This book is about foundations: What are the most important structural features of China’s contemporary tax system, and what are its fundamental determinants? In 1994, Chinese political leaders engineered a major tax reform that continues to define the key elements of the country’s public finance system. This book focuses almost entirely on taxation in China since 1994,Footnote 1 an approach that invites some questions. The 1994 tax reform may itself be an important part of the explanation of China’s success in collecting revenue in the past quarter century. If so, would an account of Chinese taxation not need to situate the post-1994 tax system within a longer time horizon? In other words, what are the key decisions of the 1994 tax reform, and how do they correspond to the structural features of the system that emerged?
Past commentators have habitually given extensive attention to the 1994 tax reform. In fact, among educated people – who are not tax specialists – whom I have talked to in China, the single institutional feature of Chinese taxation that they are most likely to identify is the revenue-sharing system (fenshuizhi) established in 1994. Under this system, the national (central) government agreed with provincial governments on sharing ratios for revenues from different types of taxes; these ratios varied across taxes but were uniform across provinces. Provincial governments then proceeded to agree with their own sub-provincial units on ways of sharing revenue not claimed by the central government. The sharing percentages for different taxes have changed over time.
One problem with privileging this aspect of the 1994 tax reform, however, is that revenue-sharing arrangements are quite common around the world, especially among federalist countries (such as Canada, Australia, and Germany). While intergovernmental revenue transfers are important to the politics of taxation in many countries, there is no reason to believe that they are more important in determining China’s tax system than they are in defining other tax systems. Moreover, while there are long-standing bodies of scholarship that study intergovernmental transfers, subnational tax competition, and other related topics, they are relatively specialized subjects that assume a background of state capacity for revenue extraction. If revenue sharing were the most remarkable feature of the Chinese tax system, the system itself would not be that remarkable.
There is also the problem, of course, that revenue must be raised before it can be shared. One aspect of the 1994 tax reform that obviously requires comment before we turn to revenue sharing is the taxes China decided to impose. Given that China had only just decided to abandon its planned economy in the early 1990s, one might expect that the taxes it introduced in 1994 were new. This was indeed the case. The 1994 tax reform launched many taxes, all of which were relatively unfamiliar to the country. These included, among others, a unified corporate income tax (CIT) that would apply to all domestically owned firms (a parallel CIT regime for foreign-owned firms was retained); a personal income tax (PIT) that would theoretically apply to the entire populace; a value-added tax (VAT) applicable to the production and sale of goods; and a broad-based turnover tax applicable to all services (Business Tax, or yingyeshui).Footnote 2 Specialists refer to this portfolio of tax policy instruments and their relative importance as the “tax structure” of a country; China’s current tax structure is a product of this 1994 tax reform.
However, tax structure is a topic that tends to be relegated to specialists. Few people outside public economics or the tax and business professions seem to think that the relative importance among a country’s PIT, payroll taxes, CIT, or VAT is of fundamental institutional significance. Also, what a specialist can say is that China’s chosen tax structure is, at least on its surface, somewhat conventional from a comparative perspective. Most importantly, the PIT plays a relatively small role in raising revenue, while the CIT, VAT, and Business Tax lift most of the weight in filling the government coffer. There are generally accepted explanations for the prevalence of similar tax structures among developing countries, to which we will turn in Chapter 6. But because of such similarities, it follows that describing this tax structure when explaining the 1994 tax reform would still not identify what is unique about Chinese taxation.
Finally, the 1994 tax reform set up two separate branches of tax administration: the State Tax Bureau (guojia shuiwuju, which I abbreviate as guoshui or GS) and Local Tax Bureau (difang shuiwuju, dishui, or DS) systems.Footnote 3 The former collected several taxes that had a greater or entire share of their revenue claimed by the central government (most importantly, the VAT), while the latter collected other taxes, including especially the PIT, Business Tax, and, in some provinces, social insurance (SI) contributions.Footnote 4 This bifurcation of tax administration, creating what has been known as the guoshui versus dishui dichotomy, was for fourteen years a key feature of Chinese tax administration. However, the bifurcation was abolished in 2018, and its legacy is unclear. While the guoshui versus dishui tax administration split has fewer counterparts in other countries, we will see that it too yields few general implications.
This chapter aims to answer the following question: Is there some aspect of the 1994 tax reform that helps define what is relatively unique about the Chinese fiscal state? I offer an affirmative answer, but the answer’s details have to do with a core vision of the reform that unraveled, was gradually abandoned by the early 2000s, and became virtually unspoken of. In other words, the most distinctive part of the reform is a part that failed, and consequently, was forgotten.
An Earlier Foundation
While the politics of the 1994 tax reform was extraordinarily complex,Footnote 5 consider one basic question that Chinese national policy makers – from Premier Zhu Rongji to his advisors in the Ministry of Finance (MOF) – must have faced. How should China go about, practically, collecting the multitude of new taxes? There are two equally stylized and contrasting answers. The first is that one needs bodies. That is, the government needed to ensure that there were tax collectors around the country ready to enforce the new taxes. This answer, focusing on personnel capacity, seems intuitive and compelling. Many scholars and commentators have followed this idea to explain subsequent developments in Chinese tax collection. For example, they have examined the guoshui and dishui tax bureaus – with some scholars positing differences in incentives between tax collectors in the two administrative branches – and used such differences to explain tax administration outcomes.
While I will consider these arguments later, it is important to note that before 1994, the “bodies” were already largely there. Figure 1.1 plots the long-term evolution of the size of the Chinese tax administration workforce, based on the China Taxation Yearbook (for years after 1993) and information gathered from various archival sources (for earlier years). What we see is a striking and steady rise in tax collection personnel beginning from 1979, the start of China’s economic reform. This workforce was not bifurcated into two systems before 1994, and in Figure 1.1 it is represented by the black line until 1993. By 1988, the overall personnel size of Chinese tax administration had already risen to around 463,000 individuals, a 160 percent increase from the 1979 level. By 1993, formal employment in the tax administration apparatus reached 581,000, or 325 percent of the 1979 level. In 1994, this workforce was split into the GS and DS branches, with the GS system keeping most existing personnel. The DS system, taking over a smaller portion of the preexisting army of tax administrators, began to hire more staff. The GS system experienced a significant personnel reduction in 2001, but the size of Chinese tax administration remained steady in the following years.Footnote 6 In 2013, the total number of tax administrators in China was 756,592, only 30 percent higher than what the country had in 1993. The basic message of Figure 1.1 is that whatever the Chinese government did to organize personnel for enforcing a new tax system, it did so upon a foundation laid down before 1993. The primary challenge facing policy makers in 1993 was not the shortage of personnel. Instead, it was how to make an existing army of tax collectors collect new taxes.
