Energy Economics Manuscript Draft Manuscript Number: ENEECO-D-15-00088 Title: Oil and Entrepreneurship Article Type: Full Length Article Section/Category: Energy Economics Keywords: Natural resource rents, Oil and gas, Entrepreneurship, Corruption Abstract: Economic theory predicts that rents produced from natural resources, especially oil and gas, can increase opportunities for entrepreneurship, but they may also reduce engagement in entrepreneurial activities as they change incentives towards rent-seeking. Using Global Entrepreneurship Monitor (GEM) annual surveys, this study provides empirical evidence that enjoying some profits (per capita) from oil and gas, first, increases entrepreneurship rates, but as the profits (per capita) rise, the negative impact of rent-seeking overcomes and reduces these rates. It also shows that corruption exacerbates this negative impact of natural resource, only, for men's entrepreneurship. The results have important implications for policy makers, especially in developing countries.
Title Page
Oil and Entrepreneurship
Mahdi Majbouri Babson College Economics Division 231 Forest St., Babson Park, MA 02457, USA email:
[email protected] phone: 781-239-5549 fax: 781-239-5239
Manuscript Click here to view linked References
Oil and Entrepreneurship Abstract: Economic theory predicts that rents produced from natural resources, especially oil and gas, can increase opportunities for entrepreneurship, but they may also reduce engagement in entrepreneurial activities as they change incentives towards rent-seeking. Using Global Entrepreneurship Monitor (GEM) annual surveys, this study provides empirical evidence that enjoying some profits (per capita) from oil and gas, first, increases entrepreneurship rates, but as the profits (per capita) rise, the negative impact of rent-seeking overcomes and reduces these rates. It also shows that corruption exacerbates this negative impact of natural resource, only, for men’s entrepreneurship. The results have important implications for policy makers, especially in developing countries. JEL Classification Codes: L26, O13, Q35 Key Words: Natural resource rents, Oil and gas, Entrepreneurship, Corruption
1. Introduction Countries whose oil, gas, or other natural resources make up a large portion of their economy are affected by both positive and negative consequences of these valuable commodities (van der Ploeg 2011). One potential impact of such resources on the economy is creating incentives for potential entrepreneurs to engage in rent-seeking activities (Mehlum et al. 2006). On one hand, by raising income, oil and gas profits increase demand for goods and services, stimulate the economy, and create new industries. Therefore, they offer new business opportunities and increase entrepreneurship. But, on the other hand, when the profits from such resources reach the government coffers – in form of taxes, or in form of revenues from ownership of those resources– the possibility of exploiting them through corruption increases. Potential entrepreneurs who are looking for opportunities in the market and are risk-takers may see the opportunity to connect themselves to the government in various ways in order to exploit these rents. In other words, profits from the resource extraction in the hands of the government create incentives for potential entrepreneurs to engage in rent-seeking activities and dissuade them from participating in entrepreneurship. This has significant ramifications for the economy, especially the growth rate, entrepreneurship, and institutionalization of corruption. When resource profit is small, the first effect dominates and entrepreneurship increases with the discovery and extraction of natural resource. But, when profit from resource extraction becomes substantial, the second effect overcomes and more people engage in rent-seeking activities rather than entrepreneurship. Using a unique dataset on entrepreneurship, covering 80 countries around the world and over time, this study offers an empirical verification of this non-linear relationship between oil and gas profits, and entrepreneurship.
2
Controlling for per capita GDP, a very important predictor of entrepreneurial activity across countries, individual characteristics, as well as country, year, and region-year fixed effects, this paper shows that having some, but not much, oil and gas profit is better than none, as it increases opportunities for entrepreneurship. But, as profit from oil and gas increases, entrepreneurial activity decreases. It also shows that this reduction is larger for male entrepreneurship in countries where corruption is more. The rest of this paper is constructed as follows: Section 2 discusses the theory that relates oil rents to entrepreneurship. In Section 3, the datasets used in this paper are explained. Section 4 provides the estimation strategy and the results. Section 5 is the conclusion which discusses implications for policy and research.
