Fiscal reform is integral to China’s broad reform plans and key for achieving more inclusive and sustainable growth. The 13th Five-Year Plan commits to deepening tax reforms to facilitate economic transition, and includes measures to collect revenues more efficiently, improve the fiscal framework, and strengthen local government finances. The authorities decided to establish the foundations of reforms by 2016 and modernize the fiscal framework by 2020.

Fiscal reform is integral to China’s broad reform plans and key for achieving more inclusive and sustainable growth. The 13th Five-Year Plan commits to deepening tax reforms to facilitate economic transition, and includes measures to collect revenues more efficiently, improve the fiscal framework, and strengthen local government finances. The authorities decided to establish the foundations of reforms by 2016 and modernize the fiscal framework by 2020.

Inefficiencies in the tax system are an important obstacle on China’s path toward a new growth model. And although the overall tax burden is lower than the Organisation for Economic Co-operation and Development (OECD) average, fiscal policy plays a very modest redistributive role. This is reflected in a relatively heavy reliance on indirect taxes and the largely regressive design of direct taxation, especially in social security contributions. Local governments often rely on costly and inefficient revenue sources to finance spending, in part because their own sources are insufficient to cover their spending responsibilities. Tax policy reforms in the coming years should help better align them.

Some progress has been made in tax policy reforms in recent years (Table 2.1). In indirect taxation, for instance, the government in 2016 completed the extension of value-added tax (VAT) to services, replacing the distortive and cascading business tax. Fundamental reform of corporate taxation in 2008 created more neutrality and improved revenue performance. Legislative work is now underway to introduce a recurrent property tax that is expected to fund local government. Yet important challenges remain. For instance, considerable work is needed to make individual income tax more progressive and raise the revenue performance of such taxation. The actual implementation of a recurrent property tax and environmental tax is also challenging.

This chapter reviews the tax system and compares it with that of other countries. It elaborates on challenges in direct and indirect taxation, with the aim of identifying options that can help improve the efficiency and equity of taxation and contribute to a successful transition toward a model of sustainable and inclusive growth.

Table 2.1.

Recent Tax Reform Initiatives in China

article image
article image
Sources: State Council; Ministry of Finance; and State Administration of Taxation.Note:The 17 percent value-added tax (VAT) rate is levied on tangible goods, although some items are subject to a 13 percent rate. The 11 percent rate is levied on transportation services, and the 6 percent rate is levied on research and development, technological, cultural, logistical, and consultative services.

Income from sole proprietorship and contractual work carries a tax rate of 5–35 percent in five progressive scales.

Employees and employers also have social security contributions that range between 10 and 30 percent of employment income, respectively. Employment incomes are subject to a threshold between 60 percent and 300 percent of the previous year’s local average income determined by local statistics departments.

A tax of 0.5–1.2 percent was introduced for single-family houses and high-end apartments in Chongqing, while a property tax of 0.6 percent was levied on newly purchased properties for residents with two or more properties and newly purchased properties for nonresidents in Shanghai.

Overview of China’s Tax System

The overall tax burden in China is comparable to those in other OECD economies in Asia, but it relies more heavily on indirect taxation. Tax revenue in China accounted for about 20 percent of GDP in 2015. This is much lower than the average 34 percent of GDP across the OECD but similar to Chile, Korea, and the United States. Indirect taxes on goods and services, including VAT and business turnover tax, account for more than half of tax revenue, compared with about one-third of revenue in other OECD countries (Figure 2.1). Direct taxes make up a relatively small percentage of tax revenue, especially for individual income tax, at 6 percent of revenue, much less than the 25 percent average across the OECD. China relies relatively heavily on nontax revenues, such as fees, fines, and income from state-owned enterprises. To make up for revenue shortages, local governments also frequently depend on revenue from land sales. Once adjusted for nontax revenue and net land sale proceeds (about 2¾ percent of GDP), the overall tax burden in China is similar to the tax-to-GDP ratio for other economies in the Asia and Pacific region, such as Australia, Japan, and Korea. Nominal revenue in China has grown about 20 percent year over year on average over the past decades. This exceeded nominal GDP growth for many years, although more recently—and as the economy is slowing—it has grown in line with nominal GDP growth.

The redistributive effect of taxes and transfers is relatively limited in China. In the mid-1980s, income inequality as measured by the Gini coefficient was low at around 0.3. Since then, it has continued to rise, in recent years to nearly 0.5.1 This widening income inequality contrasts with the experiences of other middle-income countries, where inequality has largely remained constant over the same period (Cevik and Correa-Caro 2015). Fiscal policy also appears to contribute relatively little to reducing income inequality. For instance, if one looks at the difference between the Gini coefficients of market income (that is, before taxes and transfers) and disposable income (after taxes and transfers), there is virtually no difference. Hence, fiscal redistribution in China is mostly absent, unlike in other countries (Clements and others 2015). This reflects a near absence of progress in reforming direct taxes and transfers and great reliance on indirect taxes. On the spending side, quality and access to social services and the social safety net vary according to hukou (household residency) status, which reinforces the disparities and weakens the redistributive effect of fiscal policy.

Local government revenues and spending obligations are often misaligned. Broadly speaking, tax revenue in China is assigned in one of three ways (Table 2.2):

  • Central government taxes used exclusively to finance central government spending. This includes taxes related to foreign trade, such as duties, VAT collected on imports, and taxes on luxury goods.

  • Local government taxes under exclusive control of local governments, such as property-related taxes and business operation taxes.

  • Shared taxes between central and local government, including direct taxes such as personal and corporate income taxes and VAT (excluding trade-related VAT). Revenue sharing can take various forms, but usually is based on prescribed revenue-sharing ratios.

Table 2.2.

Tax Sharing Arrangement between Central and Subnational Governments

article image
article image
Sources: CEIC; and IMF staff estimates. Note:… = not available.

Following the completion of value-added tax reform, the tax sharing arrangement is temporarily split equally between central and local governments.

Estimates based on social security contributions ratio to GDP.

The sharing of tax revenues, if excluding social security contributions, will be about 50.4 percent and 49.6 percent, respectively, for central government and local governments.

Most local governments often face significant revenue shortfalls. Local government tax revenue, including shared taxes, accounts for nearly half of government revenue. In contrast, local government spending obligations account for roughly two-thirds of government expenditure (Wang and Herd 2013). This misalignment in revenue and spending is only partially corrected by transfers from the central to local governments and forces many local governments to rely on other sources of financing to meet spending obligations (Figure 2.2). For instance, in the past many sold land and engaged in off-budget borrowing through local government financing vehicles. Reform is therefore essential to better align local revenue and spending obligations.2

Figure 2.2.
Figure 2.2.

Local Government Revenue and Tax Sharing System

Sources: CEIC; and IMF staff calculations.Note: VAT = value-added tax.

Direct Tax Reforms

This section elaborates on reforms in the individual and corporate income tax system. It starts with a brief reflection on the current system, followed by an assessment of reform options. Income tax theory provides the conceptual underpinnings of the assessment (Box 2.1).

Policy Guidance from Income Tax Theory

Optimal income tax theory seeks the most efficient way to collect revenue, taking into account distortions to behavior. A central result of the literature is that lump-sum taxes constitute the most efficient instrument because they do not involve any change in the marginal return to work. For obvious reasons, equity concerns have made this conclusion impractical; under a uniform lump-sum tax, all taxpayers are required to remit the same amount of taxes regardless of their ability to pay, which is viewed as unfair and politically unacceptable. In response to this lack of practical guidance, more detailed frameworks have been developed, and can be used to study tax policy in more realistic settings.