Figure 1.1 The evolution of China’s tax administration workforce
What were the 581,000 Chinese tax administrators in 1993 doing, if they were not collecting the taxes later introduced in 1994? It is possible to answer this question, without going into much detail into the evolution of the Chinese tax system up until 1993,Footnote 7 in three steps.
First, between the 1970s and 1993, tax policy making in China was considerably decentralized. Although the national government enacted a series of taxes after economic reform began in 1978, the obligation to enforce these taxes fell on subnational governments and was often treated as nonbinding. This was because regardless of the specific taxes in use, most provinces were only accountable to the central government for transferring fixed amounts negotiated in advance for each year, under what was known as the “fiscal contracting” system.Footnote 8 Because provincial and sub-provincial governments carried out tax collection, and because the central government’s main lever of control was negotiating targets for revenue transfers to itself, the rules by which taxes were collected were hard to dictate.
This was exacerbated by the fact that the rules requiring enforcement left much to discretion. A majority of the new taxes introduced by the central government in the 1980s were taxes on the profits of SOEs and COEs that still dominated China’s urban economy at the time. These profit taxes replaced the surrender of profits by state- or collectively owned firms to their government shareholders under the planned economy. Notably, firms under the planned economy had not kept much of their profits. Consequently, the taxes that initially replaced the profit distribution regimes were characterized by high rates, as well as pervasive firm-specific negotiations with the government regarding the applicable rate. Regional tax competition and grants of local tax preferences further fueled a tendency for tax liabilities to be determined on an individual firm basis. To a substantial extent, there was no set of “national laws” enforced by tax collectors around the country .Footnote 9 Instead, low-level tax offices often directly participated in negotiations between local governments and local SOEs and COEs about how much tax each had to pay. The line between tax administration and tax policy making was tenuous. This was one of the core practices that the central government tried to curtail by launching the 1994 tax reform.
Second, apart from the risk – a source of notable concern to the central government – that individual tax collection offices would decide what policies applied to, or could be waived for, particular taxpayers, the role of the Chinese tax collector in the 1980s generally involved intensive contact with taxpayers. Large taxpayers (many of which were SOEs and COEs) were often each assigned a dedicated tax administrator (shuishou zhuanguan yuan), an agent who determined what and how much taxes a firm needed to pay, and literally saw through payments from the firm to the government. Very early on, Chinese tax administrators recognized a stark contrast between this emerging mode of tax collection and the style of tax administration that foreign experts were beginning to describe in relation to developed countries: it was often observed that in China, there was no separation between “administration,” “inspection,” and “collection” in taxation. In the 1980s, this all-in-one role of the tax administrator had come to be known among Chinese tax officials as “nanny-style” (baomushi) tax administration: taxpayers were like children, with their tax compliance taken care of by tax administrators.Footnote 10
Chapters 3 and 4 explore how a more technologically advanced Chinese tax administration still struggles to exorcise the spirit of “nanny-style” administration. “Returning responsibility to taxpayers,” for example, has been the core principle and ringing slogan in tax administration reforms pursued by some provinces in the past decade. However, in 1980s China, “nanny-style” tax administration was a very practical arrangement. For SOEs and COEs in particular, paying tax was just a continuation of the prior practice of transferring profits to economic planners. Both tax and accounting rules were still underdeveloped, so the intervention of government agents was often necessary for determining tax liability. Budgetary units within the government were also decentralized, fragmented, and continuously changing. Having tax collectors help each business navigate such a system and shepherd tax payments through may have seemed a natural additional task.
Third, as the Chinese urban economy began to liberalize in the late 1980s, many SOEs and COEs began to fail: tax revenue from such firms plummeted.Footnote 11 Increasingly, the government needed to raise revenue from a growing population of private firms that were much more resistant to the demands of tax administrator “nannies.” Moreover, the sheer number of new firms implied that tax administrators could not remain “dedicated” to a single or a small number of taxpayers. It did not take long for the government to figure out that the “dedicated tax administrator” system could not continue. Consequently, discussions about reforming tax administration emerged just a few years after such administration was implemented. In the early 1990s, before the 1994 tax reform, two practical ideas were widely discussed.Footnote 12 One was that tax administrators should not handle payments; taxpayers should remit taxes directly at banks to government accounts. The other was that tax administrators should increasingly focus on “examining” businesses’ tax affairs. The underlying assumption was that, by then, one could speak of correct ways of determining tax liability, which taxpayers could follow themselves. Tax administrators would simply verify and monitor taxpayers’ self-assessment.
In summary, China already employed close to 600,000 tax administrators in 1993. Most of these individuals were hired in the 1980s; they enforced a transitory set of taxes that would soon become obsolete and dealt with a tax base (i.e. the profit and turnovers of SOEs and COEs) that was rapidly shrinking. While the 1994 tax reform would soon centralize both tax policy making and government budgetary affairs on the political front, there is an entirely separate question that needed to be answered: How would the work of frontline tax administrators change? This leads us back to the question, described earlier, that likely confronted Chinese tax policy makers in 1993: How should China go about collecting so many taxes that it had not collected before?
There is a second stylized answer to this question, which does not hinge on the existence of tax administrators (they were already there), and not even, at least not primarily, on tax administrators doing things in particular ways. In fact, this second answer ought to be easy to grasp, but it has eluded many commentators on modern tax administration.
Specifically, among the variety of taxes that China considered adopting in the early 1990s – which were also taxes widely observed in other countries in the late twentieth and early twenty-first centuries – some depended on the actions of tax administrators more than others. Taxes on real property and custom duties on imports, for example, required tax administrators’ intervention in the completion of compliance obligations: a tax assessor often needs to assess property value before a property owner’s tax liability can be determined and paid; the confirmation of duties paid is needed for customs to be cleared. Other taxes, however, are self-assessed: tax administrators intervene only in some cases, such as to audit a taxpayer, to correct mistakes, or to provide a refund. Generally, the taxpayers themselves determine the amount owed, make payments, and complete compliance tasks without interaction with tax administrators. The CIT, VAT, Business Tax, and the withholding of PIT and payroll taxes by employers – examples of major sources of tax revenue in post-1994 China and most developed countries – are characterized by this reliance on self-assessment.
To make this system work, then, a radical shift needed to happen: taxpayers had to start filing tax returns and follow the law themselves.
