Access the full text.
Sign up today, get DeepDyve free for 14 days.
(2009)
Cantax collection play a corporate governance role
Joseph Fan, T. Wong, Tianyu Zhang (2007)
Politically connected CEOs, corporate governance, and Post-IPO performance of China's newly partially privatized firmsJournal of Financial Economics, 84
(2008)
Corporate donations and economic incentives: An empirical study based on corporate donations following the 512 earthquakes in China
(2016)
Fiscal decentralization and corporate tax incentives: An analysis from views of local government competition
A. Leiter, H. Oberhofer, P. Raschky (2009)
Creative Disasters? Flooding Effects on Capital, Labour and Productivity Within European FirmsEnvironmental and Resource Economics, 43
M. Faccio, Ronald Masulis, John Mcconnell (2005)
Political Connections and Corporate BailoutsCorporate Finance: Governance
Ilan Noy (2009)
The macroeconomic consequences of disastersJournal of Development Economics, 88
(2011)
Media coverage and the social responsibilities of listed company: Empirical study base on the donation in Wen Chuan Earthquake
(2013)
Tax avoidance effect of political status
Joseph Fan, T. Wong, Tianyu Zhang (2013)
Institutions and Organizational Structure: The Case of State-Owned Corporate PyramidsJournal of Law Economics & Organization, 29
E. Boose, Mayra Serrano, D. Foster (2004)
Landscape and regional impacts of hurricanes in puerto ricoEcological Monographs, 74
P. Landry (2008)
Decentralized Authoritarianism in China: The Communist Party's Control of Local Elites in the Post-Mao Era
M. Pelling, Alpaslan Özerdem, S. Barakat (2002)
The macro-economic impact of disastersProgress in Development Studies, 2
E. Cavallo, Ilan Noy (2011)
Natural Disasters and the Economy — A SurveyInternational Review of Environmental and Resource Economics, 5
C. Raddatz (2005)
Are External Shocks Responsible for the Instability of Output in Low-Income Countries?Development Economics
R. Baron, D. Kenny (1986)
The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.Journal of personality and social psychology, 51 6
Zhaoyang Gu, Song Tang, Donghui Wu (2015)
The Political Economy of Labor Cost Behavior: Evidence from China
Gabriel Felbermayr, J. Gröschl (2013)
Naturally Negative: The Growth Effects of Natural DisastersSRPN: Other Built Environment (Topic)
(2016)
Economic cycle and state-owned companies’ tax avoidance
(2014)
Are Chinese enterprises’ charitable donations ‘political contributions?’ – Evidence from the replacements of the municipal party secretaries, 2
T. Atwood, Michael Drake, James Myers, Linda Myers (2012)
Home Country Tax System Characteristics and Corporate Tax Avoidance: International EvidenceSocial Science Research Network
G. Holland (2008)
A revised hurricane pressure-wind modelMonthly Weather Review, 136
(2013)
The bond effect of firms’ charitable donation: Empirical evidence from Chinese listed company
(2007)
Tax refunds, the corporate tax burden and the effectiveness of tax policy
P. Raschky (2008)
Institutions and the losses from natural disastersNatural Hazards and Earth System Sciences, 8
Olivier Dessaint, Adrien Matray (2016)
Do Managers Overreact to Salient Risks? Evidence from Hurricane StrikesERN: Econometric Studies of Corporate Governance (Topic)
(2016)
Policy uncertainty, tax collection strength and corporate tax avoidance
(2015)
Institutional investor tax enforcement and tax avoidance
(2013)
What makes 10 million local government officials tick
(2012)
Political connection and the tax avoidance of enterprises: The evidence
김양민, Albert Cannella (2004)
SOCIAL CAPITAL AMONG CORPORATE UPPER ECHELONS AND ITS IMPACTS ON FIRM PERFORMANCE
R. Bennett (1998)
Market Orientation Among Small to Medium Sized UK Charitable Organisations: Implications for Fund-Raising PerformanceJournal of Nonprofit & Public Sector Marketing, 6
(2012)
Location choice of industrial firms in China: Market potential, resource endowment and tax burden level
K. Emanuel (2011)
Global Warming Effects on U.S. Hurricane DamageWeather, Climate, and Society, 3
Ilan Noy, A. Nualsri (2011)
Fiscal storms: public spending and revenues in the aftermath of natural disastersEnvironment and Development Economics, 16
(2015)
The political dynamics of corporate tax avoidance: The Chinese experience
(2011)
Tax administration, income tax cost and earnings management
Hideki Toya, M. Skidmore (2007)
Economic Development and the Impacts of Natural DisastersEconomics Letters, 94
C. Bai, David Li, Z. Tao, Yijiang Wang (2000)
A Multitask Theory of State Enterprise ReformJournal of Comparative Economics, 28
T. Vu, David Hammes (2010)
Dustbowls and High Water, the Economic Impact of Natural Disasters in China
(2013)
Property rights, political connections and corporate tax: Evidence from listed companies in China
(2009)
Inter-governmental tax competition in China: Based on the province-level panel data
E. Goldman, Jongil So, Jörg Rocholl (2012)
Politically Connected Boards of Directors and the Allocation of Procurement ContractsPolitical Economy (Topic)
(2014)
Are Chinese enterprises’ charitable donations ‘political contributions?
R. Elliott, E. Strobl, Puyang Sun (2015)
The local impact of typhoons on economic activity in China: A view from outer spaceJournal of Urban Economics, 88
(2009)
Fiscal decentralization, local government competition and the actual tax burden of listed companies, 4
(2012)
Does ownership nature affect tax sensitiveness? Evidence based on firm’s labor demand
(2013)
Tax competition, tax enforcement and tax avoidance, 9
T. Vu, Ilan Noy (2015)
Regional effects of natural disasters in China: investing in post-disaster recoveryNatural Hazards, 75
Michelle Hanlon, Shane Heitzman (2010)
A Review of Tax ResearchCapital Markets: Asset Pricing & Valuation
Mihir Desai, Dhammika Dharmapala (2004)
Corporate Tax Avoidance and High Powered IncentivesTax Law: Tax Law & Policy eJournal
A. Muller, G. Whiteman (2009)
Exploring the Geography of Corporate Philanthropic Disaster Response: A Study of Fortune Global 500 FirmsJournal of Business Ethics, 84
E. Cavallo, Patricio Valenzuela, Eduardo Borensztein (2007)
Debt Sustainability Under Catastrophic Risk: The Case for Government Budget InsuranceSRPN: Developing World (Topic)
Jun Liu, Feng Liu (2013)
Fiscal centralization, government control and corporate tax burden: Evidence from ChinaChina Journal of Accounting Studies, 1
Thomas Kubick, G. Lockhart, Lillian Mills, John Robinson (2017)
IRS and corporate taxpayer effects of geographic proximityJournal of Accounting and Economics, 63
Stuart Gillan, L. Starks (2000)
Corporate Governance Proposals and Shareholder Activism: The Role of Institutional InvestorsJournal of Financial Economics, 57
(2014)
Institution environment, state ownership and the corporate effective tax rate
(2012)
Disaster prevention and mitigation investment, disaster control and economic growth: Economic analysis and empirical in China
Sanjeev Bhojraj, P. Sengupta (2001)
Effect of Corporate Governance on Bond Ratings and Yields: The Role of Institutional Investors and Outside DirectorsRobert H. Smith School of Business Research Paper Series
(2017)
The selfish goodwill: New evidence from corporate philanthropy in typhoon
T. Wong (2016)
Corporate Governance Research on Listed Firms in China: Institutions, Governance and Accountability
China Journal of aCC ounting StudieS , 2017 Vol . 5, no . 4, 468–509 https://doi.org/10.1080/21697213.2017.1429531 Natural disaster, fiscal pressure and tax avoidance: a typhoon- based study* a b c Liu Siyi , Weng Ruoyu and Yang Daoguang a b School of Management, Xiamen university, Xiamen, China; t he Wang Yanan institute for Studies in economics, Xiamen university, Xiamen, China; Business School, university of international Business and economics, Chaoyang district, Beijing, China ABSTRACT KEYWORDS g overnment-company Natural disasters affect not only local economic development and relationship; natural government fiscal behaviour at the macro level, but also corporate disasters; tax avoidance; tax avoidance behaviour at the micro level. From the perspective of typhoon typhoons, we investigate the effect of natural disasters on corporate tax avoidance behaviour based on the sample of China’s A-share companies from 2008 to 2015. Results indicate that the greater the damage caused by typhoons, the lower the level of tax avoidance of local companies, and this relationship is more significant in state-owned companies, political affiliates, and local leading companies, which is joined with path testing to show that natural disaster affects tax avoidance through fiscal pressure. In addition, we investigate whether this relationship varies cross-sectionally from several perspectives: institutional and economic environment, fiscal pressure, market pressure, and tax avoidance margin. Furthermore, we find that companies paying more taxes post-disaster will obtain more government grants and credit resources in the future, achieving mutual benefits between local governments and companies. In summary, this paper tests the effect of disasters on corporate tax avoidance behaviour based on China’s unique government-company relationship. 1. Introduction In recent years, the effect of natural disasters on global economic activities has increased. In addition to the direct economic losses caused by the destruction of production and infra- structure, natural disasters cause local population casualties and economic stagnation, threatening the stability of the local social order (Pelling, Özerdem, & Barakat, 2002). Therefore, understanding the effect of natural disasters and how to manage them remains a public concern (Zhuo & Duan, 2012). Scholars have conducted a series of valuable studies and discussions on the economic consequences of different types of natural disasters, such as Cavallo (2011) and Felbermayr and Gröschl (2014). China is one of the few countries to experience the most extensive and severe natural disasters in the world because of its vast territory, abundant terrain, and its climate. Since the 1990s, the proportion of economic CONTACT Yang daoguang dgyang@uibe.edu.cn Paper accepted by Kangtao Ye. © 2018 a ccounting Society of China CHINA JOURNAL OF ACCOUNTING STUDIES 469 losses caused by natural disasters to GDP has increased on average by 2.7% per year. Therefore, most research focuses on natural disasters in China, such as Vu and Hammes (2010). However, these studies mainly focus on the macroeconomic level and lack the nec- essary attention and in-depth exploration of the economic consequences at the micro level. To the best of our knowledge, only some articles examining the donation behaviour of a particular post-disaster company discuss the latter issue (Pan, Weng, & Liu, 2017; Shan, Gan, & Zheng, 2008; Xu, Xin, & Zhu, 2011; Zhang, Zhang, Wang, & Shen, 2010). In addition to the reputation factor, these studies reveal that meeting the needs of government and maintain- ing the relationship between local governments and companies are important motivations for post-disaster corporate charitable donations (Muller & Whiteman, 2009). However, from the government perspective, local governments have more control over corporate tax than over donations