A key insight of the theoretical literature on optimal income taxes is that governments face a trade-off between equity and efficiency concerns. Equity concerns are typically addressed by assuming that the government values individuals’ own welfare, but the social welfare weight placed on individuals generally decreases with ability and income. Society will therefore prefer, everything else equal, to redistribute after-tax income relatively more to those with less before-tax income. This requires that the tax system incorporate elements of progressivity. Efficiency concerns are typically modeled as a response of individuals’ labor supply—and other margins of avoidance such as fringe benefits—to changes in marginal tax rates. The extent of the response is typically expressed as the elasticity of taxable income. This parameter is meant to capture the simple intuition that collecting revenue through increased marginal tax rates will generally lead people to work less and seek tax-favored forms of remuneration.1 The optimal structure for the individual income tax usually involves marginal tax rates that follow a U-shaped pattern. Following the seminal work of James Mirrlees,2 theoretical studies have found that the basic structure includes four main features: (1) a lump-sum amount guaranteed to all taxpayers; (2) high marginal tax rates at the bottom of the income distribution to ensure that only the low-paid benefit from the lump-sum amount; (3) low marginal tax rates for middle-income earners; and (4) increasing marginal tax rates for those with higher incomes. The prescription for high marginal tax rates at the bottom of the income distribution is perhaps a surprising feature of such a system since policymakers usually consider progressivity an important feature of any sound tax system. However, high marginal rates ensure that only the neediest benefit from the guaranteed amount provided under (1). The lump-sum amount combined with high marginal rates still ensures that average tax rates increase as incomes rise. Under such a stylized tax system, the redistribution achieved through the lump sum will strictly be targeted to the lowest earners.

These basic features have intuitive appeal. Such a structure ensures first that all taxpayers enjoy a minimum standard of living, which is guaranteed by a universal and refundable tax credit—or alternatively by welfare benefits provided outside the individual income tax. Maintaining well-targeted transfer payments in turn allows lower marginal tax rates across the board and smaller overall tax distortions. This feature of optimal tax policy can coexist with initially low and increasing headline marginal tax rates by relying on a phaseout of personal tax credit or through means-tested benefits. This combination then leads to higher effective marginal tax rates on low- to middle-income earners.

Low marginal tax rates for the middle-income group ensure distortions are minimized for most taxpayers. Income distributions are usually bell shaped, so that middle-income earners constitute by far the largest group of taxpayers. Efficiency considerations will therefore favor low marginal rates for most workers. Finally, the last main feature of the optimal structure is the progressivity of marginal rates after the modal income. Increasing marginal rates on the biggest incomes ensures that individuals with the highest earning potential will contribute relatively more to revenue collection.

Despite the inclusion of income heterogeneity among individuals in optimal tax models, their direct applicability is still limited. These models generally do not incorporate other dimensions of heterogeneity, such as family structure or other indicators of need, such as number of children or disability status. The general framework described above also considers individuals in a one-period static model. This precludes the study of life-cycle considerations in labor supply and other behavior such as saving and investment in education. It also implies that the model is most applicable to the taxation of income from labor, as opposed to capital income, where saving and intertemporal substitution of consumption introduce additional elements of complexity.3

1An overview of the literature and a detailed presentation of the main results can be found in Piketty and Saez 2013.2Mirrlees 1971.3More on the optimal taxation of capital can be found in Diamond and Banks 2010.

Individual Income Tax

Current Situation and Issues

Individual income tax in China relies on a schedular system, where different sources of income are taxed according to varying structures and rates (Table 2.3). Income falls into 11 categories, including wages and salaries, unincorporated business income, income from personal services, interest, dividends and bonuses, and the leasing of property, among others. The tax unit of the individual income tax system is the individual. No provisions exist for couples to be taxed jointly, and tax credits are not available for specific family circumstances, such as the presence of dependents. The assessment period also differs across sources of income. Wages and salaries are assessed on a monthly basis, so that the tax withheld according to the rate schedule is final for most workers with only one place of employment. Individuals’ business income is taxed annually after allowable costs, expenses, and losses have been deducted.

Table 2.3.

Individual Tax Rates Applicable to Annual Wages and Salaries

article image
Source: International Bureau of Fiscal Documentation.Note: A standard deduction of RMB42,000 is allowed in calculating taxable income.

The tax schedule for employment income is broadly progressive and comparable to those of OECD economies, but with wide tax brackets that start at higher incomes (Table 2.4). The highest marginal tax rate is 45 percent at the very top of the income distribution (about 35 times the national average wage, or 17 times the average urban wage) and applies to few income earners (Figure 2.3). In contrast, on average among OECD countries, top marginal tax rates are imposed on individual income at about four times the national average wage. Moreover, according to recent calculations based on household survey data, over 80 percent of workers are not liable to pay individual income tax (Lam and Wingender 2015). This is due to a high basic personal allowance of RMB42,000, twice the average national wage. In comparison, the average basic personal allowance in OECD countries is about one-quarter of the average national wage.

Table 2.4.

Individual Tax Rates Applicable across Income Sources

article image
Source: International Bureau of Fiscal Documentation.

Individual income tax on unincorporated business activity and partnerships is also progressive, but with different rates than those on wages. Another income category defined in the legislation and taxed at different rates consists of income derived from labor service (such as medical, legal and accounting activities, consulting, and interior design). Dividends are taxable in general, although reduced rates apply to shares of companies listed on stock exchanges in China that are held for more than a month. Interest income is taxable, but interest from savings deposits and government bonds is exempt. Rental income is taxed at 10 percent after a 20 percent deduction. Capital gains are generally taxed at a 20 percent rate, although gains on the sale of shares of Chinese listed companies and houses that owners have occupied for two years or more are exempt.

Pensions and welfare benefits are exempt from taxation, and social security contributions are deductible. Since 2014, employee contributions to private company pension funds are deductible up to 4 percent of salary. Employer contributions are also exempt up to prescribed limits. Income derived from these tax-exempt contributions is taxable in retirement, although accrued gains are not. A pilot project launched in Beijing, Shanghai, and 29 other major cities on January 1, 2016, also allows employees, sole proprietors, and partners to deduct up to RMB2,400 annually for premiums paid to a qualified commercial health insurance scheme.

In addition to tax on wages and salaries, employees and their employers must contribute to social security for pensions, medical insurance, and unemployment insurance. Employee contributions are withheld from wages and remitted directly by employers, on average at about 11 percent of gross wages of employees (or net of income tax at nearly 9 percent) with the exact percentages varying between provinces and localities. Although a nominal flat rate is applied, a minimum employee contribution is also required and based on some imputed value of earnings. The imputed minimum earnings for workers, in turn, are set at 60 percent of the previous year’s average wage in a locality. Estimates put the earnings of about 30 percent of the urban labor force in several large cities below the 60 percent threshold (Cai, Du, and Wang 2011). This feature of the schedule for social security contributions, along with a cap on contributions set at three times the average wage in a locality, implies that the effective contribution rate is very regressive in employee income (Brys and others 2013; World Bank and Development Research Center of the State Council 2013).

Proposed Reforms

The current schedular approach appears to be unfair and distortive. Taxing income under different categories creates considerable fragmentation and causes inequities that are often regarded as unfair. Distortions arise as individuals adjust their activities and report toward more tax-favored options. In that context, the Chinese authorities are discussing alternative models for income tax design, most notably global income tax and dual income tax systems.

The fragmentation of the present system for collecting individual income tax has led some to argue for the adoption of a global income tax system. This would mean that all 11 categories of income would be consolidated for each individual (or household) under a single tax base, and taxed using a marginal rate schedule based on ability to pay. A global income tax system is often viewed as the theoretical ideal for many countries. Many argue that it achieves both horizontal and vertical equity and avoids the tax arbitrage that can arise from different sources of income being taxed under different schedules. Despite its appeal, however, most advanced economies have moved away from a global income tax system over the past few decades. Indeed, even arrangements that closely resembled global income tax systems were often found to be inefficient and inequitable in practice.3 For instance, high marginal tax rates on capital income were seen as highly inefficient because they distort decisions about savings and investment. The systems were also inequitable as evasion and avoidance of taxes, especially by wealthy taxpayers, made the distribution of the tax burden much less progressive than it appeared in theory.

Global income tax systems were often costly to administer, as capital income is difficult to measure accurately, and all incomes had to be attributed to individuals (or households), necessitating large-scale filing of tax returns. To simplify their individual tax systems (reducing distortions and administrative complexity), many countries have therefore adopted a schedular approach to taxing labor and capital income.

A dual income tax system provides a more solid framework for a modernized individual income tax in China. Many OECD countries have adopted variants under which all types of income are allocated to one of two broad categories: labor or capital income. In China, this would be a significant step toward a fully integrated approach to taxing income, since the number of schedules would decline from 11 to two. Starting from the current individual system, a natural classification in China would allocate to a comprehensive labor income category all income from (1) wages and salaries; (2) the business of industrial and commercial operators; (3) business operations contracted or leased from enterprises; (4) labor service; (5) remunerations to authors; (6) royalties; (7) accidental income; and (8) other types of earnings. All of these income sources would be added together and taxed under a single comprehensive rate schedule. The remaining types of income identified in the individual tax legislation, including income from (9) interest, dividends, and bonuses; (10) leasing of property; and (11) transfer of property, would be classified as capital income.