2. The Impact of Oil and Gas Rents Since the cost of production of oil and gas is considerably smaller than their market prices, they produce a large windfall of rent for the economy as a whole, and most of the time for the government in charge. This large influx of income can benefit the economy on many fronts. But, unless the rent is managed and used wisely, it may have negative consequences for these countries as well.1 One potential impact of natural resource rent may be on entrepreneurship. Oil and gas rents can boost entrepreneurship by offering new opportunities that did not exist before through creating a new industry (oil and gas), and increasing disposable income which leads to higher demand for products and services. At the same time, however, another mechanism turns on which creates disincentives for entrepreneurial activity. Profits from oil and gas rents reach the government coffers
For a review of this literature, please see van der Ploeg (2011). For some recent studies, see Apergis and Payne (2014), and Apergis et al. (2014). Majbouri (2014) calculates the opportunity cost of not managing these resources wisely. 1
3
– through taxes, or direct revenues out of ownership of these resources – and create opportunities for rent-seeking that could be more lucrative than entrepreneurship. An entrepreneur can be defined, simply, as one who is willing to take risk in order to implement a venture. Not surprisingly, Schumpeter considered an entrepreneur as a “sociologically distinct individual” (McDaniel 2005; Schumpeter 1934; see Thompson 2004, for supporting evidence.) Similarly, it is argued in the literature that there are traits and characteristics required in a person to make her an entrepreneur: for example, an insatiable need for achievement (Bygrave 1989; McClelland 1961), an internal locus of control which is the belief that the outcomes are mostly dependent upon one’s own actions (Bygrave 1989), alertness to opportunities (Kirzner 1973, 1979), risk-taking (Bygrave 1989; McClelland 1961; Shane 2004), boldness, daring, and creativity (Lumpkin and Dess 1996; Hills et al. 1999), overconfidence (Forbes 2005; Shane 2004), and stress tolerance (Rauch and Frese 2007). Thompson (2004) identifies six key entrepreneurial character themes, or natural and instinctive behaviors, and argues that techniques may help people to implement ideas, “but alone they cannot compensate for missing characteristics.”2 Moreover, evidence shows inherent personal characteristics affect career choice towards self-employment (Carter et al. 2003; Lüthje and Franke 2003). Cope (2005) argues that these are all innate abilities and permanent characteristics that evolve little over time and context. These abilities and characteristics are not found in everyone but there is a share of every society who has those traits and characteristics and can be potentially an entrepreneur. Baumol (1990) argues that potential entrepreneurs can engage in two activities: 1) entrepreneurship or what he calls ‘productive entrepreneurship’: that is creating and selling products and services that are valued in the marketplace, and 2) rent-seeking or what he calls ‘unproductive entrepreneurship’:
2
For more studies, see Puga and Garcia (2012).
4
that is creating connections with sources of rent, and competing to capture more of the rent in the economy. The latter is possible when there are rents available in the economy and particularly when rents are concentrated in a few sources like the government (for example, when government enjoys sizable profits from owning or taxing natural resources.) The potential entrepreneurs choose between these two activities based on the structure of returns from them. Figure 1 explains this graphically. The horizontal axis represents the number of potential entrepreneurs. They can participate in productive or rent-seeking activities. As more of them choose productive activity, fewer will be rent-seekers. The vertical axis on the left shows profits per entrepreneur. The number of potential entrepreneurs who choose productive activity depends on the profits earned from it. The larger are the profits per entrepreneur, the more people will engage in productive activities. It is the simple law of supply and is clear from the profits lines that are upward-sloping. [Figure 1] On the other hand, the total rent in the economy is fixed. So if more people engage in rent-seeking, each ends up with a smaller share (i.e. smaller rent). The rents curves demonstrate such phenomenon. The vertical axis on the right denotes rent per rent-seeker. The more people become rent-seekers, the smaller will be the average rent that each receives (rent per rent-seeker). The potential entrepreneur chooses productive activities if the average profits (profits per entrepreneur) is larger than the average rents. Therefore, the equilibrium number of entrepreneurs and rent-seekers happens when the average profits are the same as the average rent. This is where no one wants to switch from being an entrepreneur to a rent-seeker and vice versa. When there is no natural resource in the economy, there is still opportunity for rent grabbing for example via legal loopholes, or connections with the government. In this situation, the rent curve 5
showing rent per rent-seeker is represented by the rents curve in Figure 1 and the supply curve for the entrepreneurs is depicted by profits curve. The equilibrium, when no natural resource is in the economy, is point A on the graph. As soon as a valuable natural resource is discovered, it provides many new opportunities for entrepreneurs, since it leads to creation of a new industry (resource extraction) with all its supply chain. This leads to a substantial hike in expected profits for entrepreneurs who enter this industry. Therefore, the introduction of the natural resource to the economy is revolutionary and shifts the expected profits curve substantially to profits’. At the same time, more income in the hands of the government (through owning or taxing the resource) may increase the total rent in the economy. But the potentials for entrepreneurship is so large that it overshadows the increase in the amount of rent available. In other words, the profits curve shifts more than the rents curve. Therefore, the equilibrium moves from A to B which means the number of (productive) entrepreneurs increases. In summary, having a natural resource allows the resource extraction industry and its supply chain to develop and increases incentives for entrepreneurship. As the value of natural resource increases, it stimulates the economy more but not as much as before. This shifts the profits’ to profits’’, for instance. But, with this rise in the natural resource value, especially if this value is concentrated in the hands of corrupt governments and bureaucracies, the size of rents available for rent-seekers increases substantially. This shifts the rents curve to the left (to rents’’). The result is that large sums of natural resource value move the equilibrium towards point C. In other words, the number of entrepreneurs decreases as they switch to rent-seeking activities. One may expect that less rent would be available for rent-seekers if there is less corruption in the
6
government. In other words, the rents curve shifts less and fewer potential entrepreneurs choose rent-seeking.3 In summary, profits from extraction of natural resources, and particularly oil and gas, has an inverted U-shape relationship with entrepreneurship. First, it increases the number of productive entrepreneurs and then it reduces it. Using data on entrepreneurship and oil and gas annual profits across countries and over time, this study provides empirical evidence for this relationship. It documents that, initially, they increase entrepreneurship but as the size of oil and gas value increases in the economy, it induces more rent-seeking activity and reduces entrepreneurship. It also shows that in places where corruption exists, more oil and gas profits leads to a larger reduction in male entrepreneurship. No significant impact for female entrepreneurs is observed. The data, estimation strategy and results are discussed in the rest of this paper.
