Tax structure and entrepreneurship

Interestingly, it has been shown that institutional and policy reforms (including tax reforms) can influence the growth effects of entrepreneurship in a non-linear way. Hence, policy reforms, such as trade liberalization, have a stronger positive impact on growth when entrepreneurship is strong while some institutional reforms work best on growth where entrepreneurship is weak and can be detrimental to the effect of entrepreneurship on growth where entrepreneurial activity is high (see, for example, Baliamoune-Lutz 2007 and 2010).

Amongst 139 countries covered by the report, the U.K., Germany, France, and Italy ranked 54th, 84th, 125th, and 126th, respectively, in the component on total tax rates and 95th, 90th, 108th, and 133rd, respectively, in the component on the extent and effect of taxation.

Globalization implies increased ability to take advantage of more favorable fiscal environments and the avenues offered by transfer pricing (see, for instance, Haufler and Schjeledrup 2000) by locating the business or part of it outside the entrepreneur’s home country.

But the link is not necessarily straightforward. For instance, the link between new business formation (admittedly an imperfect proxy for entrepreneurship, as will be discussed below) and growth is both direct and indirect (e.g., displacement of incumbent businesses and byproducts of innovation) and part of the start-ups contributes little to growth. Fritsch (2011) provides a recent survey and concludes that the impact of entrepreneurship on growth is much more qualified. Wennekers and Thurik (1999) present a well balanced study of the link between entrepreneurship and economic growth. Moreover, Dejardin and Fritsch (2011, p. 381), referring to Anyadike-Danes et al. (2011)—who find the impact of growth on the level of new business formation activity to be insignificant and only weak evidence for an influence of start-ups on subsequent growth in Ireland—question “to what extent is new business formation a result of regional growth and to what extent can it be regarded as a source impulse for growth?”.

This could be referred to as the Harberger’s conjecture. See Mendoza et al. (1997).

The story about wealth creation is in fact a story about productive versus unproductive entrepreneurship. On that distinction, see Baumol (1990) and Colombatto and Melnik (2008).

The present study focuses on European countries that have reached similar levels of development and where enforcement of the rule of law is roughly the same, so we will not test for the influence on entrepreneurship of a proper definition and protection of rights. We will, however, control for the effect of economic freedom.

See Parker (1996) and references provided there (pp. 461–462).

By full loss offset it is meant here that the entrepreneur can deduce from her future tax liability the loss times the tax rates. Hence for a tax rate of t and loss of L the amount t × L can be deduced from the future tax liability. So that it is as if the loss had been limited to (1−t)L.

In this context, ‘success tax’ refers to a situation where the entrepreneur will de facto pay income taxes at a higher rate when successful at making profit than when unsuccessful.

Another, maybe more intuitive, explanation could be that higher progressivity is also associated with higher redistribution and better economic and social environments which, in turn, favor entrepreneurship. We thank Sergio Beraldo for pointing out this channel.

It is worth noting that the findings summarized in Table 1 suggest a positive impact of government tax revenues and a negative impact of social security expenditures.

See, for instance, Cullen and Gordon (2007) quoted below.

Parker (1996) used UK data from 1959 to 1991. The policy implications of his findings and interpretations are somehow puzzling if, as he claims, “higher marginal tax rates seem to encourage self-employment relative to paid employment” because “self-employment offers greater opportunities to avoid or evade taxes” (1996, p. 472).

More precisely, they build two scenarios: in one of them success means that the income of the self-employed is twice her previous income while lack of success reduces income to 50 % of what it previously was. They then look at the two corresponding marginal income tax rates. In the second scenario, success increases income to 150 % of what it was while failure reduces it to 75 % of its previous level.

However, such a link between innovation and growth should not be exaggerated. An entrepreneur who imitates some previous innovation can enhance general well being as much as a ‘pure’ innovator.

Also, tax credit for research could introduce a bias in investment decisions. But it can also have no impact in the sense that the investment would have been made independently from the fiscal policy.

We also investigate the existence of a U-shaped relationship between income and entrepreneurship (which has been documented in the literature) and report the relevant estimates in Table 8 in Appendix 1. The results indicate that there is no support for a U-shaped relationship between income and entrepreneurship in the group of European countries included in our sample.

