The response of small firms to VAT thresholds : evidence from China

This paper explores the response of small-scale value-added tax (VAT) taxpayers to the changes of VAT thresholds and VAT tax rates. Using firm-level tax administrative data for the period between 2005 and 2012, we find that firms tend to bunch around the VAT thresholds. We also find that commercial firms and firms with higher value-added ratios are less likely to register as general taxpayers. We provide an estimate of the elasticity of the VAT tax base in the range of 0.027 and 0.05.

Keywords: Value-added Tax; small-scale taxpayers; bunching

I. Introduction

The value-added tax (VAT) is the largest source of tax revenue for China, accounting for approximately 30% of China's total tax revenue over the past 20 years. Countries around the world set various VAT thresholds (Keen and Mintz [4]), setting an appropriate VAT threshold is important for the overall tax system. Higher thresholds mean lower compliance costs for the tax authority and small firms. However, a higher VAT turnover threshold reduces VAT revenue. In this paper, we investigate how the current VAT thresholds affect firms in China, and explore the impact of changes in VAT tax rates on firm behaviours.

Table 1. Voluntary registration

CE	−0.203***	−0.130***
	(0.0211)	(0.0219)
VAR	−0.322***	−0.374***
	(0.0341)	(0.0351)
Wage		−0.0417***
		(0.00885)
Asset		0.0619***
		(0.00742)
Constant	2.822***	2.702***
	(0.0865)	(0.105)
Year FE	Y	Y
Observations	35,708	32,147

1 Robust standard errors are in parentheses.

According to the sales volume and soundness of the accounting system, VAT taxpayers can be categorized as either general or small-scale taxpayers. The main difference between the two types of taxpayers is the calculation method of tax liabilities. General taxpayers are taxed at a higher tax rate only on their value-added, while small-scale taxpayers are taxed at lower rates on their total sales. General taxpayers are subject to a standard 17% VAT rate, with a reduced rate of 13% for certain products.[1] Since 2009, small-scale taxpayers have been subject to a simplified calculation method of a 3% rate on total turnover, a reduction from the prior 4% rate on commercial enterprises and 6% rate on manufacturing enterprises. No deduction is allowed for small-scale taxpayers.

The VAT sales threshold is the main criterion for dividing these two categories. If the VAT sales of a taxpayer lower than the threshold, the firm could maintain the small-scale taxpayer status. The small-scale taxpayer with a sound accounting system can also voluntarily register as a general taxpayer, which usually depends on their consideration of lowering tax burden, especially for enterprises with lower value-added. Before 2009, the threshold was RMB 1 million for manufacturing enterprises and RMB 1.8 million for commercial enterprises. The VAT sales threshold for manufacturing enterprises changed to RMB 500,000, while the threshold for commercial enterprises changed RMB 800,000 after 2009.

In general, the special VAT invoices and the enforcement system motivate firms to register as general VAT taxpayers. Although the VAT rate for general taxpayers is much higher than the rate for small-scale taxpayers, a firm with a high value-added percentage might still benefit from general taxpayer status. As a general taxpayer, a firm is able to deduct input taxes that they paid from their output taxes due. So, firms have an incentive to obtain VAT receipts when making purchases, thereby promoting greater tax compliance. Once the sales of a small-scale taxpayer reaches the minimum VAT threshold, a small-scale taxpayer will be advised by the tax authority to register as a general taxpayers. Upon the approval of the tax authority, small-scale taxpayers with sound accounting that can provide accurate taxation information can register as a general taxpayer.

When sales pass the minimum VAT threshold, firms face a decision: whether to register as a general taxpayer or remain a small-scale taxpayer by manipulating (or restricting) their sales. The financial cost, as a percentage of the total turnover, is much higher for small firms. As to lower tax burden, it may be worthwhile for firms with a turnover just above the thresholds to underreport revenue to give up input tax reimbursement and remain as small-scale taxpayers. We expect a significant percentage of firms to bunch just below the VAT minimum threshold (Devereux, Liu, and Loretz [2]; Liu and Lockwood [5]; Harju, Matikka, and Rauhanen [3]). Using the fact that firms tend to bunch around cut-off points sheds light on firm behaviour in response to tax policy. In this study, we try to investigate the impact of changes in the tax rates and thresholds on firm behaviour.