3. Data This study uses four data sets. The first is the Global Entrepreneurship Monitor (GEM) surveys which cover about 80 countries and extend over ten years. GEM is the largest study of entrepreneurial dynamics in the world and its main purpose is to measure entrepreneurial activity and understand entrepreneurial behavior, attitudes, and aspirations around the world.4 The project
The evidence also shows that rise in oil and gas profits has an adverse effect on institutional quality when there is little history of good institutions. Using a panel dataset covering ninety-nine countries during 1980–2004 and controlling for income, time-varying common shocks, regional fixed-effects, and some other covariates, Bhattacharyya and Hodler (2010) find that natural resources only induce corruption in countries that have endured a nondemocratic regime for more than 60 percent of the years since 1956. Vicente (2010) compares changes in perceived corruption in the island Sao Tome and the island of Cape Verde which have similar histories, culture, and political institutions after a natural experiment – a significant oil discovery in the island of Sao Tome. He finds that corruption increased by close to 10 percent after the announcements of a significant oil discovery in 1997–99. In a quasi-experimental setting and using data on Brazilian municipalities, Fernanda Brollo et al. (2010) show that an increase in oil profits of 10 percent for a municipality raises corruption by 17–24 percent, and increases the likelihood of the incumbent remaining to office by 7 percent. It also shrinks the fraction of its opponents holding a college degree by 7 percent. All these mean that in countries with little history of good institutions an increase in oil and gas profits deteriorates the institutional quality and shifts the rents curve more towards left. 4 More information about the GEM project may be found at www.gemconsortium.org 3
7
started in 1999 with 10 countries and grew to about 70 countries in 2012. Since 2004, the quality of the data, collected in various countries, has been professionally controlled by a team of experts, making it very reliable. Because of this, this study only uses the surveys after 2004. But because the oil and gas rents data are available until 2008 only, this research would not be able to use the surveys after 2008. Hence, only surveys between 2004 and 2008 could be exploited for this study. GEM consists of two surveys: the Adult Population Survey (APS) and the National Expert Survey (NES)5. APS is a harmonized, nationally representative random population survey gathered annually from individuals in each participating country. Within country sample size ranges from 2,000 to 45,000. In each country, a team is responsible for conducting the survey. A local survey firm is carefully chosen to collect data for a standard survey questionnaire. Survey questions are standard globally and conducted simultaneously across all countries between May and August of every year. Surveys are collected via phone and where phone penetration is low by door-to-door interviews. A standard protocol governs sampling, coding, weighting and other data collection procedures. A central coordination team of experts monitors the sampling procedure in each country and the accuracy and compatibility of data across countries. It requests for clarifications and corrections for coding and weightings, whenever necessary, so that the protocol would be followed by all national teams. At the end of collection procedure, the central coordination team harmonizes the data across countries, and offers them to public on the GEM consortium website.6 This study uses APS series to measure entrepreneurial activity for several reasons. First, APS is a population survey and asks individuals whether they engage in businesses. If the individual owns or
NES is an expert opinion survey containing opinions about nine dimensions of entrepreneurial environment in each country. These dimensions are finance, government policies and programs, entrepreneurial education and training, R&D transfer, commercial and professional infrastructure, entry regulation, physical infrastructure and services, and cultural and social norms. This study does not employ NES surveys and only uses APS data. 6 www.gemconsortium.org; Details about the procedures used to collect and harmonize GEM data can be found in Reynolds et al. (2005). 5
8
manages a business, or is in the process of starting one, she will be asked follow-up questions about the business and herself. Therefore, it covers microenterprises which are notoriously difficult to account for in firm surveys. Second, APS accounts for informal businesses which are usually neglected in firm surveys, since firm surveys create their sample based on registered businesses. Third, APS datasets are collected according to a standard protocol across countries and are qualitycontrolled and harmonized by a central team. Therefore, entrepreneurial activities are measured consistently and easily comparable across countries. This is very important in cross-country comparison and is rarely the case. For example, unlike APS datasets, some economic indicators are measured differently in each country which renders them incomparable. Fourth, APS data account for those who failed as well. Failure is an important part of entrepreneurship which is usually not measured in other surveys. All these reasons make GEM surveys the best for studying entrepreneurship across countries. Data on oil and gas rents come from World Bank Subsoil and Forest Rents dataset. It has annual rents, i.e. profits, from oil and gas for every country from 1970 to 2008. These rents are divided by population to obtain rents per capita which is the most reliable measure of the significance of natural resource in an economy (For a discussion, please see Ross, 2008). Population data and GDP per capita in 2005 constant prices are from World Bank Development Indicators. Corruption Perceptions Index, which is used as a measure of corruption, is obtained from Transparency International. This study employs two samples: one for all countries that were in the GEM surveys between 2004 and 2008, and another that only includes countries which produce oil or gas, even if in small volumes. The first sample has 65 countries and the second 50. Table 1 contains the summary
9
statistics of all the variables used for these samples. GDP and oil and gas rents per capita are deflated to account for inflation.