Reforming Administration to Enforce New Taxes
To see how vital taxpayer compliance was to the way Chinese political leaders envisioned the 1994 tax reform, one only has to read one of the two key official documents that laid out the reform’s agenda: the notice issued by the State Council on December 25, 1993, to all Chinese provincial governments and all central government ministries regarding “the Approval and Transmission of the Implementation Plan for the Reform of Industrial and Commercial Taxation Proposed by the State Administration of Taxation.”Footnote 13 About two-thirds of the document was devoted to the various taxes that the reform would introduce. The next lengthiest portion of the document, however, proposed the principles for a tax administration reform.
Practical, but New, Principles
The State Council’s careful formulation of these principles in 1993 is worth quoting at length:
1. Widely establish a tax declaration system. Making tax declarations is the first key step for taxpayers in fulfilling their tax obligations. Establishing a tax declaration system would facilitate the formation of a self-restraining mechanism among taxpayers and enhance the tax-paying awareness among citizens. It is a foundational project that enables tax authorities to implement effective tax collection and management. After the establishment of the tax declaration system, any delays or false declarations shall be regarded as tax evasion and strictly punished according to the law.
2. Actively promote a tax agent/representative system. In accordance with internationally accepted practices, a tax agent system shall be implemented, where accounting and law firms and other private intermediaries represent taxpayers in handling tax affairs. These agents will gradually become an indispensable link within the tax collection and management system.
3. Accelerate the process of computerizing tax collection and management. International experience has proven that the use of computers and other advanced technologies in tax collection is the only way to establish a rigorous and effective tax monitoring network, while also reducing the cost of tax collection. As the country has a large number of scattered taxpayers and a weak foundation in computerized management, the computerization process may begin in cities and in the collection and management of key tax types. A nationwide comprehensive computer network for tax collection and management would be gradually established.
4. Institute a strict tax inspection system. Following the wide establishment of a tax declaration system and a tax agent system, the main focus of the tax authorities would shift to carrying out routine as well as key inspections. Tax collection would be founded on the three-part configuration of declaration, assistance by tax agents, and inspection. These approaches shall be supplemented by heavy penalties for tax evasion.
5. Separate central and local tax agencies will be established to meet the needs of the revenue sharing system.
6. Establish the basic tax norms that suit the needs of the socialist market economy. At this stage, it is particularly necessary to emphasize the following. Taxpayers must pay tax in accordance with the law. Tax authorities must collect taxes in accordance with the law. Taxes must be calculated based on the rates set out in the tax law. No one shall apply a “tax farming” (baoshui) approach or arbitrarily adjust the tax rates. All ad valorem taxes must be calculated based on actual price, the practice in some industries of not changing tax after price increases must be eliminated. Except for tax reduction and exemption prescribed by the tax law, governments at all levels and various departments shall not approve additional tax reduction or exemption.
7. Lawmaking power for both central taxes and local taxes that are to be implemented nationwide will be exclusively held by the central government.
8. Strengthen the rule of law in tax collection, expedite the legislation process for tax laws and regulations; gradually establish a tax system where legislation, judicial review, and law enforcement are mutually independent and mutually constraining.
“With the implementation of the above measures, the country’s tax administration will transform into a new configuration which puts equal emphasis on legal rules, collection, inspection, and appeals and litigation, with the four aspects coordinated with and binding each other.” (Emphasis added)
Three comments must be made about these striking pronouncements. First, they describe a style of tax administration characterized by several interlinked elements, all of which were missing from the Chinese tax system when the pronouncements were made. That is, on the eve of the reform, most Chinese taxpayers did not declare their liabilities on tax returns filed with tax authorities; tax professions did not exist to assist taxpayers in such tasks; tax authorities rarely engaged in audits (since there were no declarations or returns to audit); the norm of complying with the law was not established among either tax administrators or taxpayers; and there existed only rudimentary tax rules and regulations. Additionally, while the Chinese legal system allowed citizens to file lawsuits against the government starting in the late 1980s, judicial involvement in tax collection had mainly been in the criminal procedure sphere. In short, these proposed mechanisms of tax administration were as new to China as the taxes introduced by the 1994 tax reform.
Second, as revealed by the references to international experience in the second and third principles, the State Council’s and SAT’s vision of tax administration reform borrowed from foreign practice.Footnote 14 The Chinese government received substantial technical assistance from the World Bank, International Monetary Fund, and other international organizations and foreign governments in designing the 1994 tax reform. Such foreign advice appeared to have convinced the Chinese government of the necessity of implementing a rather alien vision,Footnote 15 within which tax administration was, non-trivially, about promoting norms associated with legal systems. Specifically, it was not just that how tax liabilities were determined needed to be written into the law: it needed to be written into the law because that was how taxpayers would know what rules to follow, which in turn was important because taxpayers taking their own initiative to apply the law was how the new taxes would be collected. The tax professions were necessary to help taxpayers apply the law. Similarly, the role of tax administrators was reconceptualized: tax administrators must follow the law instead of exercising discretion, and their primary task now was to check whether taxpayers correctly applied the law to themselves, rather than applying the law for taxpayers.
This novel language of law may nevertheless have sounded practical to Chinese government officials. Compare the reference to law with the idea of computerization: China was still a poor country in 1993; the widespread use of computers could only be described in aspirational terms. Law, by contrast, was not some expensive technology, but a social practice that a government might adopt should it fit its country’s needs. Although the State Council did allude to what one might call the superstructure of law (i.e. legislatures and courts), its focus was on such grassroots appliers of law as tax administrators and taxpayers. Moreover, since the 1980s, the Chinese government had already increasingly emphasized rule-based tax collection, and had especially wanted to use law to limit local government discretion. Consequently, policy makers already possessed some familiarity with the invocations of law.Footnote 16
Third, and most importantly, the State Council’s 1993 pronouncements conveyed a vision that has not been realized: China currently does not have the system of tax administration contemplated in 1993; it has instead taken a different route, one that will preoccupy us for much of this book. Readers previously exposed to discussions about the 1994 tax reform will likely be surprised to learn that tax administration was a core of the reform agenda,Footnote 17 or that the implementation of legal norms was considered central to the transformation of tax administration. The fact that the administrative reform described in 1993 did not fundamentally transform China’s tax collection and compliance explains why. This history in itself should also puncture any presumption that the principles the State Council endorsed in 1993 are anodyne recitations of what would more or less happen automatically. Instead, the principles articulate choices about fundamental institutions. The abandonment of these choices, as the rest of this book will show, has had radical consequences.