This paper makes several contributions. First, this paper adds to the existing evidence regarding developing countries by using Chinese data to investigate changes in VAT thresholds on firm behaviour. Second, this paper provides information on changes in tax rates for two categories of firms, offering more robust results by examining the effect of both VAT threshold changes and tax rate changes.

II. Data and empirical methods

Data

In this paper, we use administrative tax investigation data from China for the year between 2005 and 2012 to study firm behaviour in response to the VAT tax changes described above.[2] This dataset contains firm-level tax and financial information. It covers all provinces, with more firms in the relatively more developed provinces. We keep only the observations of general VAT taxpayers and small-scale taxpayers,[3] dropping records of firms with negative sales. The key variable in our analysis is tax liability.[4]

Empirical method

To quantify firms' behavioural responses to the VAT threshold, we employ the nonparametric approach by Chetty et al. ([1]), which determine the counterfactual density by running a local polynomial regression on binned data (bin width is 30 k in this paper). The counterfactual density can be derived by excluding dummy variables indicating the range of VAT sales may be affected by the threshold. By comparing the true density and the counterfactual density, we can estimate the excess mass around the bunching point. Because manufacturing firms and commercial firms face different thresholds, we estimate excess bunching mass separately.

The key for using the bunching method is to find an excess mass of firms located just below or at the threshold. We need to calculate a counterfactual distribution of reported revenue in the absence of a VAT threshold and compare it with observed distribution. Counterfactual density is derived by fitting a polynomial to the empirical density excluding observations around the thresholds. Firms around the VAT thresholds tend to do tax planning. To measure the response of firms to changes in VAT thresholds, we use the bunching approach utilized by Saez ([7]). Applying the aforementioned method, we can graph both the empirical distribution and counterfactural distributions, as shown in Figure 1.[5]

PHOTO (COLOR): Figure 1. Frequency plots

To analyse the determinants of voluntary registration as a general VAT taxpayer, we use the following regression equation:

(1)

Graph

$R_{it}=\alpha +{\beta }_{1}C{E}_{it}+{\beta }_{2}VA{R}_{it}+{\beta }_3{X}_{it}+{\eta }_t+{\epsilon }_{it}$

where

Graph

$R_{it}$ represents whether firm

Graph

$i$ registered as a general taxpayer in year

Graph

$t$ , it takes the value of 1 if the firm is registered as a general taxpayer and 0 if the firm is a small-scale taxpayer.

Graph

$C{E}_{it}$ is a binary variable that takes the value of 1 if firm

Graph

$i$ is a commercial enterprise in year

Graph

$t$ and 0 if firm

Graph

$i$ is a manufacturing enterprise in year

Graph

$t$ .

Graph

$VA{R}_{it}$ is the value added ratio, which is the ratio of valued added to output, for firm

Graph

$i$ in year

Graph

$t$ . Wages and assets are also included as other firm-level controls. Year fixed effects

Graph

${\eta }_t$ are added.

Graph

${\epsilon }_{it}$ is the error term. The regression results are shown in Table 1.

Tax rates and bunching

If a firm registers as a general taxpayer, it pays

Graph

$T_g={\tau }_{g}VAR$ , where

Graph

${\tau }_g$ is the tax rate on general taxpayers. If a firm is a small-scale taxpayer, a firm pays

Graph

$T_s={\tau }_s(1+VAR)$ , where

Graph

$T_s$ is the tax that a small-scale taxpayer pays and

Graph

${\tau }_s$ is the tax rate for small-scale taxpayers. The indifference

Graph

$VAR$ for registering either as a general taxpayer or a small-scale taxpayer occurs when

Graph

$T_g=T_s$ , which yields the indifference VAR.

(2)

Graph

$VAR=\frac{{\tau }_s}{{\tau }_g-{\tau }_s}$

If the value-added ratio is lower than the critical point, the firm will register as a general taxpayer. For example, the indifference ratio is 21.4% for general taxpayer that subject to a 17% VAT rate after 2009. If the ratio is greater than 21.4%, a firm pay less tax by chosen being as a small-scale taxpayer.

The indifference ratio is 54.5%for manufacturing firms before 2009, while the ratio is 30.8% for commercial firms before 2009.[6] This means that a manufacturing firm will choose to be a small-scale taxpayer if the value-added ratio is greater than 54.5%, which is unlikely for a manufacturing firm. It is apparent from Figure 1(a) that a few firms bunched around the thresholds between 2005 and 2008, while Figure 1(b) shows that many firms bunched just below the threshold after 2009. Similar patterns exist for commercial firms, which can be seen in Figure 1(c) and 1(d).

