Land lease revenue windfalls and local tax policy in China

Abstract

This study examines how land lease revenue, a fiscal resource windfall available to local governments, has shaped local tax policy in China. We follow the literature to argue that the presence of resource revenue incentivizes governments to substitute the more distortive tax policy with the resource revenue. Studying a city-level dataset and a large manufacturing firm-level dataset from 2000 to 2013, we find evidence for this argument by showing that land lease revenue available to city governments is negatively associated with the effective tax rates faced by the firms in the cities. Furthermore, we show that the effect of land lease revenue is likely to be weakened in larger cities, cities with more agglomerated industries, and cities with lower capital mobility. Finally, we show that the effect of land lease revenue on tax rates is more salient for firms that are under the direct control of local governments and for firms that have stronger bargaining power with local governments.

Appendices

Appendix 1

See Tables 10 and 11

Appendix 2

In this Appendix, we establish a simple theoretical model to highlight the relationship between land lease revenue and effective tax rate imposed on firms under the tax competition model. The basic setup of the model builds on the fundamental work of Zodrow and Mieszkowski (1986) and Wilson (1986), and one that has been widely adopted in the tax competition literature (e.g., Wilson 1999; Wilson and Wildasin 2004).

Consider an economy consisting of $N$ localities, indexed by $i$. In each locality, a representative firm produces a numeraire output that can be used either for private or government consumption. In particular, the production function of the representative firm (in per-capita terms) is presented as a neoclassic production function in intensive form$f\left( {k_{i} } \right)$, satisfying$f^{\prime}\left( {k_{i} } \right) > 0 ,{ }f^{\prime\prime}\left( {k_{i} } \right) < 0$, where $k_{i}$ is the amount of per-capita capital. Assuming absentee capital owners and labor supply corresponding to population size, per-capita private consumption $c_{i}$ is determined by labor income,

$$c_{i} = f\left( {k_{i} } \right) - k_{i} f^{^{\prime}} \left( {k_{i} } \right)$$

(2)

Following Buettner (2006), we assume the objective function of local government is to maximize the utility of a representative resident, which is a combination of private consumption $c_{i}$ and public expenditure $e_{i}$ in a quasi-linear form,

$$u_{i} = c_{i} + \alpha_{i} v\left( {e_{i} } \right)$$

(3)

where $v^{\prime}\left({e_{i}}\right) > 0, v^{\prime\prime}\left({e_{i}}\right) < 0$; $\alpha_{i} > 0$ measures the government’s preference for public expenditure relative to private consumption. To be consistent with the Chinese institution and to simplify the model, we assume that local governments use their capital tax revenue ($\tau_{i} k_{i}$) and land lease revenue ($w_{i}$) to finance public expenditure ($e_{i}$ ).²⁸ Thus, the budget constraint of each locality in per-capita terms is presented as,

$$e_{i} = \tau_{i} k_{i} + w_{i}$$

(4)

where $\tau_{i}$ is the local tax rate on capital $k_{i}$.

Since capital is assumed to be perfectly mobile across localities, the market-clearing condition implies an allocation of capital across localities such that net return in all localities is equalized to the given economy-wide net return to capital ($r$),²⁹ that is,

$$f^{\prime}\left( {k_{i} } \right) - \tau_{i} = r$$

(5)

Immediately, we have

$$\frac{{\partial k_{i} }}{{\partial \tau_{i} }} = \frac{1}{{f^{\prime\prime}\left( {k_{i} } \right)}} < 0$$

(6)

As indicated, capital investment in locality $i$ is negatively associated with the capital tax rate selected by local governments. The problem of each locality is to choose $\tau_{i}$ independently so as to maximize its objective function (Eq 3), subject to its budget constraint (Eq 4). The resulting equilibrium of optimal tax rate on capital $\tau_{i}^{*}$ appears at the point that the marginal benefit from public expenditure equals the marginal cost of collecting public funds from capital taxation. That is,

$$\alpha_{i} v^{\prime}\left( {e_{i} } \right) = \frac{{k_{i} }}{{k_{i} + \tau_{i} \frac{{\partial k_{i} }}{{\partial \tau_{i} }}}}$$

(7)

Finally, applying the Envelope Theorem to equation (Eq 7), we have,

$$\frac{{\partial \tau_{i}^{*} }}{{\partial w_{i} }} = \frac{{k_{i} }}{{\alpha_{i} }}\frac{{ - v^{^{\prime\prime}} \left( {e_{i} } \right)}}{{\left( {v^{^{\prime}} \left( {e_{i} } \right)} \right)^{2} }}\frac{1}{{\frac{{\partial k_{i} }}{{\partial \tau_{i} }}}} < 0$$

(8)

Therefore, equation (Eq 8) provides the following hypothesis that is consistent with our main argument in the text.

Hypothesis B1

In equilibrium, an increase in land lease revenue available to local governments ($w_{i}$) will reduce the capital tax rate selected by local governments ($\tau_{i}^{*}$).

In the context of China, this hypothesis conveys a clear message that windfall revenue, in the form of land lease revenue, is likely to intensify local tax competition and, hence, lead to local governments selecting a lower tax rate.³⁰