The labor income tax schedule under a dual income tax system should include progressive marginal tax rates. Such a system would present many advantages from policy and tax administration perspectives. Income redistribution can be achieved by increasing marginal tax rates for high-income earners. Taxing all sources of labor income under the same schedule is also regarded as improving fairness. The design of the labor income tax schedule still leaves scope for many specific design options. These include the following:

  • Tax unit. Economically and administratively it would be desirable to continue to tax individual income, as opposed to family income. Family-based taxation can be more equitable since households with the same incomes would be taxed the same, regardless of the distribution of income between spouses. Yet this approach is associated with important costs. First, under family-based taxation, individuals are taxed according to marital status, so the system is not neutral in that respect. Second, individual-based taxation is simpler to administer since withholding depends only on earnings and not on marital status, which the tax administration must verify. Third, high tax rates on secondary earners tend to discourage both labor-force participation and work effort, potentially undermining economic efficiency and growth.

  • Rate structure. Labor income would be subject to a progressive marginal tax rate schedule but with scope for simplification. The number of marginal tax rates itself is not necessarily a source of added complexity, and reducing the number of rates can point to ways to simplify the system further. A schedule with marginal rates of 10, 20, and 35 percent, for instance, could strike a good balance between equity and efficiency. This schedule would raise the current minimum rate of 3 percent and reduce the top tax rate of 45 percent, which is higher than the average top rate in Asian countries. A somewhat higher rate at the bottom would efficiently expand the tax base—the first bracket collects revenue from all taxpayers with earnings above a certain level—while generating minimal efficiency costs on lower earners.

  • Threshold. The current approach is to deduct a personal allowance of RMB42,000 a year—RMB3,500 a month—from taxable wages and salaries. While this shields the lowest-income earners from having to pay personal income tax, it also favors people with higher incomes, since the value of the allowance depends on marginal tax rates. The value of the basic personal allowance is, for instance, RMB1,260 for an individual in the first tax bracket (3 percent of RMB42,000), but RMB18,900 for individuals in the highest tax bracket of 45 percent—15 times more than for those with the lowest taxable incomes. An alternative approach that maintains favorable tax treatment for the low-paid group is to transform basic personal allowances from a deduction into a tax credit. This change would ensure that, for a given marginal tax rate structure, the allowance is constant across all incomes and does not benefit high-income earners disproportionately.

  • Targeted allowances. The current individual tax system is simple to administer but difficult to tailor to the specific circumstances of taxpayers. It would in principle be possible to achieve more equity using selected targeted allowances that better reflect ability to pay. However, targeted allowances should be carefully considered: they often cause significant distortions, usually do not achieve their intended purpose, and are extremely difficult to repeal. Targeted allowances also often disproportionately benefit higher-income households, which would be especially true in China considering that only the highest-income individuals pay any tax under the current individual income tax system. To support equity considerations, China could consider introducing a targeted tax credit for children. This could be a sensible approach since the cost of raising children significantly affects the ability to pay tax, and a child tax credit would have few unintended consequences. It would not be advisable to introduce other allowances tied to spending on activities that are sometimes given preferential tax treatment (such as charitable giving, medical expenses, mortgage interest, and other expenditure), which are very easy to abuse and exceedingly costly to enforce.

  • Annual assessment. Defining taxable labor income on an annual basis would result in a more equitable system. This would ensure a consistent definition of taxable income, in particular between business and labor income, which would be consolidated as labor income under a dual income tax approach. It would also ensure equal treatment of workers with different earning patterns; for example, due to seasonal employment. Tax collection can remain largely based on monthly withholding, although end-of-year filing will increase somewhat for some taxpayers (such as professionals).

Under a dual income tax system, investment income is taxed uniformly at a flat rate. In principle, rates should be equal across all sources of investment income to maintain tax neutrality. In practice, however, several factors limit the desirability of this specific feature. Income from dividends, interest from deposits and government bonds, capital gains realized on real property, shares, and rental income would all be taxed at a proportional rate. The proportional tax rate on investment income should not differ too much from the top marginal tax rates applicable to labor income, taking into consideration any tax burden from corporate income tax. Lower effective tax rates on income from capital would create a strong incentive to recharacterize income from labor as income from capital, a problem that is especially relevant for small businesses. This incentive exists in many tax systems, especially when considering payroll or social security taxes that substantially increase effective marginal tax rates on labor income but leave the tax rates applied to income from capital unchanged. Moreover, given that investment income disproportionately accrues to the highest-income households in China, it would be appropriate to set this rate in relation to the top marginal rate on labor income. Maintaining a flat rate enables the continued use of final withholding, which ensures collection is administratively efficient.

  • Dividends. Dividends are currently subject to individual income tax through final withholding from shareholders, after being paid out of profits already taxed as corporate income. The resulting double taxation of income from equity-financed investment is often an important source of nonneutrality because, unlike dividends, interest expenses can be deducted from corporate profits—and thus taxed only once. A simple way to mitigate these distortions is to impose lower withholding tax rates on dividends received. A reduced rate simply needs to ensure that dividends are taxed at the desired effective rate. A lower dividend tax rate of 10 or 15 percent, say, would be appropriate in China, given the double taxation of distributed corporate profits.4 Since dividends disproportionately accrue to the top 10 percent of household incomes, this would imply broad consistency with a proposed top marginal tax rate of 35 percent on labor income. A slightly higher combined tax rate on dividends than on interest income would also achieve additional redistribution, since investment income of the lowest-income households is mostly composed of interest.

  • Capital gains. Capital gains should in principle be taxed in the same way as other forms of investment income—on an annual accrual basis. In practice, however, taxing capital gains raises a number of practical problems. First, capital gains, especially gains from some forms of innovative financial instruments (such as derivatives), can be difficult to identify and measure. Second, compliance and enforcement in taxing capital gains from most assets on an accrual basis is difficult, and it would be undesirable from the perspective of investment neutrality to tax on an accrual basis only those assets where accrued gains would be easy to measure (for example, assets traded on public exchanges). For these reasons, virtually all systems across the world where capital gains are taxed do so upon realization. However, this can distort the decision of when to realize gains (or losses), which can result in inefficiencies and distort asset portfolios and risk sharing. As a result, taxes on capital gains typically make up only a small share of tax revenues—less than 1 percent on average in OECD countries. However, their taxation remains important to protect the revenue base of other sources, given the ease of transforming other forms of investment income (and in some cases labor income) into capital gains. A somewhat reduced tax rate can address several issues involved in taxing capital gains. It can mitigate the “lock-in” effect associated with the taxation of realized gains, decrease the potential tax advantage of realizing capital losses earlier than otherwise, and better reflect the taxation of the underlying appreciation over time. A lower tax rate also reduces the double taxation of enterprise profits—just as for dividends—since it is the after-tax profit that tends to be capitalized in share prices. Finally, taxing capital gains at a lower rate compensates investors for the nondeductibility of losses when collected under a final withholding regime.

  • Interest, royalties, and rental income. These incomes could also be taxed at a rate of 20 percent, say. While this would be somewhat lower than the top marginal tax rate on labor income and the combined tax burden on dividends or capital gains, taxing incomes at much higher rates may not yield more revenue in light of their elastic base, which itself is in part due to international mobility. Exemptions for interest on government bonds and on deposit account savings with a Chinese bank should be repealed. Rental income should be taxed at about the same (effective) rate as other sources of investment income, following the neutrality principle. Currently, preferential treatment is given to gross rental income, which is taxed at a reduced individual income tax rate of 10 percent, with an additional cost imputation of 20 percent, making the effective tax rate on rental income 8 percent. This type of tax preference favors investment in real estate and can lead to distortions and overinvestment in the sector. Costs for the provision of rental services (such as financing or maintenance) might warrant a somewhat lower rate than that applied to interest income. With the 20 percent cost imputation maintained, the effective tax rate on rental income could be brought to 16 percent to achieve rough neutrality.