and more specific autonomy after tax collection. As a result, the demand for post-disaster tax revenue is even more urgent. From the company point of view, com- pared with donations, the amount and importance of taxes is much greater. Companies are more flexible in adjusting the extent of tax avoidance, so the scope of influence is broader. Therefore, this article focuses on the following questions: do natural disasters affect corporate tax avoidance behaviour at the micro level? Is corporate tax avoidance behaviour related to the severity of the disaster? Does this relationship reflect some form of interaction between companies and local governments? To answer the above questions, this paper considers typhoons and uses the meteorolog- ical wind field model and open data from the National Weather Bureau to quantify the degree of typhoon damage in prefecture-level cities. We use typhoons for several reasons. The first reason is their frequency and severity. A typhoon is one of the most frequent and severe natural disasters in the world, with losses of about US$26 billion a year. China suffers an average of seven typhoons each year, resulting in an economic loss of about US$5.6 billion (Elliott, Strobl, & Sun, 2015). Indeed, the typhoon-affected coastal areas are the most densely populated areas in China, with the most concentrated production and trade activities. Therefore, testing the effect of disasters from the perspective of typhoons is of great eco - nomic importance. The second reason is the separability of the effect. On the one hand, compared with typhoons, large-scale disasters such as earthquakes are likely to spread to other areas due to their emotional, political, and cultural causes and even affect the whole country. This spillover effect hinders the distinction between disaster areas and non-disaster areas, and differences in government and corporate tax behaviour between affected regions and less affected regions. On the other hand, earthquakes frequently occur in areas with poor natural environments and in economically underdeveloped areas; hence, the discov- ered effect is likely to be a fixed effect of specific areas. The third reason is data availability. Typhoons occur more frequently than other natural disasters and the number of typhoons affecting coastal areas increases each year. Therefore, meteorological models can be used to calculate the specific damage caused by typhoons in different regions each year to build panel data. Based on the sample of China’s A-share companies from 2008 to 2015, we find that the greater the damage caused by typhoons in coastal cities, the lower the level of tax avoidance in China, a corporate donation is always donated to the people afflicted by a natural disaster directly or via the red Cross after the disaster. So, the local government has limited influence on it (Zhang, Ma, & Zhang, 2013). 470 L. SIYI ET AL. of local companies, and this phenomenon is more significant in state-owned companies, political affiliates, and local leading companies, and this is joined with path testing to show that a natural disaster affects tax avoidance thorough fiscal pressure. At the same time, local fiscal pressure, economic growth pressure, and institutional environment all affect this rela- tionship to varying degrees. This paper also reveals that typhoons have a smaller effect on corporate tax avoidance in companies with higher institutional holdings and analyst follow- ing. The reduction of post-disaster tax avoidance mainly exists in companies with an effective tax rate lower than their nominal tax rate and a greater tax avoidance margin. In addition, we examine the economic consequences of corporate tax behaviour during natural disasters. Results show that when local typhoons are more destructive, if companies pay more taxes they will receive more government subsidies in the future and get debt financing at a lower cost. Results also demonstrate that if companies share the post-disaster financial pressure of local governments, they will receive government feedback or compensation afterwards, reflecting a two-way beneficial relationship between local governments and companies. The contributions of this paper are as follows. First, we extend current research on the economic consequences of natural disasters from the macro level to the micro level, and further expand the evidence at the corporate micro level. Current research mainly discusses the effect of natural disasters on macroeconomic growth, or focuses on the regulatory role of institutional factors in different countries and regions in the relationship between disasters and economic consequences. However, little research analyses how disasters affect micro - economic activities. Although some studies use a disaster event as a window period to study corporate charity donation behaviour, all of them use one disaster event cross-sectionally to study other factors affecting business behaviour during the disaster. They do not consider the characteristics of natural disasters, such as the extent of damage, as factors influencing the explanatory variables of business behaviour. By quantifying the degree of typhoon dam- age in die ff rent years and regions, this paper distinguishes die ff rent degrees of disaster effect to separate the characteristics of the disaster itself from the disaster event. This enables us to test the effect of typhoons with various degrees of damage on corporate tax avoidance behaviour. Second, this paper enriches the literature on the determinants of corporate tax avoidance behaviour. Research in this area mainly focuses on the basic characteristics of companies (such as firm size, financial leverage, and overseas operations), individual managers, and the characteristics of corporate management and governance. Hence, the discussion of macro- economic factors is clearly insufficient. From the macro perspective, some studies focus on government tax collection, macroeconomic cycle, and political uncertainty. Yet, the literature rarely discusses how the interaction between local governments and companies (or interest exchange) affects corporate tax behaviour. Based on the government-company relationship and tax avoidance behaviour, this paper constructs the logical chain of “natural disaster-local economy-local government motivation-corporate tax,” thus greatly complementing the current literature. Finally, this paper deepens the understanding of local government behaviour and the relationship between local governments and companies in the context of China’s special system. Among the official promotion mechanisms in China, social stability has a one-vote veto effect, while economic growth, investment levels, and fiscal revenue are hard targets CHINA JOURNAL OF ACCOUNTING STUDIES 471 (Green, 2013). Natural disasters affect not only local economic growth, but also local social stability, which is closely related to the personal promotion of local officials. As a result, officials are motivated to participate in post-disaster rescue and reconstruction. However, in the institutional setting of budget constraint and fiscal decentralisation, officials must seek other sources of income to ease the fiscal strain caused by emergencies, such as disas- ters. To establish and maintain their relationships with local governments, companies actively reduce their tax avoidance level. As a result, this paper takes typhoons as its research object and provides incremental evidence of the two-way interaction between local governments and companies. As companies share government fiscal pressure by reducing their tax avoid - ance, they also receive more resources, such as government grants and financing conces- sions, in subsequent periods. The remainder of the paper is organised as follows. Section 2 reviews the related literature and develops hypotheses. Section 3 describes the research methodology. Section 4 presents the main empirical results. Section 5 presents the additional analysis and robustness tests. Section 6 concludes the study. 2. Related literature and hypothesis development Following a natural disaster, local economic growth suer ff s a significant negative short-term effect (Raddatz, 2007). This effect is reflected in the following aspects. First, severe damage to local materials and infrastructure directly affects current economic output and growth. Second, casualties and declining post-disaster employment rates threaten local social sta- bility (Leiter, Oberhofer, & Raschky, 2009; Toya & Skidmore, 2007). Vu and Hammes (2010) find that regional economic output is closely related to the number of people affected by an earthquake in a specific region. More directly, Felbermayr and Gröschl (2014) construct a disaster damage index, and find that GDP per capita decreases by at least 6.83% in the most disastrous 1% sample, yet does not exceed 0.01% in 25% of the lowest sample. Elliott, Strobl, and Sun (2015) obtain similar results based on typhoon destructiveness and urban night-time lighting data – the areas most affected by typhoons suffer a significant 20% drop in economic activity during the year. However, due to differences in institutional environment and economic structure, the effect of disasters on the economy varies from country to coun - try. In general, governments can mitigate and respond to the adverse effect of disasters through effective fiscal policies in countries and regions with a high degree of education, a relatively strong legal system, a certain level of economic development, and a high degree of openness to the outside world (Noy, 2009; Raschky, 2008). Therefore, the varying fiscal policy and behaviour of governments are major topics in the study of the economic conse- quences of disasters. Based on the fiscal data from 22 developed countries and 20 developing countries, Noy and Nualsri (2011) find that in developed countries, governments usually play the role of ‘hand of support’. They often ease the pressure through tax cuts after a as green (2013) summarised, the metrics used to evaluate local leaders’ performance consist of three tiers: ‘veto ’ targets, including social stability and birth rates; ‘hard’ targets, including economic growth, investment, and fiscal revenues; and ‘soft’ targets, including education and health, cultural activities, and pension coverage. Based on proprietary performance appraisal data for municipal leaders, landry (2012) shows that meeting the ‘hard’ targets results in higher scores and determines leaders’ career success, meeting the ‘soft’ targets has little effect on the overall performance assessment, and the ‘veto’ targets are more ‘make or break’ in nature, which means failing such targets may result in demotions, punishments, or dismissal (Chen, t ang, Wu, & Yang, 2015, pp. 7–8). 