Of course, the State Council’s policy pronouncements are sometimes mere rhetoric, echoed by subordinate government units for some time, but not necessarily pursued in practice.Footnote 18 However, the State Council’s 1993 pronouncement about reforming the nation’s tax administration was no empty talk. If its fundamental logic has not been reflected in China’s current tax administration practice, it was not because no action was taken. In January 1997, the Office of the State Council released a plan for reforming tax administration proposed by the SAT.Footnote 19 The 1997 plan called for “deepening tax administration reform.” Reform would be gradually rolled out in the country, first in urban centers and then in rural areas. The goal was to “strive to largely complete” the plan by 2010. This formulation and the fourteen-year time horizon indicate that what the SAT had in mind was ambitious. It could not be quickly executed, as it was a matter of fundamental institution building.
Building a System of Self-Declarations
The SAT summarized its administration reform program succinctly, in what some Chinese tax administrators came to call the “thirty-character strategy.” This new approach would have “as its foundation taxpayer declarations” (i.e. self-assessment) and “enhanced taxpayer service”; it would “rely on a computer network”; “tax collection would be centralized” (i.e. taken away from individual tax collectors); and the emphasis in tax administration would “shift to audits.” The SAT then outlined the concrete measures for implementing each element of this strategy. For instance, it set five targets for urban tax administrators for 1997. First, all firms and sole proprietors were to begin filing tax returns that year. Second, tax agencies needed to open public-facing offices that would allow taxpayers to submit returns, make payments, and handle other tax matters, instead of relying on tax administrators to perform these tasks for them. Third, based on the transition to self-assessment and more centralized payments and services, grassroots tax administrators would begin to specialize. In particular, at least 40 percent of tax administration personnel should be assigned to audit tasks. Fourth, new internal divisions would be created in tax bureaus to replace existing ones. Fifth, and finally, while no particular target for computerization was stipulated for 1997, tax bureaus were responsible for setting up files or accounts for each taxpayer, so that the compliance activities of each could be recorded and monitored.
Both in the State Council–approved version of the 1997 SAT reform program, and in subsequent directives for implementing the program, the SAT provided further instructions about return filing, taxpayer services, and centralized collection.Footnote 20 I will turn to the implementation of these instructions shortly. It is useful to note, however, that in comparison to these aspects of tax administration, the SAT’s instructions for computerization and audits were vague. With respect to computerization, the SAT expected that technological upgrades would be gradual as well as decentralized. Specifically, there was no concrete plan for information systems to be developed at the provincial, let alone national, level.Footnote 21 Indeed, the primary role for computerization delineated in 1997 was one of assisting the new daily functions of Chinese tax administration: maintaining taxpayer accounts, handling tax returns, and tracking payments as well as other interactions between taxpayers and tax bureaus. Insofar as these administrative tasks would remain decentralized, it only made sense for the cost-saving, efficiency-enhancing use of computers to be decentralized as well. While the decentralized building of information systems would forego the benefits of the economy of scale, it allowed flexibility and avoided waste in a fast-evolving administrative environment. A nationwide information system that would allow data sharing across regions was not part of the plan.
With respect to building audit capacity, the SAT’s 1997 reform program offered several general guiding principles. First, like computer networks, audit units would be established in urban areas at the district and city levels. The lower audit units would regularly examine taxpayer affairs, while the city units would deal with important cases as well as process whistleblower claims. Second, tax authorities would begin to conduct routine audits (potentially of any taxpayer), special issue audits (where select issues would be examined in connection with a group of taxpayers), and special case audits (focusing on important cases of noncompliance detected through the two previous types of audits). Third, only one measurable outcome of audit capacity-building was prescribed: at least 40 percent of all tax administration personnel should be assigned to audit units.
The actual implementation of the 1997 tax administration reform involved the mobilization of extensive public and private resources. A tangible example is the construction and opening of taxpayer service halls (nashui fuwu ting) across the country. By the mid-2000s, these government premises had become ubiquitous features of China’s urban landscape, often housed in impressive buildings occupied by tax bureaus. While the advent of online filing and e-government in the past decade means that taxpayer service halls are no longer the bustling places they used to be, they were the high-speed railways of yesteryear in Chinese taxation: they allowed a massive population of Chinese taxpayers to perform compliance obligations in an unprecedented fashion. Not only could taxpayer registration, return filing, payments, tax invoices, and applications for tax preferences all be handled at these centers, but taxpayers would also often obtain information about tax law and policy. Indeed, they arguably made taxpayer service one of the strongest components of Chinese tax administration.Footnote 22
Less visible than the hardware of tax service centers, but no less important, is the system of processing taxpayer information. Tax returns, along with their respective instructions, were designed and published according to the 1997 reform plan. Often, the next step required proactive problem-solving on the part of local tax bureaus. Consider the story of one lowly tax bureau in the district of Huanggu, one of the ten districts in Shenyang, the capital city of Liaoning Province in northeast China. In 1995, the Huanggu GS branch bureau had jurisdiction over a taxpayer population that included close to 8,000 private firms. The branch bureau boasted of a 200-square-meter taxpayer service hall, but even this could not accommodate the crowd when all taxpayers came to file tax returns in person. Even with twenty staff members and ten desktop computers, handling in-person filing required the service center to open “without lunch break and late into the evening.” Long queues lined up.
To solve this problem, the Huanggu branch bureau entered into an agreement with the district’s postal bureau so that taxpayers could submit tax returns by mail. Special arrangements had to be made for this to work. For one, if taxpayers had to weigh their heavy envelopes stuffed with tax returns at post offices, the latter would themselves be clogged! In response, the Huanggu postal bureau agreed to offer a special, single rate for all mail sent to the local tax office regardless of weight. Indeed, the tax and postal bureaus found it worthwhile to print a new kind of envelope for mailing tax returns within the district. Special daily delivery services between taxpayers and the tax bureau were also launched. With these arrangements, the Huanggu branch bureau was able to invest in automatic envelope openers and scanners to process tax returns, in lieu of building a larger taxpayer service hall. This purportedly increased the speed of return processing by twenty times, helping to reduce the service center personnel from twenty to five.Footnote 23
This special arrangement with the postal service was adopted in the entire province of Liaoning by 1997. The SAT included a report of the practice in a national circular, promoting tax and postal bureau cooperation across China in the processing of tax returns.Footnote 24 Likewise, the coordination of grassroots tax bureaus and bank branches (in setting up taxpayer accounts that allowed direct bank transfers to government accounts) formed another theme of local experimentation.Footnote 25 Many similar stories can be told. The general point is that establishing a taxpayer self-declaration system was itself no simple feat. It required not only the undertakings of multiple government agencies but also massive mobilization among taxpayers. At least in the late 1990s, tax administration reform was not just talk. It was a radical transformation.