4. Estimation and Results The rest of this paper demonstrates if rents have any impact on entrepreneurial activity. Consider the following regression model: 𝐻𝑖𝑘𝑡 = 𝛼 + 𝛽𝑓(𝑅𝑒𝑛𝑡𝑘𝑡 ) + 𝛾 ′𝑍𝑘𝑡 + 𝜃 ′𝑋𝑖𝑘𝑡 + 𝑐𝑘 + 𝛿𝑡 + 𝜇𝑟𝑡 + 𝜀𝑖𝑘𝑡
(1)
in which, 𝐻𝑖𝑘𝑡 is a dummy showing whether individual 𝑖 in country 𝑘 in year 𝑡 was an entrepreneur, 𝑅𝑒𝑛𝑡𝑘𝑡 , and 𝑍𝑘𝑡 are respectively oil and gas rents per capita, and observed characteristics of the population in country 𝑘 in year 𝑡. 𝑓(𝑅𝑒𝑛𝑡𝑘𝑡 ) is a functional form of 𝑅𝑒𝑛𝑡𝑘𝑡 . 𝑋𝑖𝑘𝑡 contains the characteristics of individual 𝑖 in country 𝑘 in year 𝑡. 𝑐𝑘 is the country 𝑘 fixed effect which controls for all characteristics of country 𝑘 that are fixed over time. These include but not limited to culture, history, traditions, social norms, political, social and economic institutions, laws and regulations, and time-constant demographic characteristics. 𝛿𝑡 is the year 𝑡 fixed effect. It controls for any unobserved factor that affects all countries in year t, for instance, a global boom or bust. 𝜇𝑟𝑡 represents region-year fixed effects, containing all factors that are affecting all the countries in a region (such as Western Europe and the Offshoots, or Middle East and North Africa) in a specific year (year 𝑡), for instance, a regional trade agreement. The regions are Western Europe and the Offshoots (US, Canada, Iceland, Australia, and New Zealand), Latin America and the Caribbean, Eastern Europe and Russia, Middle East and North Africa, Sub-Saharan Africa, and East and South Asia. 𝜀𝑖𝑘𝑡 are independently distributed disturbances.
10
Table 2 presents the results for the regressions based on Equation (1), when the dependent variable is whether an individual was an entrepreneur in the preceding year. In this study, an entrepreneur is defined as someone for whom any of the following four statements in the GEM surveys is true:
You are, alone or with others, currently trying to start a new business, including any selfemployment 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-- an effort that is part of your normal work.
You are, alone or with others, currently the owner of a company you help manage, selfemployed, or selling any goods or services to others.
You have, in the past 12 months, sold, shut down, discontinued or quit a business you owned and managed, any form of self-employed, or selling goods or services to anyone.