472 L. SIYI ET AL. disaster. Conversely, in developing countries, government actions lead to a significant increase in fiscal revenue after a disaster, which is a completely opposite result. In China, local governments are responsible for post-disaster rescue and reconstruction. The effective response to natural disasters is linked to social stability and economic growth, which are key political tasks of local governments (Bai, Li, Tao, & Wang, 2000). If people affected by a disaster cannot get timely and effective rescue and resettlement, it will lead to local social upheaval, which in turn will lead to a veto of local government officials in future promotions. In GDP-focused promotion championships, economic losses due to dis- asters may affect the promotion of officials. To maintain social stability and reduce direct and indirect economic losses caused by disasters to local communities, local governments must provide fiscal resources to resettle affected people and increase government invest - ment to rebuild damaged infrastructure (Vu & Hammes, 2010). For example, Vu and Noy (2015) test the post-disaster investment behaviour of local governments, and note that to avoid the current or next year’s economic slowdown, local governments mitigate the disaster effect by increasing the fiscal expenditure and government investment for post-disaster reconstruction. This phenomenon is more pronounced in economically underdeveloped areas. At the same time, a greater number of casualties also significantly increase post-dis- aster government investment and fiscal expenditure. Natural disasters cause extensive and deep damage. Moreover, natural disasters have a greater effect on the economy, and thus the required fiscal expenditure is enormous. Under the current budgetary constraints, in addition to a small part of the central government’s earmarked funds, local governments are under tremendous financial pressure during the short-term rescue and reconstruction process (Zhang, Ma, & Zhang, 2013). Local govern- ments must find ways to increase their revenue. As the most important component of fiscal revenue, tax revenue is likely to be the main target of local governments. In the context of the current shared tax regime in China, corporate income tax remains the main source of revenue for local governments (Chen et al., 2015). By obtaining incremental tax revenue, local governments can unilaterally achieve their fiscal goals by adjusting their tax regulatory efforts, or they may take the initiative of companies to meet the fiscal needs of local gov - ernments (Chen et al., 2015). We ask therefore, whether the effect of disasters on corporate taxes stems from the tightening of local tax regulation or from businesses’ initiative to meet them. Considering these two possible explanations, this paper argues that the latter is more dominant. First, from the government point of view, the latter explanation is more in line with eco- nomic rationality. Although there is no consistent conclusion about the relationship between tax revenue and economic growth, the generally accepted view is that a high tax burden negatively affects the inflow of regional investment and economic growth (Ye & Ye, 2012). Therefore, it is not difficult to understand why local governments are likely to start a fierce tax competition. After the 1994 tax reform, the central and local governments levied their respective income tax. As different regions and industries benefited from different tax rates, local governments generally adopted tax policies, such as ‘return after levy’, to attract external capital inflows (Wu & Li, 2007). When the State Department banned this policy in 2000, the disguised reduction of the statutory tax rate ceased to exist. Therefore, local governments can only reduce the corporate burden by adjusting tax collection and administration, espe- cially the corporate income tax, which has more room for regulation (Fan & Tian, 2013). Current research shows that relaxing tax supervision creates a friendly business environment CHINA JOURNAL OF ACCOUNTING STUDIES 473 and an enlightened government image, thereby attracting more companies to invest locally (Cao, Liu, & Zhang, 2009; Guo & Li, 2009). In addition, the accounting and tax systems are complex, and there is some uncertainty about the information exchange between corporate management and tax regulators (Kubick, Lockhart, Mills, & Robinson, 2017). As local gov- ernments tighten tax administration, it is also possible for firms to conceal real information to avoid tax, rigging the game between the two parties and eventually leading to a decrease in overall efficiency. Under normal circumstances, local governments do not tend to strengthen tax supervision, or at least do not perform more severely than other regions. Yet, when there is an urgent need for fiscal resources (such as the need for a large amount of government funding after a disaster), local governments may quickly obtain fiscal support by tightening tax supervision. Nevertheless, this will only happen if local governments have no other better choices. In fact, local governments can ensure that some companies volun- tarily reduce tax avoidance or even pay more taxes through their ‘tacit understanding’ with local companies. Although tax collection may also be increased for specific companies (such as local state-owned companies), it is essentially the choice of a formal government system and will certainly affect the expectations of other companies or potential companies, result - ing in additional costs. Therefore, as local governments can achieve this through more dis- creet and moderate private channels, they are even less likely to choose a formal system. Second, from the company point of view, the potential benefits of the initiative are higher. The relationship between governments and companies is widespread across countries and regions, and differs only in terms of scope and invisibility. Governments at all levels still control the distribution rights of most resources, including investment opportunities, financ - ing channels, and government grants (Dai, Pan, & Feng, 2014). The companies’ ability to establish contact with local governments directly affects their advantages in the competition for scarce resources (Faccio, Masulis, & McConnell, 2006; Goldman, Rocholl, & So, 2013; Zhang et al., 2010). However, access to government resources is not without cost. Indeed, companies must undertake more social and political tasks for local governments in addition to their daily operations (Fan, Wong, & Zhang, 2013). For example, companies must consider charity (Zhang et al., 2013) or job security (Gu, Tang, & Wu, 2016). Wise businesses understand that local governments may face fiscal pressure when they decide to tighten tax collection and management; thus, businesses take the initiative to adjust their level of tax avoidance as a practical action to support post-disaster reconstruction. In the short term, reducing tax avoidance increases the tax burden, resulting in lower cash flow and net profit in the income statement, which seems economically irrational. However, in the long run, governments generally compensate companies in other forms (such as government subsidies and credit resources) and, more importantly, this exchange strengthens the relationship between local governments and companies. Indeed, if local governments tighten tax collection and admin- istration before companies take any action, it will not only be difficult for companies to avoid tax, but it will also weaken their relationship with local governments. In summary, as typhoons affect the local economy and social stability, local governments increase the fiscal pressure. As a result, to establish and maintain a good relationship with local governments, companies take the initiative to reduce tax avoidance to meet the gov- ernment financial needs. Therefore, the more severe the damage caused by typhoons, the higher the government tax pressure and the lower the degree of tax avoidance. Based on this observation, the first hypothesis of this paper is stated as follows. 474 L. SIYI ET AL. H1: Ceteris paribus, the higher the extent of typhoon damage in the area where the companies are located, the lower the degree of corporate tax avoidance. Compared with private companies, state-owned companies are naturally connected to the government (Luo & Yang, 2013). After the disaster, state-owned companies must assume the corresponding responsibilities naturally. State-owned companies not only aim at max- imising profit, but also undertake more social and political responsibilities than private com - panies while assuming economic responsibilities (Bai, Li, Tao, & Wang, 2000). Cooperating with local governments in post-disaster rescue and reconstruction is both a social respon- sibility and a political task. In terms of tax behaviour, it is more important to share govern- ment responsibilities than to retain corporate profit. From an executive-level perspective, the dual attributes of ‘management’ and ‘political bureaucracy’ determine that profit is not the only consideration when making decisions; executives must also take into account their own political career (Wong, 2016). In addition, as state-owned company executives have hardly any shareholdings in the company, the relationship between personal wealth and performance is not as close as in private-owned companies. The tax initiative during disasters is stronger than that of private companies (Wang, Wang, & Peng, 2012). Therefore, state- owned company executives are strongly committed to meeting the government fiscal needs. By sharing post-disaster government responsibilities, state-owned company executives can seek the leaders’ trust and better prospects. In addition, regardless of the level of the com- pany or the executive, the close relationship with local governments helps companies under- stand government needs more accurately and conveniently, and reduces the risk in the interest exchange process. To sum up, compared with non-state-owned companies, state- owned companies have stronger incentives and more opportunities to reduce tax avoidance activities in response to the government fiscal needs when the region is under severe fiscal pressure. Likewise, companies with political connections have similar motives and opportunities. First, politically related companies also have strong motives to maintain their relationship by meeting government needs. When governments encounter difficulties and crises, pro - viding timely and urgent help to governments is undoubtedly a more effective way to help companies maintain beneficial bilateral relations with local governments and enhance gov - ernment trust. For executives, political affiliation is not only an important social capital that can bring benefits to individuals (Fan, Wong, & Zhang, 2007; Kim & Cannella, 2008), but also an important manifestation of personal social status. Politically connected executives can build closer personal relationships by reducing corporate tax avoidance, and they can even get a higher connection level in the future. Second, the chairman or CEO of a politically connected company usually holds or has held positions in the local government, the National People’s Congress (NPC), or the Chinese People’s Political Consultative Conference (CPPCC) and other departments and inevitably has direct or indirect contact with local government officials. Therefore, after the disaster, government officials can communicate more directly with managers of politically connected companies. Managers also understand the govern- ment fiscal needs quickly and more clearly. Therefore, compared with companies without political connections, politically connected companies are more motivated and have more opportunities to reduce their tax avoidance level to support local governments. As a result, the effect of the typhoon’s destructive power on tax avoidance is even more pronounced. In addition, whether a company occupies a local leading position also affects its post-dis - aster tax avoidance. Leading companies in the region usually benefit from more public CHINA JOURNAL OF ACCOUNTING STUDIES 475 services provided by local governments and have closer ties with them. To maintain their benign interaction with local governments, they are willing to offer them more tax support. Therefore, for local leading companies, the negative effect of typhoon damage on the degree of tax avoidance is more significant. Based on the above analysis, we present the following hypothesis. H2: Ceteris paribus, the negative relationship between the extent of typhoon damage and the degree of corporate tax avoidance is more pronounced in state-owned companies, politically connected companies, and local leading companies. 