Current Situation and Issues with Social Security Contributions

Social security contributions are strongly regressive in China. For most households, social security contributions constitute over 90 percent of direct tax liabilities. The average statutory tax rate across all sources of income, including business and capital income, also places a heavy burden on the poorest households. This is also reflected in the average effective tax rate (measured as the difference between gross market income and net disposable income after taxes and transfers) reported by households. As a result of both the imputed earnings base for social security contributions and the very high levels at which marginal tax rates on an individual’s income start to be applied, the total tax burden on employment income is very high for the low-paid. Data from the household income survey suggest that households in the lowest income quartile pay a much higher effective tax rate than higher-income households (Lam and Wingender 2015). The effective tax schedule is regressive at the range of employment income below 70 percent of average wage income (in Figure 2.3, for example, the tax wedge in China is shown to be much higher than average OECD effective rates for single taxpayers and working couples with two children).

Proposed Reform of Social Security Contributions

Reform of social security contributions focuses on the efficiency, portability, and sustainability of the social security system (see also Chapter 5 on pension reform). The current structure and, in particular, the imputed taxable earnings for low-income earners reduce labor market flexibility and tend to favor informality and self-employment, which are not broadly covered by the contribution rules. This has probably led to significant numbers of low-income workers opting out of the system, since about a quarter of the urban labor force was not covered by social security contributions in 2010 (Hinz and others 2011). An important change would be to ensure that actual income is used in the calculation of insurance premiums, as opposed to the current system, which imputes a minimum level of earnings in the calculation of social security contributions. This change would avoid the sometimes extraordinarily high contribution rates for workers who earn below the minimum wage. Earnings of about one-third of the urban workforce in several major cities are well below the imputation threshold for social insurance contributions. Moreover, for about half of those workers the effective tax wedge in the formal sector (that is, the difference between gross and after-tax income) was in excess of 65 percent (Cai, Du, and Wang 2011). The authorities have committed to reducing the rate at a time and pace that ensure the sustainability of the social security system.

Corporate Income Tax

Current Situation and Issues

Before 2008, two parallel systems with different rates and incentives were applied to corporate income: one for domestic and another for foreign operations. The revision of the Enterprise Income Tax Law in 2008 eliminated this distinction and unified payments for all companies. It now operates a classical system, implying that a tax is applied on corporate profit, while distributed profit is taxed again on the individual and without relief. The law stipulates that taxes are due on the international business of the worldwide income of a resident company. Relief for double taxation is granted through tax credits for the foreign tax paid.

Between 2008 and 2012, corporate income tax accounted for 3.4 percent of GDP on average. This is somewhat higher than the 2.9 percent OECD average, but lower than in some other countries in the Asia and Pacific region (Figure 2.4). With a statutory rate of 25 percent, revenue productivity—measured as corporate income tax revenue in percent of GDP divided by the tax rate—is 0.14 percent; each percentage point of the corporate income tax rate generates about 0.14 percent of GDP. This is high compared to other countries. India and Japan raise almost the same revenue to output as China, but with much higher corporate income tax rates.

Figure 2.4.
Figure 2.4.

The Chinese statutory corporate income tax rate of 25 percent is competitive.5 It is slightly higher than in some of the smaller countries in the region, yet similar to Indonesia and below the OECD average. It is also somewhat lower than in other large economies in the region, such as India and Japan. Large countries generally have higher tax rates than small countries—as witnessed by the Group of Seven’s average corporate income tax rate of 31.3 percent in 2014 (Figure 2.4). That large countries set higher tax rates than small countries is a common feature of tax competition, as large countries have more to lose and less to gain from undercutting their neighbor’s tax rates when competing for tax bases (Keen and Konrad 2013).

The effective tax rate on new investment in China is also lower than in other countries. Investment distortions do not depend strictly on statutory corporate income tax rates, but on effective tax rates.6 The marginal effective tax rate is derived directly from the user cost, and it is a measure of the tax imposed on the marginal investment. A higher rate makes fewer investment projects worthwhile and will reduce investment. Tax rates can be calculated for different types of investment (buildings, equipment, intangibles) and alternative sources of finance (debt, retained earnings, new equity). The overall distortion of corporate income tax in China is modest compared to other countries, according to an international comparison of the weighted average of marginal effective tax rates (Chen and Mintz 2015). The rate in China is similar to the average in 61 non-OECD countries and slightly below the averages for the OECD and Group of Seven economies (see Figure 2.4).

Rationalizing tax incentives in 2008 was a key step in the modernization of China’s tax system. The elimination of tax incentives improved its neutrality and has contributed significantly to a more buoyant revenue base.7 While having played a useful role in attracting foreign direct investment, tax incentives create significant distortions between industries, regions, and businesses when in place for too long. The abolishment of tax incentives in the 2008 reform might serve as an example for other countries wishing to boost revenue mobilization. The new set of tax incentives is much more limited in scope and scale, and some incentives are justified on the grounds that they support activities that generate positive externalities for other industries, such as super deductions for expenses on research and development.

Small and medium-sized enterprises are commonly hard to tax. Beyond tax evasion (understating income by underreporting receipts or overstating deductions), entrepreneurs often respond to opportunities for tax arbitrage, as they can easily choose to convert heavily taxed income into lightly taxed income (Saez 2010). This makes it important to design an individual income tax system that provides for as much neutrality as possible in the taxation of entrepreneurial income, so that little room is left for entrepreneurs to engage in such arbitrage. Neutrality should be obtained in particular for decisions (1) to work as an entrepreneur or as an employee, which is especially relevant in sectors where tasks can be easily outsourced, as is the case in many service sectors; (2) to choose the legal form for a business, as either a separate legal entity (such as a limited liability corporation) or as a natural person (a sole proprietorship or partnership); (3) to distribute income earned by the entity or retain it in the company (so that it accumulates as an accrued capital gain); and (4) to distribute income from the legal entity to its owner(s) as labor remuneration, dividends, or a realized capital gain.

Currently, business income in China is taxed through a wide range of schedules. Sole proprietorships and partnerships are taxed under a special progressive rate schedule of the individual income tax system, with marginal rates of between 5 and 35 percent (Table 2.5). Closely held corporate businesses are taxed by the corporate income tax at a rate of 20 percent on annual income of less than RMB300,000 and at 25 percent for incomes above that. The distributions of closely held corporations can be taxed as wage income when paid as a director’s remuneration (at the schedule for ordinary salary income), or as dividend income taxed at a flat rate of 20 percent. The combined burden of corporate income tax plus dividend tax therefore lies somewhere between 36 and 40 percent.8 Several professional services are taxed at flat rates of 20, 30, or 40 percent on a perpayment basis—depending on the size of the payment—after a deduction of 20 percent for deemed expenses. Finally, micro businesses that are unable to keep books and records may be taxed on the profit they are deemed to have made.

Table 2.5.

Corporate Income Tax Schedule for Small and Medium-Sized Enterprises

article image
Source: International Bureau of Fiscal Documentation.

Proposed Reform

The current regime creates horizontal inequities, distortions, and opportunities for tax avoidance. Some distortions are discrete in nature, such as the choice of legal form or decisions regarding entrepreneurial entry. Others affect marginal behaviors, such as the retention or distribution of income and the way in which income is distributed (see Figure 2.4).

  • Discrete distortions. Average tax burdens vary for unincorporated business. For the corporate business that pays a salary to its owner-employee as long as it is taxed at a lower rate than dividends (the residual income is then paid as dividends), the tax rate will be different from those that pay above RMB50,000. Taxes on corporate businesses are always lighter than on noncorporate businesses in China, while professionals can be taxed more or less depending on their gross revenue and the size of payments.9

  • Marginal distortions. The marginal tax burden also differs on corporate capital gains, corporate dividends (assuming a tax rate on dividends of either 20 percent or 5 percent), and salary payments, such as a director’s remuneration (see Figure 2.4). Dividends are almost never tax preferred when taxed at 20 percent, but are charged at a lower marginal rate than is imposed on high salaries. Capital gains are least taxed at high incomes. Wages to low-income individuals accrue the lowest burden in light of the progressive structure of taxes on salaries.