The issue of endogeneity of taxes has been raised in the empirical literature using micro-level data (e.g., Bruce 2000; Stabile 2004). However, we think endogeneity may also be a problem in macro-level data if policymakers respond to shocks in levels of entrepreneurship by altering next period’s tax structure. Growth theory suggests that entrepreneurship is an important determinant of growth (Grossman and Helpman 1991; Romer 1990; Jones 1995). At the macro-level, start-up costs could also be influenced by the level of entrepreneurial activity. The role of socially embedded ties that entrepreneurs form with bankers (Uzzi 1999) and, in some cases, with public employees in business registries can contribute to reducing start up costs. In addition, higher entrepreneurial activity may lead to lower start up costs per registered firm through a scale effect. Note that treating the tax variables, income and start-up costs, as exogenous does not change the magnitude or significance of the results.

Germany is excluded due to the lack of data on tertiary education in the World Bank Word Development Indicators database.

We also estimate the model using OLS on pooled data (with robust standard errors). We report the results in Tables 7 and 8 in Appendix 1 and discuss them briefly in Sect. 5.

We perform tests for second-order autocorrelation (results are omitted to save space but may be obtained from the authors) and the Sargan test of overidentifying restrictions. The test results indicate that all the reported estimations pass both tests.

In addition, the Arellano-Bond GMM estimator was designed for small T and large N panels. In our sample, T (number of years) varies between 5 and 9.

The OECD defines the average wage/income (in national currency) as “the average annual gross wage earnings of adult, full-time manual and non-manual workers in the industry” (see http://​www.​oecd.​org/​tax/​taxpolicyanalysi​s/​oecdtaxdatabase.​htm for more details).

We also use net personal average (and marginal) tax for a single person at 67 and 167 % of average earnings (not shown) and the results are very similar to the ones reported in Table 3. We recognize that the self-employed could decide to incorporate in order to benefit from a lower taxation. However, the goal of this paper is to focus on the transition from being a worker to being self-employed, assuming that, at least in early stages, the self-employed will not incorporate.

We used four alternative indicators of the cost in time and procedures to start a business, including number of procedures to enforce a contract, number of procedures to register property, number of days to enforce a contract and number of days to register property, but they all turned out to be mostly nonsignificant. Thus, we omitted those results.

More accurately, ‘downside’ convexity in Gentry and Hubbard (2004a) refers to the case of failure. Here, we use it to refer to lower levels of success or no improvement in income (relative to wage income).

See, for example, Taxing Wages 2009–2010 Special feature: Wage income tax reforms and changes in tax burdens (www.​OECD.​org).

The results indicate that the effect of economic freedom is, for the most part, statistically nonsignificant. This may be explained by the low variation across the countries in our sample. The median and the mode for economic freedom data have the same value (7.4) and are not very different from the mean (7.38). The standard deviation is very low (0.5).

At a theoretical, micro-level, the study of the impact of taxation on entrepreneurship partly resembles the study of the impact of a change in wages on the supply of labor, with similar conclusions. It is well known, indeed, that lower wages can lead employees to reduce or to increase their supply of labor according to whether the substitution effect (leisure is more attractive when wages are lower) or the income effect (the need to work more to maintain a certain purchasing power) dominates. The same reasoning applies to entrepreneurship, taxation being equivalent to a wage reduction.

While it is not possible to account for all relevant taxes using macro-level data, we have tried to consider other taxes and contributions that create a wedge between gross income and net income by including tax wedges.

Low taxation of capital gain can also promote entrepreneurship to the extent that (i) the entrepreneur can report part of the profits as capital gains and (ii) financing through venture capital will be available at lower cost. Also, what will be reported as corporate income and what will be reported as personal income obviously depends on how the respective rates compare. See Bacher and Brülhart (2010, p. 4).

Self-employment increases opportunities for tax evasion (e.g., it is easier to hide part of business income) and tax avoidance (e.g., deduction of business related consumption). Those strategies are sometimes called “tax-sheltering”. This could explain why higher marginal income tax rates are often associated with an increase of self-employment, at least for high-income brackets (see Schuetze 2000; Bruce 2000, 2002). Hence, tax sheltering could explain what is happening in France where a new legal statute for self-employed was created by law on August 4th 2008, and has been implemented since January 1st 2009. A ‘self-entrepreneur’ (in French, autoentrepreneur) avoids paying corporate tax and social taxes and pays only a single flat tax (13 % of turnover if the activity consists in selling goods, 23 % if selling services). Also, the amount of red tape was reduced relatively to what an ‘ordinary’ business has to face. As a consequence, the number of ‘new entrepreneurs’ soared in the following years to exceed half a million in 2011.

This temptation has become more acute after the recent crisis, since public opinion believes that high-income earners have actually been the main beneficiaries of the bail-out programs financed by the rest of the population.