There is strong evidence from official statistics that the late 1990s saw several highly unusual years in Chinese tax administration. The government reported some of the highest numbers of taxpayer inspections in the country’s history, as if tax collectors around the country answered the call to increase audit activity.Footnote 26 Tax auditors also appeared to have had ample material to work with: in 1998, an aggregate of 220,519 whistleblower reports about tax evasion were filed across 5,879 whistleblower centers in the nation.Footnote 27 Even the State Council’s reference to tax dispute resolution as one of the four equal pillars of tax administration resonated in practice: close to 6,000 first-instance lawsuits were filed by taxpayers against tax agencies between 1998 and 2000.Footnote 28 This may not seem much in a country with over a billion people. But for comparison, the annual average quantity of tax cases in China for the 2006–2013 period was fewer than 400. The late 1990s, with an annual average of 2,000 cases, was practically the golden age of tax litigation in China.Footnote 29
There is a rich story to be told here. The activities of processing tax returns, providing services to taxpayers, and facilitating tax payments may seem quotidian, but they are foundational to modern tax systems. The institutionalization of such practices in China in the late 1990s is critical for understanding the system’s ensuing success. Or, at least, they are of equal significance as other features of Chinese tax administrations that have received more attention (e.g. the division between GS and DS bureaus). I have dwelled on the 1997 tax administration reform, however, not because of its richness and significance. Instead, I have done so because, despite its role in implementing practices that form the bedrock of China’s taxation system, very few people in China talk about the reform in the years that followed, and most scholars do not even acknowledge that it happened.
Reform Abandoned
There exists little to no scholarship, from Chinese or non-Chinese writers, on China’s 1997 tax administration reform. By contrast, less sweeping tax administration changes have attracted more commentary. For example, in 2002, the central government adjusted its revenue-sharing agreement with provinces to allocate a greater share of CIT revenue to itself. In connection with this change, for the period between 2002 and 2008, all new firms registered in China were required to remit the CIT to GS bureaus instead of DS bureaus. This change in CIT administration is examined in several scholarly studies, even though it is more piecemeal relative to the 1997 reform.Footnote 30 Similarly, several recent studies have probed the impact of the Golden Tax Project between 2000 and 2002 on firms’ VAT compliance.Footnote 31 Generally, Chinese scholars postulate that tax enforcement has steadily improved after 1994, and most attribute this improvement to the changing jurisdictional scope between DS and GS bureaus, the Golden Tax Project, or simply various government units’ “tax effort.”Footnote 32 The onset of the self-declaration system in 1997 is rarely mentioned.
I believe the Chinese government’s abandonment of key objectives in the 1997 tax administration reform may be a principal reason for the reform’s academic oversight. In China, the national media is dominated by the voice of the government. If the national government wants to elevate some reform to public prominence, such as the 2002 modification of CIT revenue sharing or the Golden Tax Project, the successes and societal impact of such reforms naturally occupy scholarly attention. By contrast, if the national government decides to relegate a policy that has fallen from favor to obscurity, scholars are unlikely to take notice.
There are various ways of tracing the gradual abandonment of the 1997 reform. For instance, in 1999, the Chinese government decided to embark on an overall reduction in government personnel across agencies. This exogenous shock certainly impacted tax administration: the GS system’s staff size shrank by over 48,000 in 2000. When the SAT set out the organizational and personnel plan for the GS system in 1997, implementing the tax administration reform was a central objective.Footnote 33 Yet the SAT’s next organizational plan, issued in 2000, made no mention of the reform (which was supposed to be ongoing until 2010), and focused instead on the new objective of reducing personnel and handing mass government layoffs.Footnote 34 From this point on, the goal of dedicating at least 40 percent of total staff to audit tasks ceased to receive mention in SAT documents.
But a more important development, referred to only obliquely in official documents, was the emergence of a critique of the 1997 reform agenda. According to this critique, the 1997 reform vision led to “the neglect of management, and the weakening of accountability” (shuyu guanli, danhua zeren). This critique never received an official exegesis. Commentaries circulated within Chinese tax administration suggest that it had two components. First, the “neglect of management” referred to a problem that arose from the separation of tax collection, audit, and the supervision or “management” of taxpayer affairs. This practice, as we saw earlier, had been advocated since the 1980s. The critique argued that while the 1997 reform entrenched the roles of tax collection and audits, tax administrators were becoming too removed from taxpayers: they simply waited for taxpayers to file returns and remit taxes, but no longer had the capacity to routinely control the level of taxpayer compliance. In essence, although audits were being conducted, they came only after infringements of tax laws had already occurred. Taxpayer compliance needed to be monitored more “dynamically,” according to the critics, and tax administrators needed to do more than completing audits (or providing taxpayer services) to ensure that taxpayers complied. But because of the “regrettable” emphasis on allocating personnel to audits and taxpayer services, taxpayer “management” had become thinly staffed.
Second, “the weakening of accountability” referred to the dearth of incentives within tax administration. Critics observed that when collection, audit, and taxpayer supervision were separated, it became harder to motivate tax administrators themselves. Administration outcomes now depended on multiple functions and agency units, making it harder to pin down the under-performance of individual persons and units. Additionally, performance metrics were harder to design. This problem of internal management was apparently aggravated by demographic factors. Many employees in the tax collection workforce had been hired in the 1980s. They had unsurprisingly low levels of education – the country’s education system was suspended during the Cultural Revolution – and were getting older. Career incentives were thus weak for a sizeable portion of the staff. Moreover, the general personnel reform being pushed across the government was affecting employee morale.
These arguments are puzzling. The “weakening of accountability” amounted to a management or human resource problem and did not seem to impugn the basic design of the new tax system. The accusation of the “neglect of management” appears even stranger. Chinese tax revenue grew at an average annual rate of 16 percent between 1997 and 2001, despite the Asian financial crisis in 1998 (see Figure 0.2). This rate of revenue growth was sustained during a period when an unfamiliar set of taxes and system of self-declarations were freshly introduced. Certainly, revenue could have grown even faster if Chinese taxpayers had been more compliant, but the reformers in 1997 had anticipated a fourteen-year timeline for building a modernized tax administration. Compliance was not supposed to improve overnight. Finally, in both the 1993 and 1997 reform statements, audits and penalties were envisioned as the main way (aside from taxpayer services) for ensuring taxpayers’ compliance, presumably through the power of deterrence. The suggestion that this was an inadequate form of taxpayer monitoring thus seems surprising.