Reforms to a dual-income-tax-inspired system would bring business income and various labor income under the same tax schedule. This would restore horizontal equity and eliminate distortions. For instance, applying the same schedule for taxable business income and salaries would eliminate differences in average tax burdens between employees, nonincorporated businesses, and corporate businesses that award a director’s remuneration to the owner. Distortions in the choice between entrepreneurship and salaried employment would therefore be eliminated. Taxing professionals (and authors) on their annual income, rather than on a per payment basis, would also align the average tax burden with that of other businesses.

The overall tax burden on profit distributions for a closely held business would be roughly aligned at a dividend tax rate of 15 percent and a 35 percent top rate on salaries. If the top rate on salaries were 35 percent, a dividend tax rate of 13.3 percent would ensure complete neutrality for companies paying 25 percent corporate income tax; a dividend tax rate of 18.75 percent would achieve neutrality for companies paying 20 percent corporate income tax.10 A dividend tax rate of 15 percent would therefore be “roughly” neutral for all entrepreneurs. Such treatment largely eliminates opportunities for tax arbitrage in the distribution of income by entrepreneurs between director’s remuneration and dividend payments.

Indirect Tax Reforms

This section elaborates on indirect tax reform and is guided by the conceptual framework shown in Box 2.2. The focus is mainly on options for reform of VAT and environmental taxation.

The Transition to VAT

Current Situation and Issues

The transition to VAT for goods and services was completed in 2016. China applied VAT only to the supply of goods and imports before the transition to VAT beginning in 2012.11 The standard VAT rate of 17 percent was complemented by a reduced rate of 13 percent for certain items. Small businesses became subject to a simplified VAT scheme, whereby they charge a reduced rate of 3 percent on sales but do not have the opportunity to claim refunds for input VAT.

Proposed Reform of VAT

The recent transition of business turnover tax into VAT is a very positive step toward raising the effectiveness and efficiency of China’s tax system, but continued reform is needed. The transformation is expected to eliminate major distortions, yet important challenges remain that warrant ongoing reform. These cover the need to do the following:

  • Unify the refund mechanism. Input tax on exports of all goods is currently not refunded in full. Refund rates vary between the full refund at the 17 percent VAT rate, to 5 percent or even to nothing for certain items. This violates an important principle of VAT as a destination-based tax, which requires full zero rating of exports with full crediting of input VAT. The practice in China imposes trade distortions that hurt the competitiveness of certain industries and distorts international trade.

  • Unify VAT rates. The current VAT has four different rates: 6, 11, 13, and 17 percent. This departs from the widely accepted standard of a single VAT rate for all items—both to avoid distortions and to ease tax administration. Indeed, reduced rates forgo revenue and create compliance problems such as with how to classify certain items.

Policy Guidance from Indirect Tax Theory

A comprehensive and uniform tax on consumption can be shown to be broadly equivalent to a tax on wages and profits. That is, lifetime consumption will match lifetime income for most individuals. So taxing them on either income or on consumption leads to similar revenue collection. However, beyond this observation, and based on very stylized models, a reform that replaces income-based taxation with consumption-based taxation can also effectively tax individuals’ capital endowment. This constitutes an efficient lump-sum tax on endowments if the reform is unanticipated. Further extensions, such as on international trade, can break down this equivalence, with potential short-term benefits for terms of trade from the move to destination-based consumption taxes such as VAT, from origin-based labor and production taxes (de Mooij and Keen 2013).

The seminal contribution by Atkinson and Stiglitz (1976) established that under general conditions, an optimal collection policy will rely on both direct and indirect taxes. In a setting with heterogeneous agents, consumption tax will be used even when the government has access to a nonlinear income tax. The reason is that when consumption patterns differ systematically across people with different ability to earn income, these consumption patterns can be used to identify individuals’ inherent ability. Moreover, when some goods are intrinsically related to either leisure or labor supply, varying tax rates can generate further gains in efficiency. However, the literature gives relatively little guidance about how to determine the specific composition.

Beyond these important theoretical considerations, the share of direct and indirect taxes is also a matter of administration. In particular, relying on both broad-based consumption taxes and income taxes reduces the risk of revenue leakage through evasion and noncompliance since it increases the opportunities to observe tax bases and collect taxes due. For example, taxes on wage income are effectively levied through withholding by employers. However, it is much harder to tax self-employed individuals through such schemes, so using indirect consumption taxes would be warranted to ensure some revenues are collected.

Individual commodity taxation should depend on the relative size of demand elasticities. This was shown in an early formulation of the problem of setting optimal tax policy by Ramsey (1927). Specifically, the ratio of the optimal ad valorem rates on two different commodities is determined by their cross-price elasticities with respect to the untaxed good (often assumed to be leisure for simplicity). Corlett and Hague (1953) also note this: a “second best” approach to indirect taxation is achieved by taxing goods and services that are most complementary to leisure. This also has the benefit of reducing distortions in labor supply incentives.

While recommendations from theory are straightforward, in practice the welfare gains of differential commodity taxation are not worth the higher administrative and compliance costs of more complex consumption tax systems (Crawford, Keen, and Smith 2010). Moreover, the distributional benefits of differential rates for certain necessities or luxury commodities are generally thought to be best achieved through income tax. The reason is that as long as the government uses an optimal nonlinear income tax, redistribution can be achieved without distorting the consumption choices of individuals, as would be the case with differential consumption tax rates.

Differential commodity tax rates may address externalities in consumption. A tax that accounts for the external effect incurred by other agents will cause the emitter to internalize the additional social cost of his action and choose the socially optimal level for his own private consumption (Pigou 1920).

Important examples are pollution, alcohol, and tobacco. Recent literature has also looked at how people’s lack of consistency, self-control, or access to information may be grounds for using indirect taxes to encourage welfare-improving consumption. The case for correcting these “internalities” has been made for tobacco as well as for the overconsumption of sugar and fat.

  • Ensure neutral treatment of property and financial services. Margin-based financial services are commonly VAT exempt, which creates inevitable distortions. To offset these, financial institutions may be taxed under a substitute Financial Activities Tax (IMF 2010). Newly constructed immovable property may be subject to full VAT, in lieu of tax on the added value that such property would generate over its lifetime.

  • Introduce a VAT registration threshold. Businesses with sales below the threshold are not obliged to register and will no longer have to charge 3 percent VAT on their sales. They can be offered to register voluntarily, for example, if they mainly export goods that put them in a net refund position. Voluntary registration may be denied for very small traders, so as to avoid abuse of VAT refunds for items bought for personal use.

A key challenge associated with the transition to VAT is the revenue impact for local governments. Revenue from the business tax was exclusively allocated to local governments, while only 25 percent of VAT revenue is shared. By splitting domestic VAT revenues equally between the central and local governments, which is expected to slightly increase the central government’s take, the reform has therefore reinforced the misalignment between local revenue and spending. This intensifies the need to reform local government financing more widely.

Environmental Taxation

Current Situation and Issues

China’s economic success has come at a high environmental cost. The Chinese Academy of Environmental Planning estimates that the cost of environmental damage amounted to 3½ percent of GDP in 2010. The World Bank has reported an estimate of 9 percent of GDP, with air and water pollution accounting for the largest shares due to pollution-related health damage, global climate change impacts, and resource depletion (World Bank and Development Research Center 2013). The World Health Organization estimates 1.7 million deaths were attributable to ambient air pollution in 2012 for the Asia and Pacific region, the highest per capita rate in the world.

The authorities have taken steps to reduce the economy’s reliance on natural resources and energy. The 13th Five-Year Plan built on environmental protection efforts adopted in 2011. The recent revision of the environmental protection law was also an important step toward setting a sound legal basis for further measures. However, the quantity-driven administrative approach has relied on targeted interventions, mandates, and regulations. Policy objectives are often pursued by mandating reductions in pollution and by setting targets without regard to current energy consumption and pollution. This can be especially burdensome for new plants and investments that already operate to the highest environmental standards.

China levies a variety of resource and vehicle taxes, and charges on air and water pollution (Figure 2.5). However, revenues are low compared to OECD countries and only a fraction of what is required to fully internalize the costs of climate change, local pollution, and congestion externalities. Energy prices are low, in particular for coal, which provides over 70 percent of the country’s energy. Deregulation of prices for water, oil, natural gas, and electricity is ongoing; however, it is important that taxes are also employed to ensure prices reflect the full cost of externalities from energy and resource use.