Therefore, the dependent variable in the regressions is a dummy equal to one if any of the above four statements in the GEM surveys was true for the individual and zero otherwise. The fourth statement is particularly necessary to be included as failure is an important part of entrepreneurship. One of the strengths of GEM surveys is that they account for those who failed as well. First, we analyze the sample that contains all countries. Like GDP per capita, oil and gas rents per capita has a large variation across countries. Hence, similar to GDP per capita, in place of 𝑓(𝑅𝑒𝑛𝑡𝑘𝑡 ) in Equation (1), one can use a function of natural log of oil and gas rents per capita. Since some countries are not producing any oil and gas, their oil and gas rents per capita is zero. But it is not possible to take natural log of zero. Therefore, rents per capita is added by one to get 𝑙𝑛(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎 + 1). As Figure 2 shows, the distribution of this variable is discontinuous
between 0 and 13. About twenty percent of the observations have the value of zero and the rest are
11
above 13. This is one issue to be aware of when we use the sample of all countries whether they produce oil or gas or not. Table 2 reports the regressions based on Equation (1) for the whole sample. The first column only includes a function of log of rents per capita as well as country and year fixed effects. The function contains both 𝑙𝑛(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎 + 1) and its squared, since there is a quadratic relationship between rents per capita and entrepreneurship.7 Consistent with the theory discussed in Section 2, the relationship is concave (an inverted U), as the coefficient of log of rents per capita is positive while the coefficient of its squared is negative. This implies that having a little bit of oil and gas rents is better than nothing. In the second column, 𝑙𝑛(𝐺𝐷𝑃 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎) and its squared (i.e. country level variables) are added to the regression, and in the third column, gender, age, and age squared (i.e. individual level characteristics) as well as region-year fixed effects are included. The concave relationship is robust in all these regressions. In the third column, the coefficient of log of rents per capita is about 17 times larger than that of its squared. This means that if natural log of oil and gas rents is less than 17 (the break-even point), the impact of oil and gas on entrepreneurship is positive. That is producing oil and gas is better than not producing it. This positive impact is maximum when log of per capita rents is about 8.5, where it starts to fall. The summary statistics for oil or gas producing countries in Table 1 shows that, the minimum value for natural log of oil and gas rents per capita in the data is 13.1, which belongs to Belgium. For Belgium, entrepreneurship rate is 0.05 percentage point (= 0.017 × 13.1 − 0.001 × 13.12) higher because of oil and gas profits. As the rents increase, this positive impact declines. For Greece, whose
If one does not include (𝑙𝑛(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎 + 1))2 in the regressions, the coefficient of 𝑙𝑛(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎 + 1) becomes insignificant in all regressions. 7
12
log of real rents per capita is about 17.3 (a bit above the break-even point), there is no positive benefit to having oil and gas rents. After this, the impact becomes negative. These results are robust to many omitted variables. The country fixed effects control for any feature of a country that is constant over time. Therefore, all time-constant factors including but not limited to history, culture, religion, social norms, time-constant traits and attitudes, political and economic institutions, time-constant laws and regulations, as well as demographic features that changed little between 2004 and 2008 (such as average population education,) do not bias the results. The year fixed effects control for all factors that affected all countries in a specific year; for example, something, like a global boom or bust that influenced the whole world. Region-year fixed effects were added in Column (3) to control for any time-varying factor that affected countries in a geographic region in a specific year, such as a regional agreement. Although these country, year, and region-year fixed effects control for many omitted variables, there might still be country-specific time-varying factors that would be correlated with GDP and rents per capita as well as entrepreneurial activity, and hence, bias the results. It is not possible to control for such factors with country-time fixed effects as rents vary across country and time. Coefficients of log of real GDP per capita and its squared, in Column (3), show a convex relationship with entrepreneurship. This is also documented in Allen and Langowitz (2011) who showed that entrepreneurship rates first decline and then rise as GDP per capita increases. Female dummy has a negative coefficient showing that on average entrepreneurship rates are lower for women. Allen and Langowitz (2011) describe this gender difference in detail. Similar to the results in Kautonen et al. (2014), Age and Age2 show a concave relationship with entrepreneurship rates. As age increases, initially, entrepreneurial activity increases first, but it falls afterwards.
13
If one estimates the same regressions, as those reported in Table 2, using only individuals in the fifty oil and gas producing countries in the sample, she can find a linear relationship between the log of real rents per capita and entrepreneurship. Since linear relationships are easier to use and interpret, Table 3 reports them in the same structure as Table 2. The first column of Table 3 reports the result when log of rents per capita as well as the country and year fixed effects are the explanatory variables. The coefficient of log of rents is negative and significant, showing that larger rents reduce entrepreneurship. In the second column, ln(𝐺𝐷𝑃 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎) and its squared are included as dependent variables. Since these are country-specific time varying factors, the country and year fixed effects cannot control for them. The coefficient of log of real per capita rents remains the same confirming the findings in Column (1). In the third column, the individual characteristics available in the GEM surveys: gender, age and age squared as well as region-year fixed effects were added to the regressions for robustness check. The coefficient of log of rents per capita is still negative and significant and has not changed. Similar to Table 2, coefficients of log of real GDP per capita still show a convex relationship with entrepreneurship. Similarly, female dummy has a negative coefficient and Age and Age2 show a concave relationship with entrepreneurship rates. Even the size of these coefficients are the same between these two tables. Based on Mehlum et al. (2006), one may argue that the impact of oil and gas rents should be larger in countries that are more corrupt. In order to test this conjecture, one can interact oil and gas rents with a measure of corruption in the country. Transparency International measures corruption globally every year and reports it in a measure called Corruption Perceptions Index which is between 0 (the most corrupt) and 10 (the least corrupt). In other words, the larger is this measure, the less the 14
country is corrupt. In order to get a measure that is increasing in corruption, one can subtract Corruption Perception Index from 10. This way the least corrupt environment gets a zero grade and the most corrupt one is marked 10. This new variable, entitled Corruption, is included in the regressions to simplify the interpretation of the results. Table 4 reports these results using the sample of oil or gas producers, for all the individuals (the first column), women (the second column), and men (the third one). We observe that corruption increases the impact of oil rent only for male entrepreneurial activity. This is a very interesting result. It can be argued that women do not participate in rent-seeking activities as much as men do. Could it be because women are more likely to follow ethical ways of earning income and hence, less likely to become rent-seekers? Or could it be because women have less access to the corrupt sources of rent? These are questions that require further analysis. The impact of oil and gas rents on male entrepreneurship is positive if corruption index is below 2 (i.e. Transparency International CPI larger than 8). In other words, profits from oil and gas increase entrepreneurial activity in countries such as United Kingdom, Canada, and Norway, which have very low corruption. The turning point is at corruption index of 2. As corruption increases, entrepreneurship is discouraged. Adding the interaction with corruption to the regressions in Table 2 for all countries is a bit complex, since the squared log of rents per capita is also included. But, interestingly, if one runs the same regressions as those reported in Table 4 for the sample of all countries, she finds similar results as those in Table 4. In other words, if we do not include squared of log of rents in the regressions in Table 2 and include an interaction of log of rents per capita with corruption, we find that corruption increases the negative impact of oil and gas rents for men. The coefficient of log of rents per capita would be 0.06, while the coefficient of its interaction would be -0.02. For countries with corruption
15
index larger than 3 (or Transparency International CPI smaller than 7,) oil and gas rents reduce male entrepreneurship. As corruption increases this reduction in entrepreneurship becomes larger in absolute value.