3. Research design 3.1. Sample and data source China revised its Enterprise Income Tax Law of the People’s Republic of China in 2007, effective 1 January 2008. The new law adjusted the corporate tax rate and pre-tax deduction methods. To mitigate the effects of the corporate tax reform, our sample period starts in 2008. As typhoons mainly affect coastal cities, our sample consists of A-share listed companies in China’s coastal areas from 2008 to 2015, while we use listed companies at the national level for the robustness test. Next, we filter the sample according to the following methods. Taking into account the uniqueness of the fiscal characteristics and in accordance with previous research, we exclude financial and ST companies. According to studies on tax avoidance (Chen et al., 2015), to avoid the effects of extreme values, we also exclude sample observa- tions with negative owner’s equity, negative pre-tax profit, and negative corporate tax expense and those in which the effective tax rate is lower than 0 or greater than 1. Observations with missing data for variables are also removed. Our final sample thus consists of 6999 firm-year observations. The financial data of listed companies and the market data come from the China Stock Market and Accounting Research Database (CSMAR). We obtain the applicable statutory tax rate data and GDP data from the Wind database. Other macro- economic data, such as fiscal revenue and expenditure, are collected from the CEIC database and the National Bureau of Statistics of China. We obtain the data related to typhoons from the Regional Specialized Meteorological Centre (RSMC) Best Track Data, which, since 1949, has provided six-hourly data on locations, intensities, and central pressures on all tropical cyclones in the West Pacific. These data are then converted to damage degree using mete - orological models. 3.2. Regression model We estimate the following OLS model to test our hypotheses: DETR = + TYPHOON + X + X ∗ TYPHOON + Controls + + + + (1) 1 2 3 i where DETR is the dependent variable, defined as the difference between the effective tax rate and the applicable statutory tax rate, to measure the level of corporate tax avoidance. A less negative DETR value suggests less tax avoidance. TYPHOON indicates the degree of in the Chinese stock market, St (Special t reatment) is a sign for abnormal financial conditions or other situations, e.g. two consecutive periods of loss. 476 L. SIYI ET AL. Table 1. Variable definitions. Variables Definition DETR d egree of tax avoidance. income tax expense divided by pre-tax profit minus applicable statutory tax rate TYPHOON_PROV t yphoon damage degree in province-level. Sum up t yphoon damage value f at province-level ijt TYPHOON_CITY t yphoon damage degree in city-level. Sum up t yphoon damage value f at city-level ijt SIZE t he natural logarithm of a firm’s total assets at the fiscal year end ROA t he ratio of return to total assets at the fiscal year end LEV l everage. t otal liabilities divided by total assets at the fiscal year end BM t otal assets at the fiscal year end divided by the total market value TANGIBLE t angible assets intensity.net value of fixed assets divided by total assets at the fiscal year end INTANGIBLE intangible assets intensity. net value of intangible assets divided by total assets at the fiscal year end FOREIGN an indicator variable that takes a value of one if a firm has a foreign-related operation for fiscal year t, and zero otherwise LCITYGDP regional economic level. t he natural logarithm of total gdP for fiscal year t MARKET t he natural logarithm of the marketisation index damage caused by typhoons. It is measured at the province level (TYPHOON_PROV) and city level (TYPHOON_CITY), where a higher value indicates greater damage caused by typhoons at this location during this year. In Hypothesis H2, we construct the interaction effects using variables for property rights (SOE), political connection (PC), and leading enterprise (FLAGSHIP) with the typhoon variables, and in model (1), X represents the above characteristics. Controls includes variables for corporate financial characteristics and regional level. In addition, we control the year- (μ) and industry-fixed (τ) effects. As the typhoon variables measure the local damage of listed companies in one year, we control the urban-fixed effects (ω) to avoid the potential effects of local city characteristics on tax avoidance, which also mitigate the area omitted variable bias. The detailed definitions of the variables are presented in Table 1. 3.3. Measures on degree of typhoon damage Instead of directly using the typhoon level as a measure of the degree of typhoon damage, we depend on a more objective model of typhoon damage to calculate the continuous effect of each typhoon. In this model, we consider the duration and wind speed of each time point, and examine the specific effects on different regions according to the distance. Indeed, the typhoon level generally focuses on the maximum typhoon winds at certain time points, which fails to measure the eec ff ts on dier ff ent regions. In addition, current typhoon damage models all come from objective historical typhoon data, while measures based on physical attributes help minimise the noise caused by anthropogenic statistics (Elliott et al., 2015). Following Elliott et al. (2015), who suggest that wind speed can be used as a proxy for the damage caused by a tropical storm, we use the index of the degree of typhoon damage to measure the extent of typhoon damage caused in the area where the company is located. The calculation of the index involves the wind field model and the damage function (Boose, Serrano & Foster, 2004). The wind field model determines the corresponding wind speed in the target area, while the damage function transforms the wind speed into a damage value in the target area. The specific models are as follows: � � � � B B � � jt jt ⎡ ⎤ V R R � � �� h,jt m,jt m,jt ⎢ ⎥ V = GF V − S 1 − sin T exp 1 − ijt m,jt ijt (2) ⎢ ⎥ 2 R R ijt ijt ⎣ ⎦ CHINA JOURNAL OF ACCOUNTING STUDIES 477 V ∕3.6 =−0.0027 ×Δp + 0.9 ×Δp + 11 (3) mjt jt jt Δp = p − p jt n cjt (4) lnR = 5.3259 − 0.0249Δp − 0.0161 (5) mjt jt jt mg B = b ( ) ≈ 1.6b (6) jt st st cjt x −5 2 jt (7) b =−4.4 × 10 Δp + 0.01Δp + 0.03 − 0.01 + 0.15V + 1 st jt jt jt jt Δp jt x = 0.6 × 1 − (8) jt Model (2) is the wind field model, and Models (3)–(8) are the estimated models of corre - sponding parameters in the wind field model. In Model (2), we calculate V , which represents ijt the wind speed of the target area i at time t under the influence of typhoon j. V is the mjt maximum sustained wind speed (km/h) in the zone of influence of the typhoon. T is the ijt clockwise angle between the forward path of the typhoon and a radial line from the typhoon centre to the point of interest. V is the forward velocity of the typhoon centre. R represents hjt mjt the radius of maximum wind (). R is the radial distance from the centre of the typhoon to ijt the point of intere (km). B is the typhoon pressure distribution coefficient. Other parameters jt are based on previous research: gust factor G = 1.5, friction factor F = 0.9, and asymmetry factor S = 1 (Elliott et al., 2015). We use Models (3) and (4) to calculate V . ∆p represents the pressure drop due to the mjt jt typhoon centre (hPa). p represents the atmospheric pressure constant at 1015 hPa. p is n cjt the centre pressure of typhoon j at time t. Model (5) is used to calculate R , where ψ . rep- mjt jt resents the latitude (° N) of the center of typhoon j at time t. Models (6)–(8) are used to cal- culate the parameter B , according to the method of Holland (2008). jt We employ the index, f, proposed by Emanuel (2011) that proxies the fraction of property damaged, after calculating the wind speed in the target area with the above models, we use the damage function to convert the wind speed to the damage value. The models are as follows: nijt f = . (9) ijt 1 + v nijt MAX v − v ,0 ijt thresh v = (10) nijt v − v half thresh Emanuel (2011) argues that only when the wind speed exceeds a certain critical value will it cause substantial damage. Following Emanuel (2011), we use a value of 92.6 km/h for 478 L. SIYI ET AL. Figure 1. d escriptive statistics of typhoon damage. note: Created by authors from data available in the regional Specialized Meteorological Centre (RSMC) Best t rack data. v , and a value of 277.8 km/h for v , which is the threshold at which half of the property thresh half is damaged. As the typhoon-related losses increase exponentially in parallel with the wind speed, v is used to capture this change, through which the damage value f of typhoon nijt ijt j at time t in the target area i can be obtained from Model (9). Based on the Baidu Map, we extract the central locations of China’s provinces and pre- fecture-level cities as the target area i, and add the total typhoon destructive power each year in the target area i, thus obtaining the annual observations of typhoon damage degree index at the province and city levels, respectively (TYPHOON_PROV, TYPHOON_CITY). 