Figure 2.5.
Figure 2.5.
Figure 2.5.

Environmental Taxes and Costs: China and the World

A nationwide emission trading scheme will be introduced in 2017. It will build on seven regional pilot systems launched from 2011 to 2014 and will be the world’s largest emission trading system, even as at the start it is still expected to only cover about half of all the country’s emissions. The plan is to initially cover large entities in power generation, iron and steel, chemicals, cement, and paper. Small entities in these sectors, and those in road transportation, construction, and vehicle industries, representing about 50 percent of total emissions, will not be covered by the scheme, which means that low-cost emission reductions in these sectors will not be fully exploited. Beyond coverage issues, the specific design of the system will also be of great importance. It will be crucial that prices are sufficiently high and caps are stringent enough to support the country’s mitigation plans. Given the considerable revenues at stake—potentially about 2 percent of GDP (Parry, Shang, and Wingender 2016)—initial allowances should be auctioned and not allocated for free.

Proposed Reform

Unlike targeted regulatory policies, taxes can be very effective at exploiting opportunities for improving energy efficiency and abating pollution; for example, by giving energy producers incentives to switch to cleaner fuel or by increasing prices of polluting fuels to reduce demand. Fiscal instruments can also provide especially strong incentives for power plants to install and operate emission control technologies. Countries with higher taxes on fossil fuels tend to incur lower externality costs such as local air pollution, greenhouse gas emissions, traffic congestion and accidents, and road damage.

While the national emission trading scheme is a key step in the right direction, several tax policy instruments can complement the reform. They include the following:

  • Carbon tax. An upstream tax on fossil fuels based on carbon emission rates could be considered alongside the announced emission trading scheme (Parry, Shang, and Wingender 2016).12 A tax on pollution at point of entry in the economy (coal mines, coal-powered energy plants, petroleum refineries, and the like) would reduce the administrative burden. It could also build on administrative structures already in place. A carbon tax on upstream production would provide the necessary incentives for emission-reducing behavior throughout the economy since energy prices would increase uniformly. A carbon tax and an emission trading scheme are not incompatible; they can be used in tandem to maximize tax arbitrage. Many countries have had successful experiences with carbon taxes.13

  • Tax on local air pollutants. Current pollution levies on sulfur dioxide and nitrogen oxide emissions and small particulates could also be significantly increased. The revision to the Environmental Protection Law expected in 2016 will upgrade their legal status to full-fledged taxes. This will facilitate enforcement and ensure they provide the necessary incentives to reduce emissions as the current rates of taxes—about $200 per ton of sulfur dioxide for example—are only a fraction of the estimated health costs of local pollution (Parry and others 2014). The current direct subsidies and preferential pricing policies for the use of emission-control technologies could also be integrated into the tax system through a credit mechanism. The onus of reporting reductions in emissions through continuous monitoring would remain with the operator. This should be facilitated by the recent mandate for 15,000 plants to monitor and report their air-polluting emissions in real time.

    China would need a relatively high corrective tax rate to fully internalize all costs (Parry and others 2014). The estimate is about $15 per gigajoule on energy produced from coal, which includes a carbon tax of around $35 per ton of carbon dioxide and additional taxes of approximately $20,000 for each ton of emitted sulfur dioxide, nitrogen oxides, and fine particulates. While the global costs of carbon dioxide emissions on climate change are roughly equal across countries, local pollution costs from the elevated risk of premature mortality are significantly higher in China, mainly due to the location of coal-intensive industries, the current low level of abatement among emitters, and high population density in exposed areas. Current taxes on coal are very low across all countries and, like many others, China’s industrial and pricing policies in effect subsidize the use of coal. The proposed corrective tax would be comparable to that in the United Kingdom, and would be exceeded only by those in Israel and Poland in a selected group of comparable countries. Australia, Japan, and Korea in particular would require lower rates of around $6 per gigajoule due to the lower local pollution externalities caused by sulfur dioxide, nitrogen oxides, and particulates.

    An OECD modeling exercise in 2013 found that a modest carbon tax of $1.50 per ton of carbon dioxide emissions would have a very small impact on GDP, especially if the proceeds were used to offset consumption or payroll taxes. While output in carbon- and energy-intensive sectors such as chemicals, construction materials, and steel would contract, other less energy intensive sectors like financial services, food processing, and consumer services would benefit from the relative change in prices. Moreover, employment would also increase as labor is substituted for energy (Parry, Shang, and Wingender 2016). Other models of such policy experiments have reached similar conclusions (World Bank and Development Research Center of the State Council 2013).

    A phase-in period where rate increases are announced in advance would allow firms and consumers to adapt and undertake mitigation measures. Revenue recycling—using the proceeds of a carbon tax to reduce other distortionary taxes such as payroll taxes—is also crucial to minimize the impact of higher energy prices on aggregate demand. In the early years of reform, for example, corporate income tax credits could be used to offset the cost of abatement technologies. China should also continue to develop and commercialize cleaner sources of energy such as wind, solar, and hydropower. Ultimately, it is important to maintain the principle of “polluter pays” to ensure the environmental effectiveness of the reform and to achieve the broader goal of rebalancing growth toward less-energy-intensive sectors.

  • Energy and fuel taxes. Although fuel subsidies have fallen substantially in recent years, the government retains control over several energy markets. State-owned enterprises control most generation, transmission, and distribution of electricity as well as oil refineries. Private electricity generators must sell to the grid at fixed prices, while coal suppliers are required to sell some coal to power generators at below market price; and prices for natural gas are also determined by the government. Liberalization of the sector would promote competition and increase efficiency. Market prices would also better reflect the cost of production and scarcity. In the case of electricity, the government introduced a reform in 2012 explicitly to use household tariffs as a redistributive instrument. As a result, average household electricity prices in China are much lower than in other OECD countries. Despite the distributional benefits of the reform, a preferable option could be to use targeted cash transfers to households and fiscal transfers to provinces.

    China’s fuel excise tax remains low relative to world prices. The introduction of a fuel tax in 2009 was partly offset by the removal of other taxes, such that prices remain relatively low compared to world prices. Excise duties on gasoline and other petroleum products could be further increased to cover the externalities associated with greenhouse gas emissions, local pollution, and traffic congestion and accidents. To cover these costs, the effective tax rate on gasoline could increase from 40¢ per liter at present to roughly 55¢ per liter. This would be comparable to other countries’ taxes on gasoline and be significantly lower than in Japan (75¢ per liter) and Korea (85¢ per liter). The government could also consider introducing distance-and time-based taxes, which would be far more effective at decreasing congestion since rates could vary by period and place; for example, peak-period pricing for busy roads or cordon tolls around urban centers. Such measures have proved successful in Singapore, London, and Stockholm. Congestion charges would also be preferable to the license-plate rationing now used in some Chinese cities, which can be easily circumvented by owning more than one car or registering a car in a neighboring city.

  • Vehicle taxes. China’s current sales taxes on vehicles are based on engine capacity, with rates varying from 1 percent for vehicles with cylinder capacity below 1 liter, to 40 percent for vehicles with cylinder capacity larger than 4 liters. While engine capacity is a reasonable proxy for the environmental damage caused by cars, a better alternative would be to use a graduated system based on vehicles’ emissions. This could take the form of an ad valorem excise tax on vehicle value to meet revenue objectives and a revenue-neutral “feebate” on emission rates to promote environmental objectives. The feebate structure would entail an increasingly large fee or rebate on the purchase price depending on whether a vehicle is more or less fuel efficient than some standard—for example, the average emission rate across all vehicle sales for a given category. The tax rate could be set at a higher value to provide incentives to adopt more-fuel-efficient cars. These parameters could easily be modified to take into account improvements in fuel efficiency of a new car or to increase emission reductions. Vehicle taxes make cars costlier to own and will therefore reduce the overall demand for vehicles. They also provide a good tax handle that has attractive distributional properties since higher-income households typically spend more on vehicles. However, since these taxes do not provide an incentive to reduce the amount of driving, they do relatively less than fuel or distance/place taxes to reduce externalities from emissions, congestion, and accidents.

The environmental tax needs to balance revenue generation with distributional considerations, given the importance of energy prices for low-income households. Increases in energy and fuel prices should be offset for the most vulnerable through assistance programs and improved safety net policies (Parry, Shang, and Wingender 2016). Broad public support must also be secured to ensure sustainable reforms. This requires coordinated and far-reaching communication strategies, improvements in transparency about fiscal costs and the beneficiaries of current policies, and a comprehensive reform plan that includes phased increases in energy prices.