5. Conclusion This study examined the relationship between the size of oil and gas rents (profits) and entrepreneurship. It demonstrated that this relationship is an inverted U-shape. Some rent, especially in a less corrupt environment, boosts entrepreneurship as it offers new opportunities through creating a new industry (oil and gas), as well as increasing disposable incomes and demand for new products and services. But at the same time, more income for the governments, through owning or taxing oil and gas resources, creates opportunities for rent-seeking especially in a corrupt environment. This entices potential entrepreneurs to choose rent-seeking rather than entrepreneurship because it is more lucrative. This study offers empirical evidence that as oil and gas rents increase more than a threshold, the rent-seeking opportunities rise more than the entrepreneurial opportunities in the economy, leading to less entrepreneurship. One important lesson for policy is that significant amount of oil and gas rents may not be advantageous for the economy and specifically productive entrepreneurship. It creates new sources of rents, thereby increasing the expected returns to rent-seeking activities. One solution is to reduce the amount of rent that is available to the government. Ideas such as transparent commodity funds, and lump-sum distributions are helpful in reducing rent-seeking. Governments can create transparent sovereign wealth funds to save some of the oil and gas profits especially when those profits are rising and new rent-seeking opportunities are created. Since these savings take excess profits out of the economy and into the transparent fund, fewer new opportunities for rent-seeking will be created when the oil and gas profits increase. Another way to reduce rent-seeking activities is to distribute all 16
excess profits from oil and gas among the general public. This is the idea based on Alaska Permanent Fund which by law distributes half of investment earnings equally among the population. This way less rent becomes available for rent-seekers.8 Another important lesson is that fighting corruption and creating more transparency is the key in harnessing natural resources. Since these profits can become a source of corruption and bad institutions themselves9, it is difficult to mix oil and gas with transparency but the resulting potion is the elixir of fast and continuous growth. One solution is that the whole process of extraction, transportation, refinement, and revenue management of oil and gas is audited by international auditors. Joining the Extractive Industries Transparency Initiative (EITI) provides a standard framework for governments, and the industry alike, on how to reduce corruption. Grass-roots pressure in the developed countries have pushed the international corporations to sign off on social responsibility schemes such as EITI by which they have to follow stricter rules in dealing with resource-rich governments. Under the EITI initiative, the oil companies need to publish what they pay to governments and the governments need to show their receipts. The initiative matches these two figures and publishes the result for the general public. Governments sometimes try to encourage entrepreneurship through sponsoring entrepreneurs, or subsidizing their cost of capital, labor, or taxes. But, these policies usually become a source of rent for rent-seekers themselves, and lead to worse outcomes. Many potential entrepreneurs write proposals or define projects just to acquire these rents from the government. It is always difficult for governments to choose the winners, the right entrepreneurs and the right projects. This results in many bad apples wasting resources. More research on understanding the interrelation of oil and gas
Sala-I-Martin and Subramanian (2013) and Birdsall and Subramanian (2004) suggest this idea for Iraq and Nigeria. See Bhattacharyya and Hodler (2010) for panel data evidence, Vicente (2010) and Borello et al. (2010) for quasiexperimental evidence. 8 9
17
and corruption is necessary and more ideas to encourage entrepreneurship that are feasible in developing countries are required.
Acknowledgements: I am sincerely grateful to Babson Faculty Research Fund (BFRF) for supporting this research through a summer grant. Thanks to BFRF reviewers for their comments. All the remaining errors are mine.