Despite these possible replies, the charge that tax administration at the end of the 1990s was “neglecting management and weakening accountability” became influential. It was explicitly acknowledged in a SAT proposal issued in 2001 “to experiment with using information technology (IT) to promote tax administration reform.”Footnote 35 According to this proposal, tax administration across China should begin to adopt more uniform software, standardize administrative protocols through computerization, and move toward a more integrated information system. And along with developing this information system, clearer specialization among administrative roles would be delineated. The payoff of investing in IT was framed mostly in terms of improving internal management, seemingly responding to the complaint about “the weakening of accountability.” Its intended impact on taxpayers was not clear. Other SAT documents between 2000 and 2003 also indicate that tax administration reform entered an inward-looking phase, emphasizing the enhancement of IT, reorganization and centralization of administrative functions, and improvement of specialization and internal accountability.Footnote 36 The outlines of the four fundamental pillars of tax administration described by the State Council in 1993 (i.e. legislation, collection, audit, and dispute resolution) became obscured. Likewise, the emphasis on taxpayer self-declarations, service, centralized collection, audits (and IT to serve these tasks) faded.
However, none of these developments definitively signaled the reversal of the 1997 reform. After all, setting up a self-declaration system, along with the taxpayer services and audit capacity that complemented it, involved both large upfront investments and subsequent ongoing improvements. The phase of physical construction and taxpayer mobilization in the late 1990s represented the upfront investment. As these are naturally more dramatic than subsequent improvements, a quieter ensuing phase is just what one would expect.
The main reason to believe that the 1997 tax reform met its demise was that, beginning in 2003, the SAT came to advocate administrative changes that decidedly deemphasized taxpayer services and audits, while elevating taxpayer “management” to the highest prominence. According to this new position, there existed certain tasks in tax administration that were “more foundational” than taxpayer services and audits.Footnote 37 As a result, SAT officially augmented its thirty-character reform agenda from 1997: tax administration must now also “strengthen management” (qianghua guanli). The decision to switch from practices that “neglected management” to ones that prioritized management was associated with the arrival of a new Commissioner of the SAT, Xie Xuren, in 2004. Xie was perceived to be a critic of the initial results of the 1997 reform,Footnote 38 and he championed efforts to institute a “revenue manager” system throughout China.Footnote 39 The effort culminated in the adoption of two sets of “trial measures” by the SAT in 2005: the “System of Revenue Managers” and the “System of Taxpayer Evaluation.”Footnote 40
In subsequent chapters, we examine the 2005 institutions in some detail, studying both the letters of the 2005 SAT directives and how they worked in real practice. At this point, what is important to grasp is the systems’ general significance. The SAT’s 2005 announcement of the revenue management and taxpayer evaluation systems represented a clear abandonment of the 1997 reform, because the revenue manager system largely revived the “nanny-style” tax administration of the 1980s.Footnote 41 Tax administrators returned to overseeing particular taxpayers as opposed to performing specialized tasks. Their performance also came to be evaluated based on the compliance activities of these taxpayers. The main differences between the “dedicated taxpayer managers” of the 1980s and the “revenue managers” promoted in 2005 were that the latter did not physically collect tax returns or payments, and did not enjoy discretion in determining what policies applied to taxpayers. Moreover, revenue managers worked in parallel to a separate “tax inspection” system that specialized in formal audits and were supposed to use penalties more frequently. In other words, revenue managers were to ensure that taxpayers carried out their compliance activities, generally without using formal audits and imposing penalties. And, of course, revenue managers no longer supervised mainly SOEs and COEs. Instead, they were to “manage” large populations of private firms.
I want to stress the novelty of the story I have related so far. It is unclear whether most of China’s younger generation of tax policy makers and administrators today know about the 1997 tax administration reform. The reform’s astounding obscurity, despite its dramatic immediate impact, is what prompted me to use it as a centerpiece for discussing the administrative foundations of the Chinese fiscal state. What is equally surprising is the obscurity of the revenue management system. If one examines Chinese government directives about tax administration after 2005, it would be impossible to avoid reading at length about the bureaucratic details of revenue management. These details are further elaborated on and debated among the 800,000 tax administrators across China. Yet the system receives little acknowledgement in commentaries on Chinese taxation and is ignored in social scientific and legal scholarship. I believe this is because scholars have lacked the conceptual apparatus for understanding this quintessential aspect of Chinese taxation – and of the Chinese state. This book will gradually develop such an apparatus to bring out the practice’s social scientific significance.
The foregoing story raises the following key questions: What is the revenue management system, and why is it felt in China to be superior to a system that relies on voluntary taxpayer compliance and audits? The 1997 national tax administration reform and its subsequent demise illustrate the competition between these two paradigms of tax administration. The revenue management paradigm, having its roots in the 1980s (when China was still substantially closed to the world), may be considered native to China. The self-declaration-plus-audit paradigm, in contrast, may plausibly be viewed as imported: it is a style of tax administration that came with the new taxes that China decided to introduce in the early 1990s. The imported paradigm is easier to talk about, even in China – and most certainly for the readers of this book – because it is more internationally prominent. Yet it is the native paradigm that has the most tenacious hold in real-world practice in China. Indeed, the fate of the 1997 national reform is not the only example illustrating the triumph of the native paradigm. In Chapter 4, we will see that the competition between the two paradigms was staged again during 2012–2015, in a self-initiated tax administration reform in Jiangsu province. The native paradigm again turned out to be the winner in that competition.
Before diving into these topics, however, the remainder of this chapter comments on three features of Chinese tax administration that may be somewhat more familiar to readers because of existing scholarship and policy discourse: (1) the GS versus DS distinction, (2) bureaucratic decentralization, and (3) the use of IT. These are important features of the Chinese fiscal state, and I will have much to say about them in the rest of the book. But it is important to put them in place and define their relationship to the dominance of the revenue management system.