Recurrent Property Taxes

Current Situation

Recurrent property taxes are largely absent in China. Real estate tax revenues amount to 1.6 percent of GDP, which is comparable to the OECD average. However, taxes in China are mostly levied on transactions, whereas recurrent taxes take a much larger share of revenue in other countries (see Figure 2.1). Property taxes can raise up to 2 or 3 percent of GDP—although the average across countries is about 1 percent. This depends on exemptions, rate structure, and compliance. Rates usually vary between 0.1 percent and 1 percent of the market value of properties, depending on the type of property (such as commercial versus residential; or special rates applied for agricultural buildings).

Following pilot programs in Chongqing and Shanghai in 2011, a recurrent property tax could be extended to other municipalities and provinces.14 The tax base should also be extended to ensure that all urban owner-occupied housing—over 85 percent of households in China own their homes—is taxed, with appropriate exemptions for the lowest-valued houses. Over time, the recurrent property tax could replace the numerous transaction-based taxes and fees on land and property levied by local governments. Development of a nationwide registry is underway, but much remains to be done to improve local government capacity for data registration, appraisals, and valuations.

Proposed Reforms

Recurrent property taxes are helpful because they:

  • Do not discourage mutually beneficial transactions because of cash flow constraints. Macro regressions typically find that recurrent property taxes are among the least distortive to growth (Norregaard 2013). They are also less volatile than most other taxes, and are therefore particularly attractive as a local revenue source. Finally, these taxes are progressive, since high-income households usually tend to also have higher property wealth. Progressivity can be strengthened by the appropriate use of basic allowances.

  • Enhance the autonomy of local government finances. Local governments in China have almost no control over their own revenue sources. Recurrent property taxes could change that, since they are usually local taxes, often with some restrictions set by the central government, such as national guidelines for property valuation methods, for the types of tax exemptions allowed, and bands for rates. Tax bases are also usually defined nationally to prevent local jurisdictions from competing for marginal footloose investment. However, tax competition is usually limited due to the relatively low mobility of the base.15

  • Are consistent with the benefit tax principle. A well-designed recurrent property tax can be linked to the quality of local public goods delivery. Local amenities financed by property taxes, such as service delivery, public infrastructure, and environmental outcomes, will be capitalized in the prices of properties. By extension, efficient provision of local services and amenities would be reflected in increased property values, and ultimately in higher tax revenues. This could provide strong incentives for local governments to improve their delivery of services.

  • Contribute—when well designed—to urbanization and help contain real estate risks. Barnett and Zhang (2014) estimate that fiscally induced pressure drove about one-third of property transactions in large cities in 2012–13, although others estimate the role of taxes to be less significant. Their introduction, however, needs to be timed well and phased in to avoid triggering a slowdown in the real estate market.

Current challenges to implementing a broad-based property tax include the absence of a nationwide registry of real estate, although the Ministry of Land and Resources has recently taken steps to implement registration. As long as proper valuation records remain unavailable, property taxes can be based on the area and location of the property, possibly with corrections for certain characteristics, to approximate value. Many emerging market and developing economies use such area and proxy-based systems. Another difficulty in China is that many properties are government owned and land ownership applies only for fixed periods (typically 70 years). Special attention should also be paid to property taxes on businesses and agriculture, perhaps by setting different rates or exemptions. Transitional arrangements should be carefully considered since the introduction of a recurrent tax in addition to transaction taxes may lead to the perception of “double taxation.” The ongoing transition from the business tax to VAT also has important implications for local government revenue.

The central government may decide to set national guidelines for property valuation methods, for the types of tax exemptions allowed, and for bands for rates. Such bands would in principle set a strictly positive bottom threshold. Finally, prices of immovable property vary widely across provinces, so the tax-free basic allowance could be set differently across provinces and cities.


Tax reforms are an important part of China’s efforts to modernize the fiscal framework. Reforms can improve the efficiency of the tax system, and at the same time contribute to more inclusive and environment-friendly growth. The country has considerable scope to increase the redistributive role of the tax system. Key steps to achieve this would include shifting away from the heavy reliance on indirect taxes toward direct taxes, transitioning indirect tax toward a well-designed VAT (that is, building on the extension of the VAT to services by replacing the distortive and cascading business tax), strengthening the progressivity of individual income tax, and reducing social security contribution rates, especially for low-income workers.

For direct taxes, a dual income tax for individual income could serve as a model for China. It would be a significant step toward a more integrated income tax by reducing the current 11 categories of taxable income to two broad categories: labor or capital income. Such a dual system would present many advantages from policy and tax administration perspectives. The labor income tax schedule should include progressive tax rates so that it can support redistribution purposes. Taxing all sources of labor income under the same schedule is also regarded as fairer, since individuals with same income are treated the same irrespective of where that income is earned. Sources of capital income would all be taxed at a proportional rate, which should not be too different from the top marginal tax rates applicable to labor income, taking into consideration any burden imposed by corporate income tax. This would strike a good balance between efficiency considerations and equity objectives.

For indirect taxes, the transition to VAT has raised efficiency considerably. Remaining reforms should focus on unifying the multiple VAT rates and improving the refund mechanism to avoid distortions and ease tax administration. At the same time, measures to ensure neutral treatment of property and financial services are important. Introducing a VAT registration threshold could further enhance VAT efficiencies by reducing administrative and compliance costs.

Environmental taxes can be effective for improving energy efficiency and abating pollution. Fiscal instruments can also provide especially strong incentives for power plants to install and operate technology to control emissions.

Ongoing work on the registration of properties and preparatory legislation would contribute to the launch of recurrent property taxes. The introduction of property taxes would be consistent with the benefit principle and facilitate reforms to strengthen local government revenue sources and support urbanization.

Policy sequencing and coordination among agencies are important for tax reforms to be successful. Many of the reforms are intertwined and therefore require careful sequencing to mitigate their impact on fiscal revenue and growth, particularly in light of the misalignment of revenues and spending by central and local governments. For instance, the transition from business tax to VAT would likely reduce tax revenues for local governments, even if the reform is revenue neutral in aggregate. Careful sequencing, however, should not cause delay. Steadfast implementation of reforms is necessary to reach an inclusive and sustainable growth path. Other reforms, such as social security and property tax, would involve several ministries. Policy coordination and the sharing of information among agencies would be key to successful implementation.


  • Atkinson, A. B., and J. E. Stiglitz. 1976. “The Design of Tax Structure: Direct versus Indirect Taxation.Journal of Public Economics 6 (1976): 5575.

    • Search Google Scholar
    • Export Citation
  • Barnett, S., and Y. Zhang. 2014. “Fiscal Vulnerabilities and Risks from Local Government Finance in China.Working Paper No. 14/4, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Brys, B., S. Matthews, R. Herd, and X. Wang. 2013. “Tax Policy and Tax Reform in the People’s Republic of China.OECD Taxation Working Paper No. 18, Organisation for Economic Co-operation and Development Publishing, Paris.

    • Search Google Scholar
    • Export Citation
  • Cai, F., Y. Du, and M. Wang. 2011. “Overview of China’s Labor Market.Working Paper, Institute of Population and Labor Economics, China Academy of Social Sciences, Beijing.

    • Search Google Scholar
    • Export Citation
  • Cevik, S., and C. Correa-Caro. 2015. “Growing (Un)equal: Fiscal Policy and Income Inequality in China and BRIC+.Working Paper No. 15/68, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Chen, D., and J. Mintz. 2015. “The 2014 Global Tax Competitiveness Report: A Proposed Business Tax Reform Agenda.University of Calgary, the School of Public Policy, University of Calgary Publication Series 8 (4).

    • Search Google Scholar
    • Export Citation
  • Clements, B., R. de Mooij, S. Gupta, and M. Keen, eds. 2015. Inequality and Fiscal Policy. Washington: International Monetary Fund.

  • Corlett, W. J., and D. C. Hague. 1953. “Complementarity and the Excess Burden of Taxation.Review of Economic Studies 21 (1): 2130.