References Allen, I. Elaine, and Nan S. Langowitz (2011). “Understanding the Gender Gap in Entrepreneurship: A Multicountry Examination.” Chapter in The Dynamics of Entrepreneurship: Evidence from Global Entrepreneurship Monitor Data, (Ed.) Maria Minniti,
Oxfor University Press, pp. 31-56. Apergis, Nicholas, and James E. Payne (2014). “The oil curse, institutional quality, and growth in MENA countries: Evidence from time-varying cointegration.” Energy Economics, 46: 1-9. Apergis, Nicholas, El-Montasser, Ghassen , Sekyere, Emmanuel, Ajmi, Ahdi N., and Rangan Gupta (2014). “Dutch disease effect of oil rents on agriculture value added in Middle East and North African (MENA) countries.” Energy Economics, 45: 485-490. Baumol, William (1990). “Entrepreneurship: Productive, Unproductive, and Destructive.” Journal of Political Economy, 98(5): 893-921.
Bhattacharyya, Sambit, and Roland Hodler (2010). “Natural Resources, Democracy and Corruption.” European Economic Review, 54(4): 608–21.
18
Birdsall, Nancy and Arvind Subramanian (2004). “Saving Iraq From Its Oil.” Foreign Affairs (July/August): 77-89. Bolton, W.K. and John L. Thompson (2003). The Entrepreneur in Focus: Achieve Your Potential. Thomson, London. Bygrave, W. D. (1989). “The entrepreneurship paradigm (I): a philosophical look at its research methodologies.” Entrepreneurship Theory and Practice, 14(1), 7–26. Carter, N. M., Gartner, W. B., Shaver, K. G., and E. J. Gatewood, (2003). “The career reasons of nascent entrepreneurs.” Journal of Business Venturing, 18(1), 13–39. Cope, J. (2005). “Toward a dynamic learning perspective of entrepreneurship.” Entrepreneurship Theory and Practice, 29(4), 373–397.
Brollo, Fernanda, Tommaso Nannicini, Roberto Perotti, and Guido Tabellini. (2013). “The Political Resource Curse.” American Economic Review, 103(5): 1759-96. Hills, G. E., Shrader, R. C., & Lumpkin, G. T. (1999). “Opportunity recognition as a creative process.” Frontiers of Entrepreneurship Research (pp. 216–227). Wellesley, MA: Babson College. (Carter et al. 2003; Lüthje and Franke 2003): motivation for self employment
Kautonen, Teemu, Down., Simon, and Maria Minniti (2014). “Ageing and Entrepreneurial Preferences.” Small Business Economics, 42(3): 579-594. Kirzner, I. M. (1973). Competition and entrepreneurship. Chicago: University of Chicago Press. Kirzner, I. M. (1979). Perception, opportunity, and profit: Studies in the theory of entrepreneurship. Chicago: University of Chicago Press. 19
Lumpkin, G. T., and G. G. Dess (1996). “Clarifying the entrepreneurial orientation construct and linking it to performance.” Academy of Management Review, 21(1): 135–172. Lüthje, C., and N. Franke (2003). “The ‘making’ of an entrepreneur: testing a model of entrepreneurial intent among engineering students at MIT.” R&D Management, 33(2): 135– 147. Majbouri, Mahdi (2014) “Calculating the Income Counterfactuals of Oil Exporting countries of the MENA Region,” forthcoming in Resources Policy. Minniti, Maria (2011). “Data Appendix” Appendix in The Dynamics of Entrepreneurship: Evidence from Global Entrepreneurship Monitor Data, (Ed.) Maria Minniti, Oxfor University Press, pp.
307-312. McClelland, D. C. (1961). The achieving society. Princeton: Van Nostrand Reinhold. McDaniel. Bruce A. (2005). “A Contemporary View of Joseph A. Schumpeter's Theory of the Entrepreneur.” Journal of Economic Issues, 39(2): 485-489. Mehlum, Halvor, Karl Moene, and Ragnar Torvik (2006). “Institutions and the Resource Curse.” Economic Journal, 116(508): 1–20.
Puga, Jorge López , Juan García García (2012). A Comparative Study on Entrepreneurial Attitudes Modeled with Logistic Regression and Bayes Nets. The Spanish journal of psychology, 15:03 Quill, M., Bosma, N., and Minniti, M. (2007), 2006 Global Entrepreneurship Monitor Technical Assessment. London, UK, and Babson Park, MA: London Business School and Babson
College.
20
Rauch, A., and M. Frese (2007). “Let’s put the person back into entrepreneurship research: a metaanalysis on the relationship between business owners’ personality traits, business creation, and success.” European Journal of Work and Organizational Psychology, 16(4), 353–385. Raynolds, P.N., Bosma, E., Autio, E., Hunt, S., De Bono, M., Servais, I., Lopez-Garcia, P., and Chin, N. (2005). “Global Entrepreneurship Monitor: Data Collection Design and Implementation, 1998-2003,” Small Business Economics, 24(3): 205-31. Ross, Michael L (2008). “Oil, Islam, and Women.” American Political Science Review 102(1): 107123. Sala-i-Martin, Xavier, and Arvind Subramanian (2013). “Addressing the Natural Resource Curse: An Illustration from Nigeria,” Journal of African Economies, 22(4): 570-615. Schumpeter, J. A. (1934). Theory of economic development: An inquiry into profits, capital, credit, interest and the business cycle. Cambridge: Harvard University Press.