Bifurcated Administration, Decentralization, and Technology Use
Guoshui versus Dishui
The bifurcation of Chinese tax administration into the state tax bureau (GS) and local tax bureau (DS) systems lasted between 1994 and 2018. Both systems implemented tax law and policies set by the national government (as elaborated by the MOF and SAT). The best-known distinction between the two systems is political. The central government determined the budget, personnel, and leadership appointments of GS units in the country, whereas provincial governments controlled these matters for DS units within their respective provinces. In Chinese terminology, the GS system was subject to “vertical control” by the national government, while the units in the DS system were “vertically controlled” by provincial governments. There is no a priori answer to the question of how the national (as opposed to the provincial) determination of agency budgets might have affected administrative capacity. However, it is reasonable to postulate that appointments matter. Leaders of DS units were ultimately answerable to political bosses in their respective provinces. Leaders of GS units, by contrast, participated in a nationwide bureaucratic hierarchy with the SAT at its head. Senior bureaucrats in the GS system thus enjoyed career opportunities (e.g. taking up positions in different provinces or at the SAT) not available to their DS counterparts.
Does this political distinction between GS and DS bureaus carry significant implications? Many Chinese scholars have viewed “vertical control” by the national government as a solution to problems arising from excessive decentralization in many policy areas. In viewing the GS versus DS distinction through this lens, they suggest that the GS system exemplifies a superior form of organization.Footnote 42 But this is based on misunderstanding. These same scholars usually contrast “vertical control” by the central government with “tiered control,” where a government agency at a low (e.g. county) tier of government is accountable – for example, because of appointments – primarily to the political bosses at the same low tier. But the DS system does not present an instance of “tiered control”: DS units at all tiers in a province are accountable to the provincial leadership, which is at best a slightly lower level of leadership than the SAT. The better view is that the differences in incentives for lower-tier bureaucrats in the DS and GS are multidimensional, and not dominated by whether the SAT wields the power of appointment.
There are, in fact, many significant but lesser-known institutional differences between the GS and DS systems. Figure 1.1 showed that overall the GS system is larger in terms of staff size. The discrepancy is even larger in terms of revenue collected. The upper-left panel in Figure 1.2 shows that from 1997 to 2009, the GS system consistently raised about two times the revenue as the DS system. Furthermore, although the budgets of the DS and GS systems are not easily located, the GS system appears to command better resources in many ways. The upper-right panel shows that more of GS personnel is concentrated in tax bureaus at the county level and above, and the lower-left panel shows that, at these levels, GS bureaus are allowed to host more internal divisions. Both facts suggest higher levels of specialization for GS employees in these bureaus and potentially better career opportunities. The GS system also enjoys a far greater number of information centers (lower-right panel), as well as employees working in nonadministrative units that provide training and logistical support.
Figure 1.2 Comparisons between GS’ and DS’ resources
The DS system, however, has some of its own advantages. Figure 1.1 showed that the DS system experienced a much smaller personnel reduction than the GS system during the government reorganization in 2000. Because the DS system was built starting from 1994, it also benefitted from a workforce that is somewhat younger and better educated, as shown in Figure 1.3.Footnote 43 Moreover, since budgets are controlled by provinces in the DS system, there is likely to be greater variation in pay across provinces. This means that incentive pay for tax administrators can be arranged in ways that would not be feasible in the GS system.
Figure 1.3 Education and age composition in GS versus DS bureaus
Historically, the DS system was substantially more decentralized: it had more branch bureaus and outposts than GS bureaus and assigned more of its staff to these grassroots units. This is significant, especially when it interacts with the operation of the revenue management system. For these and other additional reasons,Footnote 44 without careful research design, it would be hard to attribute any observed differences in administrative outcomes between GS and DS bureaus to a single institutional factor (such as the fact that DS bureaus are more accountable to local political leaders). The most one can say about Chinese tax administration’s bifurcation into the GS and DS branches before 2018 is that it is always one of the factors to consider while analyzing Chinese tax administration; however, its predictive value is sensitive to context. And, most importantly, the competition between the native and imported styles of tax administration discussed earlier applies to both the GS and DS systems. Understanding either system requires a better grasp of these basic approaches to tax administration.
Administrative Decentralization
The decentralization of tax administration is one of the most striking features of Chinese taxation. Unfortunately, this feature has long been obscured by the way scholars (both in China and elsewhere) use the term “local” (difang) to describe Chinese governance. This term is often used to draw a line between national and subnational governments. When used so broadly, “local” may denote the provincial, prefectural/city, or district/county levels of government, which are the next three tiers of government (in descending order) in China.Footnote 45 Or it could denote governance activities at even lower levels, such as neighborhoods/townships. Failing to distinguish between these four tiers of subnational governance, however, omits essential facts about China’s administrative decentralization.
The GS and DS systems (before they merged in 2018) each established tax bureaus at the provincial, prefectural/city, and district/county levels. This implied totals of about sixty provincial bureaus in the thirty-one provinces, close to 700 city bureaus in the more than 300 prefectures across the country, and close to 6,000 district/county bureaus. But this count of bureaucratic units only includes full-service bureaus (quanneng ju) or bureaus that manage subordinate units (ju jiguan). Such bureaucratic entities supervise a much larger number of specialized or subordinate units: specialized units include inspection bureaus (discussed in Chapter 2), and subordinate units include branch bureaus and tax offices assigned to finer geographical jurisdictions within districts and counties. The number of subordinate outpost units is staggering. The left panel in Figure 1.4 shows that in 2013, the GS system had 10,128 below-county units and the DS system had 16,791. In 1995, these numbers were 26,715 and 32,723, respectively.
Figure 1.4 Decentralization in GS versus DS bureaus
Virtually each of these below-county units is a physical premise at which taxpayers interact with tax administrators to fulfill compliance obligations. What this means is that in 1995, there were close to 60,000 grassroots tax offices (i.e. not counting the full-service or management-oriented higher-level bureaus) that Chinese taxpayers could visit. This was comparable to the total number of post offices (61,898) China had that year! Although the number of tax outposts declined significantly over time (while the number of post offices grew), in 2012, Chinese taxpayers still enjoyed one tax office open for every three post offices. And China’s postal network is denser than many other countries. The US Postal Service, for example, had 30,825 post offices and locations in 2017, close to the total number of tax bureaus and offices in China. This staggering comparison with post offices is what one needs to understand the sweeping meaning of “local” tax administration in China.
This extraordinarily dense network of grassroots tax offices has enormous implications for both government personnel arrangements and taxpayer experience. China’s workforce of tax administrators comprised approximately 757,000 civil servants in the last fifteen years. Although this is certainly the largest tax bureaucracy in the world by absolute size, it is not so in per capita terms: the number of tax administrators relative to the size of the country’s general workforce is much lower than the OECD average.Footnote 46 However, by being organizationally bottom heavy, China’s tax administration permits most taxpayers to easily contact tax collectors.