    • Search Google Scholar
    • Export Citation
  • Crawford, I., M. Keen, and S. Smith. 2010. “Value-Added Tax and Excises.” In Dimensions of Tax Design: The Mirrlees Review, 275362. Oxford: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • de Mooij, R., and M. Keen. 2013. “Fiscal Devaluation and Fiscal Consolidation: The VAT in Troubled Times.” In Fiscal Policy after the Financial Crisis, edited by Alberto Alesina and Francesco Giavazzi. Chicago: University of Chicago Press.

    • Search Google Scholar
    • Export Citation
  • Diamond, P. A., and J. W. Banks. 2010. “The Base for Direct Taxation.” In Dimensions of Tax Design: The Mirrlees Review. Oxford: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Hinz, R., M. Pallares-Miralles, C. Romero, and E. Whitehouse. 2011. “International Patterns of Pension Provision II: A Worldwide Overview of Facts and Factures of the 2000s.Social Protection Discussion Paper, World Bank, Washington.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (IMF). 2010. “A Fair and Substantial Contribution by the Financial Sector—Final Report for the G-20.International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Keen, M., and K. Konrad. 2013. “The Theory of International Tax Competition and Coordination.” In Handbook of Public Economics Vol. 5, edited by A. Auerbach, R. Chetty, M. Feldstein, and E. Saez. Amsterdam: North Holland.

    • Search Google Scholar
    • Export Citation
  • Lam, W. R., and P. Wingender. 2015. “China: How Can Revenue Reforms Contribute to Inclusive and Sustainable Growth?Working Paper No. 15/66, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Mirrlees, J. A. 1971. “An Exploration in the Theory of Optimum Income Taxation.Review of Economic Studies 38 (2): 175208.

  • Norregaard, J. 2013. “Taxing Immovable Property: Revenue Potential and Implementation Challenges.Working Paper No. 13/129, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Parry, I., D. Heine, E. Lis, and S. Li. 2014. Getting Energy Prices Right: From Principles to Practice. Washington: International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Parry, I., B. Shang, and P. Wingender. 2016. “Climate Mitigation Policy in China: Assessing the Options.Working Paper No. 16/148, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Pigou, A. C. 1920. The Economics of Welfare. London: Macmillan.

  • Piketty, T., and E. Saez. 2013. “Optimal Labor Income Taxation.” In Handbook of Public Economics Vol. 5, edited by A. Auerbach, R. Chetty, M. Feldstein, and E. Saez. Amsterdam: North Holland.

    • Search Google Scholar
    • Export Citation
  • Ramsey, F. P. 1927. “A Contribution to the Theory of Taxation.Economic Journal 37 (145): 4761.

  • Saez, E. 2010. “Do Taxpayers Bunch at Kink Points?American Economic Journal: Economic Policy 2 (3): 180212.

  • Wang, X., and R. Herd. 2013. “The System of Revenue Sharing and Fiscal Transfers in China.OECD Economics Department Working Paper No. 1030, Organisation for Economic Co-operation and Development, Paris.

    • Search Google Scholar
    • Export Citation
  • World Bank. 2014. State and Trends of Carbon Pricing 2014. Washington: World Bank.

  • World Bank, and Development Research Center of the State Council. 2013. China 2030: Building a Modern, Harmonious, and Creative Society. Washington: World Bank. http://documents.worldbank.org/curated/en/2013/03/17494829/china-2030-building-modern-harmonious-creative-society.

    • Search Google Scholar
    • Export Citation
  • Zodrow, G., and P. Mieszkowski. 1986. “Pigou, Tiebout, Property Taxation, and the Underprovision of Local Public Goods.Journal of Urban Economics 19 (3): 35670.

    • Search Google Scholar
    • Export Citation

We are grateful for comments from the Ministry of Finance and State Administration of Taxation. CUI Lu and Sung Jung provided excellent research assistance.


Income disparities and limited redistribution are mirrored by the average income by group, with individuals in the highest-income decile earning more than 16–18 times those in the lowest decile, according to the World Bank and Development Research Center of the State Council (2013).


Local governments can be broadly classified into provinces, prefectures, counties, and townships. Each administration is required to maintain a balanced budget and receives transfers from an upper level, while providing transfers to its lower levels of government. County governments typically have the highest spending obligations and are therefore the most dependent on transfers.


Pure global income tax systems do not exist, even in countries that have tried to approximate them, such as the United Kingdom and the United States.


A statutory rate of 15 percent would imply that dividends are effectively taxed at a combined rate of 36.25 percent [0.25 + (1 – 0.25) × 0.15], given the regular enterprise income tax rate of 25 percent on corporate profits.


A reduced rate of 20 percent applies to companies with annual income less than RMB300,000.


Investment theory suggests that investment projects continue to be undertaken until the project that just breaks even. The price at this point is the user cost of capital, which depends not only on the corporate income tax rate but also on elements of the tax base, such as depreciation allowances, inventory valuation, interest deductibility, and investment tax credits.


Most tax incentives have been successfully phased out. Between the mid-1980s and mid-2000s, China introduced a range of industrial policy instruments, including tax incentives for special economic zones, reduced tax rates for foreign direct investment, and tax holidays for strategic industries. These schemes are regarded as having succeeded in achieving their objectives, including to attract foreign direct investment. Most tax incentives were phased out in the new Enterprise Income Tax Law, but some were retained, including a reduced rate for high-technology and research and development investment, new enterprises in the western region of the country, and selected services in the Shanghai free trade zones.


This reflects a 20 percent corporate income tax plus (1 – 0.2) × 20 percent dividend tax, and a 25 percent corporate income tax plus (1 – 0.25) × 20 percent dividend tax.


Further differences in effective tax burdens between professionals can arise due to variation in actual costs incurred. For instance, if actual business expenses are 50 percent of revenue, the 16 percent tax on gross receipts is equivalent to an income tax rate of 32 percent.


The formula to maintain neutrality between salary payments and dividends is TD = (TP – TC)/(1 – TC), where TD is the dividend tax, TP is the top personal income tax rate on salaries, and TC is the corporate income tax.


The government started to extend the VAT to services as a replacement for the business tax in 2012. A pilot program was launched in Shanghai for transportation services (except railway transportation) and modern services industries. The pilot was extended to eight municipalities and provinces in August 2012. Two new VAT rates of 6 and 11 percent were added to the existing rates of 13 and 17 percent on goods. The Ministry of Finance and State Administration of Taxation extended VAT on these services to the entire country in August 2013. Beginning in January 2014, VAT for railway transportation and postal services was introduced nationwide. In June 2014, it was broadened to cover telecommunications services. The transition to the remaining services was completed in May 2016.


The tax set on each fuel would equal the carbon dioxide coefficient (that is, the carbon dioxide produced by combusting a unit of the fuel) times the externality cost of climate change associated with the additional carbon dioxide emitted.


Successful examples of carbon taxes across the world include the Canadian province of British Columbia, which introduced such a tax in 2008. The carbon tax covers approximately 70 percent of all greenhouse gas emissions of the province, and the rate was increased gradually over time and is now set at $30 per ton of carbon dioxide emitted. An interesting feature is that the tax is meant to be revenue neutral as the proceeds are remitted through tax cuts and credits for low-income households and corporate income tax credits. It is estimated that the province reduced carbon emissions by 10 percent over the first three years of the tax. Along with Finland, Norway, and Sweden, Denmark introduced a carbon tax in the early 1990s. To mitigate the tax burden, other energy taxes were decreased. The tax currently covers about 45 percent of all carbon emissions at a current rate of $31 per ton of carbon dioxide. Industries subject to the European Union emissions trading scheme are generally exempt from the carbon tax. Gradual implementation and differentiation of effective rates between energy uses minimized the impact of tax competitiveness. Primary energy intensity has declined by 26 percent since the introduction of the carbon tax, and carbon dioxide emissions fell by 25 percent during the first decade (World Bank 2014).


The tax is currently applied to the ownership of a second property. In Shanghai, it applies to newly acquired second properties exceeding 180 square meters for a family of three but is subject to multiple exemptions. For example, the tax is not levied if married children of the owner use a house as their primary residence.


The so-called new or capital view sees property tax as a tax on capital (including mobile capital) within a local jurisdiction, with the distortions that may imply. See the seminal discussion in Zod-row and Mieszkowski (1986).

Investing in Soft Infrastructure