3.4. Measures on tax avoidance degree and other variables The degree of tax avoidance in a company represents a reduction in income tax compared with pre-tax profit, thus capturing the ultimate consequence of various corporate tax avoid - ance strategies (Hanlon & Heitzman, 2010). Considering that some companies may enjoy tax incentives and that their scope differs, following Chen et al. (2015), we use the difference between the effective tax rate (income tax expense divided by pre-tax profit) and applicable statutory tax rate (DETR) to assess the degree of corporate tax avoidance. A less negative difference indicates a lower degree of corporate tax planning and tax avoidance. Following previous studies of tax avoidance (Hanlon & Heitzman, 2010; Liu & Wu, 2014), we further control for the firm size (SIZE), book-to-market ratio (BM), profitability (ROA ), leverage (LEV), tangible asset intensity (TANGIBLE), intangible asset intensity (INTANGIBLE), foreign-related operation (FOREIGN), regional economic level (LCITYGDP), and regional mar- ketisation level (MARKET). The definitions of the variables are presented in Table 1. Baidu is the largest Search engine in China, and the Baidu Map offers a geographic search service including longitude and latitude similar to the g oogle Map service. CHINA JOURNAL OF ACCOUNTING STUDIES 479 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Number of Typhoons Number of Typhoons Cause Damage Figure 2. Monthly distribution statistics of typhoon (2008–2015). note: Created by authors from data available in the regional Specialized Meteorological Centre (RSMC) Best t rack data. 4. Empirical results 4.1. Descriptive statistics Figure 1 presents the descriptive statistics of typhoon damage. When the typhoon damage index in a certain province or city is greater than 0, this location is considered damaged by typhoons. The figures indicate that during the period 2008–2015, typhoon damage occurred on average in 48 cities and nine provinces each year, indicating that typhoon disasters are more common in the coastal areas of China. Typhoons were more frequent in 2012 and covered the largest area. During that year, 14 provinces and as many as 92 cities were affected by the devastating effect of typhoons. Nevertheless, not all typhoons affect China. By com- paring the annual total number of typhoons and the actual number of typhoon victims, we find that there are on average 23 typhoons a year, while only about five typhoons with destructive power affect China each year. Figure 2 shows the monthly distribution of typhoon damage in the 2008–2015 period. Typhoons mainly occurred between June and October. According to Emanuel (2011), typhoons exceeding the damage threshold and causing actual damage are distributed between June and November (Typhoon_City is greater than 0). Current tax law requires corporate income tax to be paid in advance on a monthly or quarterly basis and the settlement must be made within 5 months of the end of the tax year. As typhoon damage is concentrated from June to November, for the companies’ actual tax planning and the tax bureau’s supervision, the effects of typhoons on tax are concentrated in one year. Therefore, selecting typhoon disasters as the research object ensures the tem- poral correspondence between typhoon damage and corporate tax avoidance to avoid cross-year problems. Table 2 shows the descriptive statistics of the variables. t he figures in our paper (f igure 1 to f igure 4) are drawn from the typhoon damage index calculated as described in Section 3.3. 480 L. SIYI ET AL. Table 2. d escriptive statistics. Variables Observations Mean STD 1/4 Median 3/4 DETR 6999 –0.03 0.20 –0.07 –0.01 0.02 TYPHOON_CITY 6999 0.02 0.05 0.00 0.00 0.00 TYPHOON_PROV 6999 0.01 0.04 0.00 0.00 0.00 SIZE 6999 21.86 1.33 20.96 21.68 22.53 ROA 6999 0.07 0.09 0.03 0.06 0.10 LEV 6999 0.45 0.27 0.26 0.44 0.60 BM 6999 0.51 0.25 0.31 0.48 0.69 TANGIBLE 6999 0.25 0.20 0.10 0.21 0.36 INTANGIBLE 6999 0.05 0.07 0.01 0.03 0.06 FOREIGN 6999 0.57 0.49 0.00 1.00 1.00 LCITYGDP 6999 6.57 0.91 5.98 6.71 7.32 MARKET 6999 10.04 3.80 7.31 10.88 12.30 To eliminate the influence of extreme values, we winsorise all of the continuous variables at their bottom and top percentiles. The mean (median) of DETR is –0.03 (–0.01), indicating that, on average, the effective corporate tax rate is lower than the applicable statutory tax rate. Less than half of the effective corporate tax rate exceeds the applicable statutory tax rate, which suggests that tax avoidance is a relatively common phenomenon. The mean degree of typhoon damage at the province level is 0.01 and that at the city level is 0.02, while the 1/4 and 3/4 quantile values are both 0, indicating that the degree of typhoon damage is clearly a regional feature. The distribution of other control variables is similar to that of previous research (Chen et al., 2015). The average return on assets used to measure profit - ability is about 7.2%. The average book-to-market ratio is 51%, indicating that the average debt level of the sample does not exceed 50%. The mean of book-to-market ratio is 0.51, in line with China’s overall market environment, in which the market value of net assets is the double of the book value. The average proportion of tangible assets is 25% and that of intangible assets is 5%. In addition, the mean of foreign-related operation is 0.518%, indi- cating that 57% of the sample has foreign businesses. 4.2. Correlation analysis Table 3 reports the Person correlation coefficient matrix. Results reveal a significant positive correlation between the degree of typhoon damage at the province and city levels, and highlight the difference between an effective tax rate and an applicable statutory tax rate. Preliminary results show that the higher the degree of typhoon damage, the lower the level of tax avoidance; thus, to some extent, the results demonstrate the obvious regional characteristics of the typhoon ee ff ct. The damage caused by typhoons differs in different regions of the same province. The correlation coefficient between the degree of typhoon damage at two different levels is 0.624, well below 1 but significant at the 1% level, indicating that the difference in typhoon damage levels captures this variation. Therefore, paying attention to the effect of typhoons on tax avoidance at dif- ferent levels is meaningful. In addition, there is a positive correlation between the degree of tax avoidance and some control variables, including firm size, prot fi ability, book-to-market ratio, foreign-related operation, regional economic level, and region marketisation level, while leverage, tangible asset intensity, and intangible asset intensity are negatively corre- lated with the degree of tax avoidance. The correlation coefficients between control variables CHINA JOURNAL OF ACCOUNTING STUDIES 481 Table 3. Correlation analysis. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (1 )DETR 1 (2 )TYPHOON_CITY 0.024** 1 (3) TYPHOON_PROV 0.026** 0.624*** 1 (4) SIZE 0.033*** –0.036*** –0.025** 1 (5) ROA 0.090*** 0.001 –0.004 0.031** 1 (6) LEV –0.061*** 0.024** –0.006 0.285*** –0.130*** 1 (7) BM 0.068*** 0.024** 0.0100 0.590*** –0.200*** 0.263*** 1 (8) TANGIBLE –0.058*** –0.026** –0.040*** 0.070*** 0.132*** –0.037*** 0.089*** 1 (9) INTANGIBLE –0.011 –0.005 –0.006 –0.038*** 0.146*** –0.036*** –0.047*** 0.217*** 1 (10) FOREIGN 0.026** 0.001 –0.008 –0.054*** –0.047*** –0.098*** –0.053*** 0.098*** 0.0140 1 (11) LCITYGDP –0.013 –0.103*** –0.081*** 0.152*** –0.011 –0.008 –0.038*** –0.160*** –0.088*** –0.105*** 1 (12) MARKET 0.044*** –0.156*** –0.154*** 0.064*** 0.001 –0.112*** –0.113*** –0.121*** –0.023* 0.079*** 0.364*** 1 notes: t able 3 reports the Person Correlation Coefficient among main variables. *, **and *** indicate statistical significance at the 10, 5, and 1% levels, respectively. 482 L. SIYI ET AL. are not large, indicating that there is no serious multicollinearity concern; thus, the variables can be included in the same model. 4.3. Main regression results Table 4 shows the regression results based on the coastal sample. In columns (1) and (2), we use the typhoon damage index (TYPHOON-CITY) at the city level as the independent variable. Results show that the coefficients of typhoon damage index at the city level are significantly positive at the 5% level after controlling for the year-, industry-, and urban-fixed effects. Columns (3) and (4) present the regression results of the province-level typhoon damage index (TYPHOON-PROV) as the independent variable, and the results are not significant. By comparing test results at different levels, we conclude that the typhoon effect generally has clear regional characteristics and physically damages only coastal areas. As previously mentioned, municipal governments are responsible for respond- ing to natural disasters rather than the entire province, which in turn affects enterprises within their tax jurisdiction. Therefore, we only test typhoon damage at the city level in the following analysis. The above results support hypothesis H1, which suggests that the higher the degree of typhoon damage, the lower the extent of tax avoidance of local companies. Table 4. t yphoon damage and tax avoidance. City-level Province-level (1) (2) (3) (4) DETR DETR DETR DETR TYPHOON_CITY 0.102* 0.110** (0.057) (0.043) TYPHOON_PROV 0.004 0.013 (0.968) (0.896) SIZE –0.004 –0.002 (0.453) (0.723) ROA 0.251*** 0.243*** (0.000) (0.000) LEV –0.059*** –0.054*** (0.000) (0.000) BM 0.118*** 0.104*** (0.001) (0.004) TANGIBLE –0.058*** –0.039* (0.006) (0.070) INTANGIBLE –0.028 –0.016 (0.353) (0.704) FOREIGN 0.012* 0.006 (0.076) (0.235) LCITY_GDP –0.003 –0.036 (0.942) (0.624) MARKET 0.002 0.001 (0.389) (0.708) CONS –0.241*** –0.136 –0.176*** –0.018 (0.000) (0.459) (0.001) (0.955) YEAR FE Yes Yes Yes Yes INDUSTRY FE Yes Yes Yes Yes CITY FE Yes Yes Yes Yes N 6,999 6,999 6,999 6,999 R 0.032 0.056 0.041 0.062 notes: t he p-value, reported in parentheses, are clustered at the firm level. *, ** and *** indicate statistical significance at the 10, 5, and 1% levels, respectively. t he same as in the following tables. CHINA JOURNAL OF ACCOUNTING STUDIES 483 For other control variables, the debt-to-asset ratio is significantly negatively correlated with DETR, indicating higher levels of tax avoidance for highly leveraged companies and supporting the expectation that interest rates have a tax shield function. Fixed asset intensity is significantly negatively correlated with DETR, indicating that companies with a higher fixed asset intensity have a higher tax avoidance degree, and the depreciation of fixed assets reduces the tax burden. The above results are consistent with the findings of Chen, Chen, and Dong (2016) and Chen, Kong, and Wang (2016). In addition, there is a significant positive correlation between profitability and tax avoidance, with a significant negative correlation between the book-to-market ratio and tax avoidance. Foreign operation is significantly neg - atively correlated with the tax avoidance level, which is inconsistent with the expectation of tax avoidance by foreign businesses. The reason for this discrepancy may be that previous research accounts for the 2008 income tax system reform (Chen, Chen, & Dong, 2016; Chen, Kong, & Wang, 2016; Chen et al., 2015), while our research period starts in 2008 after the implementation of the reform. Compared with the previous income tax system, the post-re- form income tax system obviously changes the income tax treatment and offshore income tax credit. In addition, it strengthens the anti-avoidance provisions, which may weaken the motivations for tax avoidance in high profitability companies, thus reducing the use of for - eign businesses for tax avoidance. To further test whether there are differences in the tax avoidance behaviour of state- owned companies, political affiliates, and local leading companies in the aftermath of dis- asters, we examine the interaction variables between TYPHOON_CITY and enterprise characteristics. SOE is the dummy variable for the property right, which