers 1999 to 2003 (Reynolds, Autio, & Hechavarria, 2008). The 300-page code book for this data set includes frequency distributions for all items by year as well as the questionnaires utilized for these five years. It is available at the University of Michigan ICPSR archives and Research Gate (Reynolds, Autio, & Hechavarria, 2008).

Notes:

(1.) The data collection procedures are discussed in some detail in Reynolds, Bosma, et al. (2005).

(2.) To provide an unbiased characterization of the population estimates on the screening items, a special analysis is required. If a sub-sample composed of those that were not eligible for a special topic module is combined with a randomly selected half of those that did complete a special topic module (those that received all 14 screening items), the case weights should be adjusted. This can be done by reducing the case weights of the sub-sample where 100% received all initial items. This will ensure that those active in the business modules are not overrepresented.

(3.) This single-item indicator is considered to reflect the transition to a profitable new firm. It reflects the assumption that only after all monthly expenses, including employee wages and salaries, were paid would the owners receive payments. A more complex multi-item measure of firm profits was utilized for the longitudinal studies of the firm creation process (Reynolds, 2000).

(4.) The use of 3.5 years reflects the completion of most surveys in the summer months and the question related only to the year of initial payments of salaries and wages, providing data on period salary payments over 0.5 years, 1.5 years, 2.5 years, etc.

(5.) As the APS locates individuals involved in starting or managing a business, a business with a large team is more likely to be captured in the screening process. Division of case weights by the size of the owner-manager team reduces the impact of ventures with larger teams in the analysis.

(6.) Technically, this is the 95% confidence interval.

(7.) These are coded using the United Nations International Standard Industrial Classification, version 3 or 4 (United Nations, 1990, 2008).

(8.) For the first three years, 1998, 1999, and 2000, WEIGHT_L and WEIGHT_A were based solely on the values of WEIGHT provided by the survey operations. After 2000, weights across all national samples were adjusted to match a standard source of current population age and gender characteristics. United States Census Bureau. International Data Base (IDB). Available at http://www.census.gov/population/international/data/idb/.

(9.) In the initial years, 1999–2003, Erkko Autio was responsible for improvements in index reliability.

(10.) In some years, the GEM program had limited resources, preventing attention to the details and cross-year harmonization of the NES questionnaire data, resulting in some gaps in the archives. In recent years, Prof. Alicia Courdes has been responsible for improving cross-country and cross-year harmonization of the NES questionnaire data.

(11.) Yemen is omitted, as no new firms owner-managers were identified in the survey. Vanuatu, an extreme outlier with a TEA value of 55 per 100, is also omitted. Additional Vanuatu surveys may confirm this result.

(12.) These are the patterns of distribution assumed by the computation of a correlation.

(13.) An effort to create a multi-dimensional Global Entrepreneurship and Development Index (GEDI) that comingles GEM measures of activity, attitudes, and aspirations with hundreds of items from diverse sources tends to obscure the importance of actual business creation (Acs, Szerb, & Autio, 2015). This makes it difficult to identify the positive benefits of business creation and what policies might facilitate more new firms and economic growth.