Our results demonstrate that the elasticity is 0.027 for manufacturing firms and 0.05 for commercial firms, which are larger than the elasticities of developed countries for the United Kingdom (L. Liu and Lockwood [5]) and Finland (Harju, Matikka, and Rauhanen [3]).

III. Results

Value-added ratio and type of firm are the two main determinants of voluntary registration. Compared with manufacturing enterprises, commercial enterprises are less likely to register as general taxpayers. There are two reasons for this result. First, the value-added rate is a key factor that a firm considers on whether to register as a general taxpayer, which is confirmed by the significantly negative regression coefficients for voluntary registration on value-added rate. Commercial firms usually have a higher value-added rate relative to manufacturing firms, thus commercial firms can benefit from being a small-scale taxpayer. Therefore, the bunching mass for commercial firms is greater than the manufacturing firm.

Second, the proportion of sales to the final consumer is another key factor. Consumers buy the goods for final use, so they do not need to get VAT invoices to deduct input tax, which is unlike selling goods to firms use as intermediates. Commercial enterprises sell more goods and services to final consumers than manufacturing firms, they do not have to register as general taxpayers to issue special VAT invoices. Contrarily, manufacturing firms tend to register as general VAT taxpayers since only a small portion of goods sales to final consumers, they need to issue VAT invoices to the firm purchaser that requires special VAT invoices to deduct output tax.

The identified tax elasticity implies that firms are bunching around the VAT threshold. Our analysis shows that the elasticity of the VAT tax base is 0.027 for manufacturing firms and 0.05 for commercial firms, which is fit the above conclusion, commercial firms show more responses than manufacturing firms.

Acknowledgments

The authors appreciate the feedback from participants in the 2016 Southern Economic Association annual meeting, Wuhan University workshop, ZUEL and CUFE workshops.

Disclosure statement

No potential conflict of interest was reported by the authors.

1 Chetty, R., J. N. Friedman, T. Olsen, et al. 2011. "Adjustment Costs, Firm Responses, and Micro vs. Macro Labor Supply Elasticities: Evidence from Danish Tax Records. " The Quarterly Journal of Economics 126 (2): 749 – 804. 2 Devereux, M. P., L. Liu, and S. Loretz. 2014. " The Elasticity of Corporate Taxable Income: New Evidence from Uk Tax Records." American Economic Journal: Economic Policy 6 (2): 19 – 53. 3 Harju, J., T. Matikka, and T. Rauhanen. 2019. " Compliance Costs Vs. Tax Incentives: Why Do Entrepreneurs Respond to Size-based Regulations? " Journal of Public Economics 173 : 139 – 164. doi: 10.1016/j.jpubeco.2019.02.003. 4 Keen, M., and J. Mintz. 2004. " The Optimal Threshold for a Value-added Tax." Journal of Public Economics 88 (3): 559 – 576. doi: 10.1016/S0047-2727(02)00165-2. 5 Liu, L., and B. Lockwood, et al. 2016. " Vat Notches, Voluntary Registration, and Bunching: Theory and Uk Evidence." Review of Economics and Statistics, forthcoming. 6 Liu, Y., and J. Mao. 2019. " How Do Tax Incentives Affect Investment and Productivity? Firm-Level Evidence from China." American Economic Journal: Economic Policy 11(3): 261–91. 7 Saez, E. 2010. " Do Taxpayers Bunch at Kink Points? " American Economic Journal: Economic Policy 2 (3): 180 – 212. Food grains, natural gas are examples of the products subject to a 13% VAT rate. The tax rate is 0% for taxpayers exporting goods. See Y. Liu and Mao ([6]) for detail information on the tax return data of China. We drop non-VAT taxpayers, VAT taxpayers of members of consolidated tax paying entity, and VAT taxpayers of the tax consolidating headquarters. Taxable income is calculated based on the general sales of goods and services for general taxpayers, while small-scale VAT taxpayers pay their taxes according to their sales of dutiable goods and services calculated using the simplified method. More detailed method and results are available upon request. These two rates are calculated based on a regular 17% tax rate, which might be different for firms in the agriculture industry.

By Enhui Kou; Yukun Sun and Linfeng Yue

Reported by Author; Author; Author