takes a value of 1 if a firm is a state-owned enterprise, and 0 otherwise. PC is the dummy variable for the political connection of companies, using the den fi ition of political connection in the current literature (Li & Xu, 2013; Luo & Wei, 2012). If the chairman or general manager of a company has held or holds a position in the local government, the National People’s Congress (NPC), or the Chinese People’s Political Consultative Conference (CPPCC), indicating that the company has a political connection, the value of PC equals 1, and 0 otherwise. FLAGSHIP is the dummy variable for companies in a leading position. Locally classie fi d enterprises are generally used as benchmarks. FLAGSHIP takes the value of 1 if the company is ranked in the top one third of all of the local companies, and 0 otherwise. Due to the natural political relations between state-owned enterprises and local governments, we mainly focus on the political connections in private enterprises during the test. Therefore, we only use the sample of private enterprises for this test. Table 5 shows the results based on state-owned enterprises, companies with political connections, and local leading companies. Results show that TYPHOON_CITY * SOE, TYPHOON_CITY * PC, and TYPHOON_CITY * FLAGSHIP are all significantly positive at the 5% or 10% level, indicating that typhoon damage in state-owned enterprises, companies with political connections, and local leading com- panies have a more significant effect on corporate tax avoidance. These results support Hypothesis 2, which suggests that these companies have stronger incentives at the time of the disaster to reduce tax avoidance to mitigate the fiscal pressure to support local govern- ments. In addition, in columns (2) and (6), the coefficients of SOE and FLAGSHIP are signifi- cantly negative at 5% and 1%, respectively. The coefficient of SOE is significantly negative, indicating that state-owned enterprises engage in more routine tax avoidance activities than their counterparts. However, when comparing the coefficients of SOE and 484 L. SIYI ET AL. Table 5. t ests for different political attributes. Local leading Property right Political connection companies (1) (2) (3) (4) (5) (6) DETR DETR DETR DETR DETR DETR TYPHOON_ 0.110** 0.056 0.072 –0.008 0.110** 0.043 CITY (0.04) (0.36) (0.19) (0.89) (0.04) (0.36) SOE –0.013** –0.015** (0.04) (0.02) TYPHOON_ 0.149* CITY*SOE (0.10) PC 0.011 0.008 (0.13) (0.23) TYPHOON_ 0.167** CITY*PC (0.02) FLAGSHIP –0.058*** –0.061*** (0.00) (0.00) TYPHOON_ 0.218* CITY* FLAGSHIP (0.07) SIZE –0.003 –0.003 0.002 0.002 –0.014*** –0.013*** (0.58) (0.59) (0.58) (0.57) (0.00) (0.01) ROA 0.246*** 0.245*** 0.207*** 0.207*** 0.136*** 0.136*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) LEV –0.057*** –0.057*** –0.054*** –0.054*** –0.045*** –0.045*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) BM 0.120*** 0.119*** 0.105*** 0.105*** 0.121*** 0.121*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) TANGIBLE –0.056*** –0.057*** –0.052** –0.052** –0.042** –0.043** (0.01) (0.01) (0.02) (0.02) (0.03) (0.03) INTANGIBLE –0.028 –0.028 –0.046 –0.048 –0.008 –0.007 (0.36) (0.35) (0.45) (0.43) (0.80) (0.81) FOREIGN 0.013* 0.013* 0.015** 0.015** 0.012* 0.012* (0.06) (0.06) (0.02) (0.02) (0.08) (0.08) LCITY_GDP –0.005 –0.002 0.031 0.030 –0.002 –0.002 (0.92) (0.96) (0.66) (0.67) (0.96) (0.97) MARKET 0.002 0.002 0.001 0.001 0.002 0.002 (0.38) (0.39) (0.63) (0.64) (0.38) (0.38) CONS –0.157 –0.166 –0.364 –0.360 0.035 0.034 (0.38) (0.36) (0.19) (0.19) (0.84) (0.85) YEAR FE Yes Yes Yes Yes Yes Yes INDUSTRY Yes Yes Yes Yes Yes Yes FE CITY FE Yes Yes Yes Yes Yes Yes N 6,999 6,999 4,410 4,410 6,999 6,999 R 0.057 0.057 0.068 0.068 0.067 0.068 TYPHOON_CITY * SOE, the sum of the coefficients is 0.134 (–0.015 + 0.149), which is signifi- cantly positive and indicates that state-owned enterprises enjoy more favourable policies and government shelters in their daily management activities, in addition to more tax avoid- ance. When natural disasters cause fiscal pressure, state-owned enterprises bear a heavier tax burden than non-state-owned enterprises, and take the initiative to further reduce tax avoidance. Similarly, the coefficient of FLAGSHIP is significantly negative, indicating that local leading companies receive more government support for taxation; thus, the degree of tax avoidance is higher. By comparing the sum of the coefficients of FLAGSHIP and CHINA JOURNAL OF ACCOUNTING STUDIES 485 TYPHOON_CITY * FLAGSHIP, we also find that the sum of the coefficients is significantly pos- itive and equals 0.157 (–0.061 + 0.218). Although leading companies receive more govern- ment support in their daily management activities, when a typhoon occurs, their actual tax burden is also higher compared with other local non-leading companies. Therefore, for leading companies, reducing tax avoidance to support local governments during fiscal crises is more important. To sum up, for enterprises with special political attributes, although tax support can be obtained in daily operations for various reasons, they also take the initiative to assume more political tasks when events such as disasters happen, and reduce their tax avoidance behaviour to meet the government fiscal needs. 5. Additional analysis and robustness tests 5.1. Path testing: typhoon damage and local fiscal The basic logical premise of the paper is that typhoon disasters increase the pressure on local fiscal and subsequently affect the tax behaviour of enterprises that year. As a result, we first test the intermediary effect. The costs of disasters include not only direct costs, such as disaster relief and resettlement, but also indirect costs, such as subsequent reconstruction. In the current public fiscal data, there is no direct information on disaster relief and reset - tlement. Therefore, we use the fiscal deficit rate and the current year’s government invest - ment level as a test indicator. In addition, we use Baron and Kenny’s (1986) step-by-step method to test the intermediary effect. When constructing the intermediary effect test model, we first examine the effect of typhoon damage on the fiscal deficit and the level of local government investment. If we find that typhoon damage affects local fiscal variables, we add the typhoon destruction variables and local fiscal variables to the tax avoidance model at the same time. If the coefficients of the typhoon damage variables and local fiscal variables are significant, the existence of the intermediary effect is confirmed. In the model, DEFICT represents the fiscal deficit rate, which is the difference between fiscal expenditure and fiscal revenue divided by the current year’s GDP. GOVINV represents the local government investment level, measured by per capita government investment. According to current research (Chen et al., 2015; Elliott et al., 2015), the model also controls the city’s annual GDP level (LCITY_GDP), population level (CITYPOP), economic growth (GDP_ GROW), per capita fiscal revenue (REV_POP), and per capita fiscal expenditure (EXP_POP). The results of Panel A in Table 6 show that at the macroscopic level of prefecture-level cities, the typhoon’s destructive power is significantly and positively correlated with the current fiscal deficit rate and government investment level at 10% and 5%, respectively. The above results show that typhoon disasters have a significant effect on local fiscal expenditure, which is consistent with Elliott et al. (2015). Furthermore, as shown in column (1) in Panel B of Table 6, both typhoon damage and the fiscal deficit rate are significantly and positively correlated with tax avoidance at the 10% level, which indicates that the typhoon’s destructive power constitutes an incomplete intermediary effect through fiscal deficit. Similarly, column (2) shows that both the typhoon’s destructive power and the level of local investment are significant and positive at the 5% and 10% levels, respectively. These results also constitute an incomplete intermediary effect. To sum up, the results in Table 6 verify the logical premise of this paper, which suggests that the local fiscal pressure is an intermediary mechanism between typhoon disasters and tax avoidance behaviour. 486 L. SIYI ET AL. Table 6. t yphoon damage, local finance and tax avoidance. Panel A: Fiscal deficit Government investment (1) (2) DEFICIT GOVINV TYPHOON_CITY 9.139* 14.156** (0.089) (0.025) LCITY_GDP –16.747*** –14.386*** (0.000) (0.007) CITYPOP 4.104 –47.042*** (0.661) (0.000) GDP_GROW 14.579 –15.374 (0.151) (0.197) INC_POP –12.545*** 2.464*** (0.000) (0.001) PAY_POP 10.903*** 1.588** (0.000) (0.015) CONS 61.834 422.583*** (0.402) (0.000) YEAR FE Yes Yes CITY FE Yes Yes N 341 341 R 0.966 0.931 Panel B: (1) (2) detr detr TYPHOON_CITY 0.138* 0.119** (0.06) (0.02) DEFICIT 0.001* (0.07) GOVINV 0.001* (0.07) SIZE –0.004 –0.004 (0.47) (0.47) ROA 0.253*** 0.248*** (0.00) (0.00) LEV –0.059*** –0.056*** (0.00) (0.00) BM 0.118*** 0.117*** (0.00) (0.00) TANGIBLE –0.059*** –0.056*** (0.00) (0.01) INTANGIBLE –0.030 –0.046 (0.33) (0.21) FOREIGN 0.012* 0.013* (0.09) (0.06) LCITY_GDP 0.003 –0.041 (0.95) (0.55) MARKET 0.001 0.002 (0.65) (0.58) CONSTANT –0.194 –0.038 (0.34) (0.88) Year FE Yes Yes Ind. FE Yes Yes City FE Yes Yes N 6984 6571 R 0.059 0.056 CHINA JOURNAL OF ACCOUNTING STUDIES 487 5.2. Cross-sectional tests from multiple perspectives 5.2.1. Financial constraints perspective: fiscal decentralisation and fiscal surplus After a disaster, local governments may get fiscal support from higher-level governments including central government and provincial governments to share the fiscal pressure. The central and higher-level governments may help local governments cope with post-disaster relief and reconstruction through tax return, general transfer payment, or special transfer payment subsidies. If higher-level governments give more post-disaster subsidies, the fiscal pressure on local governments will be relatively modest. Therefore, this article tests whether this main hypothesis is based on the local fiscal pressure perspective. As the current public fiscal information does not disclose funds allocated by the central or higher-level govern- ments after a particular disaster, we choose to test the fiscal relationship between local governments and the central or higher-level governments based on fiscal decentralisation. In the current context of high concentration of fiscal power in China, the degree of fiscal decentralisation is one of the direct causes affecting the fiscal pressure on local governments. With a huge gap between local revenue and expenditure, the tax burden on enterprises increases, as it is difficult to reduce fiscal expenditure (Liu & Liu, 2014). Therefore, the degree of fiscal decentralisation is also directly related to corporate tax avoidance (Cao et al., 2009). Following natural disasters, a high degree of fiscal decentralisation indicates that the central or higher-level governments provide more fiscal allocations and subsidies to local govern- ments. As a result, local governments have more disposable income attributed to localities, and are better able to cope with the costs of post-disaster relief