Taxation on Economic Growth: A literature review (part)

It is always the interest of economists that what determines the long run economic growth. Since the last two or three decades, the effects of public sector on economic growth have received special attention. “The conventional wisdom of the neoclassical model has been that fiscal variables such as the level of taxation and the level of government spending can affect the level of income but have no impact on the rate of economic growth in the long run.”[1] However, after Romer (1986), the developments of the “new growth theory (endogenous growth theory)” have inspired much research investigation into the effects of public factors on economic growth especially for tax and government expenditure, both theoretical and empirical.
Theoretical works generally started from Lucas (1990) and King and Rebelo (1990), where the former one introduced human capital and endogenous growth and got the conclusion that the effects of taxation on growth is trivial; the latter one employed a two-sector endogenous growth model to find that the effects of taxation on growth depend on the production technology for human capital. Rebelo (1991) studied a class of endogenous growth models that have constant returns to scale technologies and drew the conclusion that the growth rate would be relatively low for countries in which there are high income tax rates. Pecorino (1993) analyzed the tax structure effects and growth given a government revenue constraint, he also added some conditions such as depreciation and endogenous labor supply. Devereux and Love (1994) did the qualitative and quantitative analysis of tax effects on growth in an endogenous growth model in which growth is from the joint accumulation of both human and physical capital, they found capital taxes, wage taxes and consumption taxes all have negative effects on growth and on the dynamic path to steady state, capital taxes has the most important effect. Razin and Yuen (1996) focused on the effect of capital taxation effects on long run growth and extended the basic model studied before with endogenous population growth and open economy, they found even larger effects of capital income taxes under free capital mobility. Milesi-Ferretti and Roubini (1998b) analyzed the consumption and factor taxation effects on long run growth with special treatment of leisure by using different formulations of leisure activity, they also treated human capital production section as non-taxable and taxable to conclude the negative effects of all the taxations. Turnovsky (2000) relaxed the labor supply assumption by replacing the inelastic labor supply with elastic one, they then found an adverse effect of consumption and labor income tax on growth rate as well as the same but trivial effect of tax on capital. Petrucci (2001) extended the endogenous growth model with finite-lived households by adopting an Overlapping Generation (OLG) model and then found consumption taxation reduces consumption and therefore raises saving to stimulation capital accumulation and economic growth. This is quite a contrary finding compared with former studies. In summary, the study suggests that taxation has negative effect on long run growth rate, especially for physical capital taxation and human capital taxation, for consumption taxation, the results are relatively ambiguous, and some suggest no effect, some negative effect and some positive effect.
Empirical ones can be generally divided into two groups through the way they deal with tax.

The empirical works from the first group put overall tax burden into the regression equation, i.e. the total tax revenue as a share of GDP or the total income (or other) tax revenue as a share of GDP. These works include Koester and Kormendi (1989), Barro (1991), Engen and Skinner (1992), Agell, Lindh and Ohlsson (1997), Fölster and Henrekson(2001) and etc. The results from these works are quite ambiguous: some found that aggregate tax rates have no effect on growth (Koester and Kormendi (1989), Agell, Lindh and Ohlsson (1997)) while others found a negative relationship ( Engen and Skinner (1992), Fölster and Henrekson(2001) and Cashin (1995) ).

Considering tax revenue-GDP ratio to be too general to be a good proxy for tax policy, the empirical works in the second group tend to use marginal tax rates as the proxy of tax policy. As well as using tax revenue-GDP ratio, Koester and Kormendi (1989) also developed a method to calculate marginal tax rate by regressing tax revenue on GDP; Padovano and Galli (2001) extended this method by adding a dummy for tax reform as well as the interaction item of it and GDP to capture tax structure change effect ; Widmalm (2001) also used this method to calculate what he called “real progressivity” except for using tax revenue and GDP in log forms; Easterly and Rebelo (1993a) used another method by combining information on statutory rates with the amount of tax revenue collected and with data on income distribution; Mendoza, Milesi-Ferretti, Asea (1997) did similar exercise using marginal tax rates calculated by employing data from tax revenue statistics and national accounts of OECD, which is described in Mendoza, Razin, and Tesar (1994). Most of the empirical works using marginal tax rates found a negative relationship between taxation and growth.
Recently, Widmalm (2001) employed a new way to see the effects of different taxes on growth. He took different taxes as a share of total tax revenue into regression using a panel data of 23 industrialized democracy OECD countries between 1965 and 1990.

[1] Fölster and Henrekson (1999)