Consider the following: at the national level, the SAT hosts just over 800 staff members – a mere 0.1 percent of the army of tax administrators. By contrast, in 2013 the US IRS National Office in Washington D.C. employed close to 5 percent of total IRS staff.Footnote 47 At the provincial level, China’s tax bureaucracy has 11,000 employees. Thus, even when the national and provincial offices are combined, employees at these levels represent only 1.56 percent of the tax administration.Footnote 48 Data compiled by the SAT shows that in 2003, 16 percent of subnational tax administrators were posted at prefectural tax bureaus, with the remaining 80 percent working at the county level or below. The right panel in Figure 1.4 graphs two longer time series capturing the proportion of the administrative workforce staffed at below-county units. Although this proportion declined generally over time – it was around 60 percent for both the GS and DS systems in the late 1990s – it was still well over a quarter in the GS system, and 40 percent in the DS system, by 2013.
Imagine what the US IRS would be like if it were structured like China’s tax bureaucracy. Most notably, the Washington D.C. National Office would essentially be empty. There would also be no counterparts to the seven IRS regional offices, each of which oversees matters across several states. Instead, the current 139 IRS district offices,Footnote 49 which are field offices that deal directly with taxpayers, would become high-level offices giving commands to three additional layers of subordinate offices, where most personnel would be located. Instead of having twenty Taxpayer Assistance Centers in the entire state of New York (or twenty-eight in California),Footnote 50 there would be twenty such centers just for upper Manhattan. Rather than having two or three hundred field offices in the entire country, the IRS would possess this same number of offices in just one metropolitan area. Instead of mailing their returns and writing checks to distant “return processing centers,” US taxpayers would be able to dial-up local IRS service hotlines where someone always answers the phone. If they wished, taxpayers could also simply handle their basic tax compliance and obtain taxpayer services in their neighborhoods.
The coming chapters examine what exactly happens at these extremely “local” sites of interaction. For now, we can label this type of state organizational structure as “administrative decentralization,” defined by the following two features. First, there is a single bureaucratic hierarchy, and government functions viz-a-viz citizens are performed at the lowest levels of this hierarchy. By contrast, higher levels of the bureaucracy do not exercise government power directly over citizens but instead issue commands to bureaucratic subordinates. Second, the lower the bureaucratic rank, the smaller the geographical reach of units in the rank. Decentralization thus implies that the jurisdiction of a citizen-facing government unit is geographically quite limited. What is unusual about China is first, how deep (i.e. multilayered) the bureaucratic hierarchy is, and second, how resolutely the tasks of government administration are placed at the bottom ranks of the hierarchy.
The Golden Tax Project
Like many other organizations, tax administration can benefit from a well-educated workforce and use of advanced IT. Chinese tax administrators are generally highly educated (Figure 1.3). China’s tax authorities have also invested heavily in IT since the late 1990s. The reason to discuss IT use here, however, is that a particular form of technology has gained special attention in recent social scientific research.
Specifically, the Golden Tax Project (jinshui gongcheng) is a nationwide IT project launched in the late 1990s that is far more ambitious than the computer networks contemplated by the SAT 1997 tax administration reform. It had a short unsuccessful pilot phase in 1994–1995 called Phase I. GTP Phase III, which attempts to integrate tax authorities’ IT systems across the country – allowing, among other things, automatic taxpayer data transmission to the SAT and across regions within the entire tax administration – did not become operational until 2017. It is GTP Phase II, which began implementation in 2000 and became fully operational across the country in 2003, that has attracted scholarly attention. GTP Phase II is a VAT anti-fraud system that requires large VAT-remitting firms to use special encrypted VAT invoices, deploy an extensive network of physical devices for the regular de-encryption and authentication of these invoices, and, to a more limited extent, perform cross-matching of invoices between business sellers and buyers.Footnote 51 The implementation of GTP Phase II helped the Chinese government staunch an alarming rise in VAT fraud in the 1990s and played a significant role in shaping China’s VAT policy.Footnote 52
But scholarly attention to GTP Phase II has arisen for a very different reason – and possibly on the basis of a misunderstanding. Researchers in developmental economics have in recent years portrayed the cross-matching of invoices as a key feature of the VAT, one that is particularly relevant to developing countries because of their low levels of tax compliance.Footnote 53 Since Chinese policy makers invested billions of yuan in GTP Phase II and believe that they have improved VAT compliance as a result, China’s practical experience seems to resonate with this new scholarly understanding. The problem is that Chinese policy makers are aware – contrary to the belief of scholars advocating extensive VAT invoice cross-matching – that invoice cross-matching is not part of the VAT’s traditional design. The vast majority of developed countries that initially adopted the VAT (including countries in Europe, Canada, Australia, and Singapore) do not have an invoice cross-matching system under their VATs. And it was based on the successful experience of VAT implementation (without cross-matching) in these rich countries that the VAT was promoted to many developing countries.Footnote 54 When China considered adopting a VAT anti-fraud system, there was little international experience to borrow from.Footnote 55 GTP Phase II, therefore, was very much an indigenous creation.
The launch of GTP Phase II was backed by Chinese Premier Zhu Rongji. It was probably the single most expensive IT project that the SAT invested in. Not surprisingly, national political leaders and the SAT gave it heavy promotion and much press coverage. An objective evaluation of the actual impact of the VAT anti-fraud system, however, still awaits thorough investigation.Footnote 56 In fact, assessment of the system by Chinese taxpayers and tax administrators is quite mixed.Footnote 57 For taxpayers, the VAT anti-fraud system created high compliance costs, even for many firms that would have been compliant already. For tax administrators, it created a technological arms race between the government and criminals that the former is never assured of winning. None of this, though, is unique to China. The VAT’s vulnerability to criminal fraud, including in developed countries, is well documented, and tax administrators in many countries must weigh the benefit of anti-fraud measures against their compliance costs. What I want to note here is that while China practices the use of IT in VAT invoice cross-matching, it is controversial domestically. The merit of VAT invoice cross-matching is also highly controversial internationally, despite its recent popularity among developmental economists.
The role of information in tax administration is a central topic for this book. In the following chapters, we will systematically investigate the institutions and practices that generate information. For example, if tax administrators enforce the norm of truthful reporting, the information transmitted from taxpayers to the government’s database is more likely to be reliable. The non-enforcement of truthful reporting, on the other hand, may render much of the government’s data useless, whatever advanced IT is used. It turns out that the revenue management system has a major impact on how information is gathered by Chinese tax agencies and the type of information gathered. To use a favorable term of economists, information is endogenous. Different approaches to tax administration will generate different types of information.