Shane, S. A. (2004). A general theory of entrepreneurship: The individual-opportunity Nexus. Cheltenham: Edward Elgar. Thompson, John L (2004). “The facets of the entrepreneur: identifying entrepreneurial potential.” Management Decision, 42(2): 243 – 258.
van der Ploeg, Frederick (2011). “Natural Resources: Curse or Blessing?” Journal of Economic Literature, 49(2): 366–420.
Vicente, Pedro C. (2010). “Does Oil Corrupt? Evidence from a Natural Experiment in West Africa.” Journal of Development Economics, 92(1): 28–38.
21
Figures
Figure 1 – Productive Entrepreneurship and Rent-Seeking
C
rents’
rents
Rents per Rent-seeker
Profits per Entrepreneur
rents’’
B
A
profits’’
profits’ profits
Entrepreneurs
Rent-seekers
Note: The profits curve shows that the larger are the profits, more people engage in productive activities. On the other hand, the total rent in the economy is fixed. The rents curves show that as more people engage in rent-seeking, each will end up with a smaller share. The equilibrium happens when the average profits are the same as the average rent, i.e. where the two lines cross. As rents increase, the rents curve shifts to the left and with it the equilibrium. Hence, it decreases the number of productive entrepreneurs. Source: van der Ploeg (2011), based on Mehlum et al. (2006).
22
0
.05
.1
.15
.2
.25
Figure 2 – Distribution of natural log of oil and gas rents per capita
0
10 20 ln(Real rents per capita + 1)
Note: Real rents per capita is the total profits (price minus the cost of production) of oil and gas in constant prices deflated and divided by population of the country.
23
30
Tables
Table 1 – Summary Statistics Mean
Std. dev.
Min
Max
Being an entrepreneur in the last year
0.2
0.4
0.0
1.0
ln(Rents per capita + 1)
18.1
8.4
0.0
26.8
ln(GDP per capita)
9.9
1.0
5.7
11.1
Female
0.5
0.5
0.0
1.0
Age
42.8
15.1
9.0
104.0
Being an entrepreneur in the last year
0.2
0.4
0.0
1.0
ln(Rents per capita)
21.5
3.1
13.1
26.8
Transparency International CPI
6.6
2.1
2.2
9.6
Corruption Index
3.4
2.1
0.4
7.8
ln(GDP per capita)
9.9
1.0
6.7
11.1
Female
0.5
0.5
0.0
1.0
Age
43.1
15.2
9.0
104.0
All countries:
Oil or gas producers:
Note: Being an entrepreneur in the last year is a dummy that is equal to one if the individual at the time of the survey was engaged in starting a business, or a new venture in an existing business, or owned an established business, or had shut down a business in the 12 months prior to the survey. ln(Rents per capita) represents annual profits from oil and gas rents deflated for inflation. Transparency International CPI is the Corruption Perceptions index which is between zero (fully corrupt) and ten (no corruption). Corruption is equal to ten minus Transparency International CPI. ln(GDP per capita) and its squared are in 2005 constant prices for each country and over time. Female is a dummy equal to one if the individual is female and zero otherwise. Age is measured in years. Number of countries: 50 countries Years: 2004 through 2008
24
Table 2 –Entrepreneurial Activity and Resource Rents (All countries) All (1)
(2)
(3)
ln(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎)
0.024*** (0.009)
0.021*** (0.007)
0.017*** (0.004)
(ln(𝑅𝑒𝑛𝑡𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎))2
-0.002**
-0.001**
-0.001***
(0.001)
(0.001)
(0.000)
ln(𝐺𝐷𝑃 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎)
-0.42 (0.94)
-2.60*** (0.95)
(ln(𝐺𝐷𝑃 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎))2
0.01 (0.05)
0.11** (0.05)
Female
-0.10*** (0.01)
Age
0.02*** (0.001)
Age2 × 10−2
-0.02*** (0.001)
Country fixed effects Year fixed effects Region × Year fixed effects Observations Adjusted R-squared
Yes Yes No
Yes Yes No
Yes Yes Yes
691,668 0.065
691,668 0.065
691,030 0.095
Note: The dependent variable is being an entrepreneur in the past year. For description and summary statistics of variable, please see Table 1. 65 countries and 5 years (2004 through 2008) are in the sample. Not all countries have data for all years. Regions are Western Europe and the offshoots (US, Canada, Iceland, Australia, and New Zealand), Latin America and the Caribbean, Eastern Europe and Russia, Middle East and North Africa, Sub-Saharan Africa, and East and South Asia. Robustheteroskedastic standard errors corrected for correlation inside countries are in parentheses. *** p