and reconstruction. In addi- tion, the fiscal pressure decreases. For local companies, the motivation to support their local governments by reducing tax avoidance is low. In addition, in parts of areas often affected by disasters, local governments may adopt budgetary provisions for disaster preparedness projects. However, earmarking part of the budget for a disaster response inevitably affects other fiscal expenditure items and, to a certain extent, restricts investment in other areas. Therefore, the extent to which local gov- ernments can build reserves for disaster response depends largely on the local finance sur - plus. The fiscal surplus reflects the difference between current fiscal revenue and expenditure. If the local fiscal surplus is high, it will show that local governments face less fiscal pressure; thus, excess fiscal funds can be used for disaster response. To sum up, the higher the degree of fiscal decentralisation and fiscal surplus, the lower the government fiscal pressure, and the smaller the effect of disasters on corporate tax avoidance. Following the above analysis, we test the hypothesis based on the fiscal decentralisation and fiscal surplus perspective. We measure fiscal decentralisation by DECENT, which is cal- culated as the per capita fiscal expenditure of prefecture-level cities as a percentage of the total per capita fiscal expenditure (Jia & Ying, 2016). SURPLUS represents the fiscal surplus and is measured by the difference between fiscal revenue and expenditure in prefecture-level cities divided by GDP. The higher the degree of fiscal decentralisation and fiscal surplus, the lower the local fiscal pressure. Table 7 shows that TYPHOON_CITY * DECENT and TYPHOON_CITY * SURPLUS are signifi- cantly negatively correlated with DETR at the 10% and 1% levels, respectively, indicating that the relationship between typhoon damage and corporate tax avoidance is weakened when the fiscal pressure is low, which supports our hypothesis. 488 L. SIYI ET AL. Table 7. f urther tests on fiscal decentralisation and fiscal pressure. Fiscal decentralisation Fiscal surplus (1) (2) (3) (4) DETR DETR DETR DETR TYPHOON_CITY 0.112** 0.879** 0.116* 0.194*** (0.04) (0.04) (0.05) (0.00) DECENT 0.163* 0.181* (0.09) (0.06) TYPHOON_CITY* DECENT –1.026* (0.08) SURPLUS 0.013 0.024 (0.48) (0.19) TYPHOON_CITY* SURPLUS –1.122*** (0.00) SIZE –0.004 –0.004 –0.004 –0.004 (0.19) (0.20) (0.47) (0.47) ROA 0.248*** 0.249*** 0.251*** 0.250*** (0.00) (0.00) (0.00) (0.00) LEV –0.060*** –0.060*** –0.059*** –0.059*** (0.00) (0.00) (0.00) (0.00) BM 0.118*** 0.118*** 0.118*** 0.117*** (0.00) (0.00) (0.00) (0.00) TANGIBLE –0.057*** –0.057*** –0.059*** –0.059*** (0.00) (0.00) (0.01) (0.01) INTANGIBLE –0.024 –0.023 –0.028 –0.028 (0.62) (0.63) (0.36) (0.36) FOREIGN 0.012** 0.012** 0.012* 0.012* (0.04) (0.04) (0.08) (0.08) LCITY_GDP –0.010 –0.017 –0.008 –0.017 (0.87) (0.77) (0.87) (0.73) MARKET 0.002 0.002 0.002 0.003 (0.35) (0.31) (0.37) (0.30) CONS –0.193 –0.177 –0.120 –0.088 (0.40) (0.44) (0.50) (0.60) YEAR FE Yes Yes Yes Yes INDUSTRY FE Yes Yes Yes Yes CITY FE Yes Yes Yes Yes N 6,999 6,999 6,984 6,984 R 0.056 0.057 0.057 0.057 5.2.2. Local institutional characteristic perspective: institutional environment and pressure on economic growth Enterprises are willing to sacrifice short-term tax revenue to help ease the post-disaster government fiscal pressure, build a good relationship with local governments, and obtain more resources in the future. The premise is that local governments have greater discretion over the allocation of resources. Differences in institutional levels between different regions may result in different capacities in resource allocation. In areas with higher marketisation levels, resource competition between enterprises is more equitable and less reliant on local governments, which reduces incentives for enterprises to obtain resource advantages by meeting the government fiscal targets (Luo & Yang, 2013). Therefore, we expect a smaller effect of typhoons on corporate tax avoidance in areas with higher marketisation levels. As mentioned earlier, typhoon disasters damage the local economy, affecting local eco - nomic growth. In the promotion competition, local officials’ urgent need for post-disaster reconstruction is closely related to the pressure of economic growth. Therefore, the more severe the effect of disasters on economic growth, the greater the fiscal pressure on local CHINA JOURNAL OF ACCOUNTING STUDIES 489 governments and the higher the demand for tax revenue. Therefore, local enterprises also have a stronger incentive to meet the government needs to maintain economic growth using fiscal means. Based on the above analysis, this paper further tests the relationship between typhoon damage and corporate tax avoidance based on the degree of marketisation and economic growth in different regions. MARKET represents the degree of marketisation. Following the previous literature (Liu & Wu, 2014), we use the ‘Fan Gang Market-oriented Index’ to measure it. We also use the province-level index from 2008 to 2015 to measure the degree of marke- tisation. The data for 2014 replace the missing data for 2015. The pressure of economic growth is measured by the absolute value of the decline in the current year’s economic growth, DGDPGROW. The higher the value, the greater the decline in the current year. If the annual growth rate is higher than the growth rate of the previous year, the value is 0. Table 8 presents the results. The results of column (2) show that TYPHOON_CITY * MARKET is significantly negative at the 10% level, indicating that the role of typhoons in corporate tax avoidance is weakened in areas with higher marketisation levels and better institutional protection. Column (4) shows that TYPHOON_CITY * DGDPGROW and DETR are significantly and positively correlated at the 5% level, indicating that the role of typhoons in corporate Table 8. f urther tests on marketisation level and economic growth pressure. Marketisation level Economic growth pressure (1) (2) (3) (4) DETR DETR DETR DETR TYPHOON_CITY 0.110** 0.353** 0.110** 0.129*** (0.043) (0.03) (0.04) (0.01) MARKET –0.004 0.002 0.002 0.003 (0.453) (0.35) (0.38) (0.27) TYPHOON_CITY*MARKET –0.030* (0.07) DGDPGROW 0.034 –0.012 (0.76) (0.90) TYPHOON_CITY*DGDPGROW 2.150** (0.04) SIZE 0.002 –0.004 –0.004 –0.004 (0.389) (0.46) (0.45) (0.45) ROA 0.251*** 0.251*** 0.251*** 0.251*** (0.000) (0.00) (0.00) (0.00) LEV –0.059*** –0.059*** –0.059*** –0.059*** (0.000) (0.00) (0.00) (0.00) BM 0.118*** 0.118*** 0.118*** 0.119*** (0.001) (0.00) (0.00) (0.00) TANGIBLE –0.058*** –0.058*** –0.058*** –0.058*** (0.006) (0.01) (0.01) (0.01) INTANGIBLE –0.028 –0.028 –0.028 –0.027 (0.353) (0.36) (0.35) (0.37) FOREIGN 0.012* 0.012* 0.012* 0.012* (0.076) (0.08) (0.08) (0.08) LCITY_GDP –0.003 –0.002 –0.000 –0.018 (0.942) (0.97) (1.00) (0.68) CONS –0.136 –0.149 –0.147 –0.084 (0.459) (0.41) (0.45) (0.62) YEAR FE Yes Yes Yes Yes INDUSTRY FE Yes Yes Yes Yes CITY FE Yes Yes Yes Yes N 6999 6999 6999 6999 R 0.056 0.056 0.056 0.056 490 L. SIYI ET AL. tax avoidance is strengthened in areas with higher economic growth pressure, and that the motivation of enterprises to actively meet government fiscal needs is also more evident. 5.2.3. Market pressure perspective: institutional holdings and analysts following For enterprises, reducing tax avoidance can ease the post-disaster fiscal pressure on local governments, help build a good relationship with local governments in the long run, and offer advantages in future resource competition. However, in the short term, the income tax expense directly affects the net profit and cash outflow in the current period, exerting per - formance pressure on enterprises. For companies, there may be a trade-o ff between building long-term relationships with local governments and meeting the target of short-term market performance when the pressure from capital markets is greater. Current research shows that institutional investors and analysts can play the role of external governance and prevent the behaviour of managers from deviating from shareholder interests (Bhojraj & Sengupta, 2003; Gillan & Starks, 2000). When institutional investors hold a higher percentage of shares and more analysts follow companies, the pressure on short-term corporate performance increases, which to a certain extent limits the managers’ ability to meet the government needs. Based on the above analysis, we further examine the role of institutional investors and analysts. We capture the role of institutional investors using the institutional holding ratio, which is measured by the proportion of fund holdings to the total number of shares (Cai & Rao, 2015). We use the analyst following to capture the role of analysts, which is measured by the natural logarithm of the number of analyst teams releasing a company analysis in the year. The results in Table 9 show that the coefficients of TYPHOON_CITY * INSTI and TYPHOON_ CITY * ANALYST are both negative at the 1% and 5% levels, respectively. This reveals that when the institutional holding ratio and the analyst following are higher, the short-term corporate performance is subject to greater pressure. Moreover, the motivation to meet government fiscal needs by reducing tax avoidance after the disaster is lower. 5.2.4. Corporate operating environment perspective: tax avoidance margin From the tax avoidance margin perspective, enterprises are likely to meet the government fiscal needs by reducing post-disaster tax avoidance, which is greatly affected by the current tax avoidance margin of enterprises. When the effective corporate tax rate is higher than the nominal tax rate, enterprises cannot make downward adjustment and can only pay more tax, which puts pressure on routine business operations. Therefore, the effect of typhoons on corporate tax avoidance mainly exists when the effective tax rate is lower than the nominal tax rate. We further examine the role of the tax avoidance margin. The sample is divided into two groups: a group in which the effective tax rate is higher than the nominal tax rate (DETR > 0) and a group in which the effective tax rate is lower than the nominal tax rate (DETR < 0). Table 10 shows the results. Columns (1) and (2) are the regression results of the first group (DETR > 0) and second group (DETR < 0). We find that typhoon damage is significant only when the effective tax rate is lower than the nominal tax rate. Therefore, the effect of typhoons on tax avoidance mainly exists when the effective tax rate is lower than the nominal tax rate. In short, enterprises with a tax avoidance margin voluntarily choose to reduce tax avoidance after the disaster. CHINA JOURNAL OF ACCOUNTING STUDIES 491 Table 9. f urther tests on capital market pressure. Institutional investors Analyst following (1) (2) (3) (4) DETR DETR DETR DETR TYPHOON_CITY 0.110** 0.183*** 0.110** 0.217** (0.05) (0.01) (0.04) (0.01) INSTI 0.002*** 0.003*** (0.00) (0.00) TYPHOON_CITY* INSTI –0.031*** (0.00) ANALYST 0.003 0.004* (0.22) (0.06) TYPHOON_CITY* –0.076** ANALYST (0.02) SIZE –0.006 –0.006 –0.006 –0.006 (0.26) (0.27) (0.33) (0.34) ROA 0.244*** 0.244*** 0.245*** 0.245*** (0.00) (0.00) (0.00) (0.00) LEV –0.057*** –0.057*** –0.057*** –0.057*** (0.00) (0.00) (0.00) (0.00) BM 0.131*** 0.131*** 0.124*** 0.124*** (0.00) (0.00) (0.00) (0.00) TANGIBLE –0.058*** –0.058*** –0.058*** –0.058*** (0.01) (0.01) (0.01) (0.01) INTANGIBLE –0.026 –0.026 –0.025 –0.025 (0.38) (0.39) (0.40) (0.42) FOREIGN 0.012* 0.012* 0.012* 0.012* (0.08) (0.07) (0.07) (0.08) LCITY_GDP 0.001 –0.002 –0.002 –0.003 (0.98) (0.96) (0.96) (0.94) MARKET 0.002 0.002 0.002 0.002 (0.37) (0.35) (0.39) (0.39) CONS –0.123 –0.112 –0.107 –0.109 (0.50) (0.54) (0.57) (0.57) YEAR FE Yes Yes Yes Yes INDUSTRY FE Yes Yes Yes Yes CITY FE Yes Yes Yes Yes N 6,999 6,999 6,999 6,999 R 0.057 0.057 0.056 0.057 5.3. The economic consequence test of tax avoidance: two-way interest exchange As mentioned earlier, enterprises sacrifice short-term taxes to support local governments and obtain longer-term economic benefits. In fact, when enterprises help local governments carry outpost-disaster rescue missions, governments generally give corresponding returns to enter- prises (Zhang et al., 2013). The incremental tax paid by enterprises after the disaster compen- sates the government fiscal gap to a certain extent, and contributes to post-disaster rescue and reconstruction while boosting the local economy. The higher the damage, the higher the taxes payable, and the greater the government support. For local governments, compensating corporate support helps establish a good government reputation and encourages more enterprises to cooperate with the governments in case of the future emergencies. Therefore, local governments use their administrative power to provide assistance to enterprises through government subsidies, government projects, bank loans, and other ways. As a result, the two-way relationship between local governments and enterprises is established. 492 L. SIYI ET AL. Table 10. f urther tests on tax avoidance margin. Tax avoidance margin (1) (2) DETR>0 DETR<0 TYPHOON_CITY 0.099 0.126** (0.13) (0.03) SIZE –0.024*** 0.013*** (0.00) (0.00) ROA –0.427*** 0.367*** (0.00) (0.00) LEV 0.160*** –0.088*** (0.00) (0.00) BM 0.046 –0.025 (0.12) (0.54) TANGIBLE 0.026 –0.045*** (0.13) (0.01) INTANGIBLE 0.116** –0.104* (0.04) (0.09) FOREIGN –0.001 0.017* (0.93) (0.07) LCITY_GDP –0.010 –0.101 (0.88) (0.12) MARKET 0.003 0.002 (0.19) (0.49) CONS 0.487 –0.075 (0.11) (0.76) YEAR FE Yes Yes INDUSTRY FE Yes Yes CITY FE Yes Yes N 2,929 4,068 R 0.180 0.118 To investigate the economic consequences of post-disaster tax avoidance and to discuss whether enterprises can get more benefits by paying more taxes to support local govern- ments, we use the government subsidies and credit resources for further tests. In addition to direct government subsidies, current research proves that the government administrative power can affect the distribution of credit resources (Zhang et al., 2010). Sub(t + 1) is the government subsidy obtained by the enterprise in period t + 1 divided by the profit before tax in period t (Chen et al., 2015). FIN(t + 1) is the debt financing obtained in the period t + 1 divided by the total assets at the end of the period t + 1. Table 11 shows the test results of the economic consequences. TYPHOON_CITY*DETR is significantly positive at the 1% and 10% levels in columns (3) and (6), indicating that the higher the typhoon’s destructive power, the higher the corporate tax burden, which can help enterprises obtain government subsidies and debt financing in the future. These results confirm the exchange of interests between enterprises and local governments. 5.4. Alternative explanations 5.4.1 Tightened government supervision or corporate initiative In the theoretical analysis of hypothesis H1, we mention that the effect of typhoon disasters on tax avoidance may stem from a tightened local tax supervision, yet it may also come from companies taking the initiative to help local governments. Although we analyse the possible drawbacks of strengthening tax regulation, we do not separate the two mechanisms in the CHINA JOURNAL OF ACCOUNTING STUDIES 493 Table 11. t he economic consequence of post-disaster tax avoidance. Government subsidy Debt financing (1) (2) (3) (4) (5) (6) SUB SUB SUB FIN FIN FIN t+1 t+1 t+1 t+1 t+1 t+1 TYPHOON_CITY –0.001 –0.001 –0.009 –0.012 (0.11) (0.11) (0.71) (0.61) DETR 0.001** 0.000 0.024** 0.018* (0.03) (0.42) (0.04) (0.09) TYPHOON_CITY*DETR 0.016*** 0.373* (0.01) (0.07) SIZE –0.001*** –0.001*** –0.001*** 0.015*** 0.015*** 0.015*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) ROA –0.003*** –0.004*** –0.004*** –0.037 –0.042 –0.042 (0.00) (0.00) (0.00) (0.19) (0.14) (0.14) LEV 0.000 0.000 0.000 0.131*** 0.132*** 0.132*** (0.20) (0.18) (0.17) (0.00) (0.00) (0.00) BM 0.001** 0.001** 0.001** 0.088*** 0.085*** 0.086*** (0.01) (0.01) (0.01) (0.00) (0.00) (0.00) TANGIBLE –0.000 0.000 0.000 0.108*** 0.109*** 0.109*** (0.96) (0.96) (0.95) (0.00) (0.00) (0.00) INTANGIBLE 0.001** 0.001** 0.001** 0.107** 0.107** 0.108** (0.05) (0.04) (0.04) (0.02) (0.02) (0.02) FOREIGN –0.000** –0.000** –0.000** 0.022*** 0.021*** 0.021*** (0.05) (0.04) (0.04) (0.00) (0.00) (0.00) LCITY_GDP –0.000 –0.000 –0.000 –0.008 –0.009 –0.008 (0.97) (0.92) (0.96) (0.86) (0.85) (0.86) MARKET 0.000* 0.000* 0.000* 0.000 0.000 0.000 (0.10) (0.10) (0.09) (0.75) (0.78) (0.77) CONS 0.011** 0.011** 0.011** –0.324 –0.319 –0.320 (0.02) (0.02) (0.02) (0.11) (0.12) (0.11) YEAR FE Yes Yes Yes Yes Yes Yes INDUSTRY FE Yes Yes Yes Yes Yes Yes CITY FE Yes Yes Yes Yes Yes Yes N 5,667 5,667 5,667 5,885 5,885 5,885 R 0.125 0.126 0.129 0.298 0.299 0.300 494 L. SIYI ET AL. main test. Therefore, we continue to compare these two possibilities by separating the two mechanisms using two methods. First, we distinguish between companies with different property rights. Following Chen et al. (2015), we distinguish whether typhoon damage affects tax avoidance in companies with different property rights based on the test of hypothesis H2. If local governments increase tax regulation to seek more tax revenue, and imposing differential tax collection and administration on different companies proves difficult, the effect of typhoon disasters on tax avoidance will exist in companies with different property rights. However, if the ini- tiative to meet government needs comes from companies because the managers of state- owned enterprises have a significant incentive to meet these needs, this phenomenon will only exist in state-owned enterprises. More specifically, as local state-owned enterprises are managed by local and not central governments, the expected results will exist only for local state-owned enterprises. In Table 12(a), we distinguish three types of enterprises with dif- ferent property rights. Results show that the effect of typhoon damage on corporate tax avoidance exists only in local state-owned enterprises, and not in central- and non-state-owned enterprises, which suggests that the effect of disasters on tax avoidance is mainly due to state-owned enter - prises’ initiative to meet government needs rather than the unified adjustment of tax regu- latory policies by local governments. Second, based on Baron and Kenny’s (1986) intermediary effect test, we test whether typhoon damage affects local tax regulation and we analyse whether local tax regulation is Table 12. (a) t est with different property rights. Local state-owned enter- Central state-owned enter- Non-state-owned enter- prises prises prises (1) (2) (3) detr detr detr TYPHOON_CITY 0.231** 0.090 0.071 (0.02) (0.45) (0.19) SIZE 0.003 –0.015** 0.003 (0.76) (0.05) (0.46) ROA 0.264*** 0.338*** 0.207*** (0.00) (0.00) (0.00) LEV –0.020 –0.072 –0.055*** (0.34) (0.14) (0.00) BM 0.086* 0.092 0.105*** (0.08) (0.11) (0.00) TANGIBLE –0.023 –0.079 –0.052** (0.44) (0.17) (0.02) INTANGIBLE –0.049 0.093 –0.047 (0.33) (0.62) (0.45) FOREIGN –0.010 0.023** 0.014** (0.23) (0.02) (0.03) LCITY_GDP –0.062 0.151 0.027 (0.59) (0.25) (0.70) MARKET 0.003 –0.001 0.001 (0.61) (0.94) (0.62) CONS 0.117 –0.689 –0.357 (0.89) (0.46) (0.19) YEAR FE Yes Yes Yes INDUSTRY FE Yes Yes Yes CITY FE Yes Yes Yes N 1,818 768 4,413 R 0.096 0.121 0.067 CHINA JOURNAL OF ACCOUNTING STUDIES 495 the intermediary factor in the relationship between typhoon damage and tax avoidance behav- iour. If we fail to find the intermediary effect, then this alternative explanation will be excluded. We calculate the tax supervision intensity TE following previous studies (Ye & Liu, 2011; Zeng & Zhang, 2009). Panel A in Table 12(b) shows that the degree of typhoon damage is not signif- icantly correlated with the tax supervision intensity. Meanwhile, the results in Panel B show that the tax supervision has no significant effect on the degree of tax avoidance. We conclude that typhoon damage does not significantly affect the tax supervision inten - sity, which excludes this alternative explanation. 5.4.2. The effect of tax rebates and government subsidies on tax avoidance motivation With the occurrence of natural disasters, affected enterprises may benefit from preferential taxation and fiscal policies, such as deferred payment, loss reduction and exemption, and fiscal subsidies. These measures may affect the corporate tax evasion incentive and tax avoid - ance channels to some extent, which in turn affect the degree of corporate tax avoidance. For example, enterprises benefiting from tax rebates and other fiscal subsidies may have lower tax evasion motivation. At the same time, tax-returnable fiscal subsidies may be directly related to tax payments and other fiscal subsidies. The mechanism may explain the phe - nomenon discovered in this article to some extent. For this reason, additional tests on this alternative explanation are needed, which we test in two ways. First, as we know, enterprises receiving more tax returns and fiscal subsidies are typically more severely affected during natural disasters. However, the initial logic of this paper focuses on enterprises’ incentive to ‘help’ local governments. Therefore, our hypothesis will be sup- ported to a certain extent if we can prove that the effect found in this paper is more evident in less affected enterprises. (We assume that enterprises located in the same area suffer different degrees of disaster.) For this reason, we divide enterprises in the sample into light- and heavy-asset enterprises to distinguish their disaster degree. In general, the damage caused by natural disasters mainly affects enterprises that own more factories and machinery, as natural disasters damage equipment and interrupt production. Therefore, if this alternative explanation is reasonable, we will find that heavily funded enterprises suffering more seri- ously from disasters expect to obtain more local tax rebates or government subsidies; thus, their tax avoidance motivation will be even more pronounced and more sensitive to natural disasters. If this alternative hypothesis is true, the results will be more evident in light-asset firms. The test results are presented in Table 12(a). In the table, LIGHT is a dummy variable for light-asset industries based on the SFC industry classification of 2012. LIGHT equals 1 if the firm belongs to the industry that does not involve manufacturing plants, machinery, and equipment, and 0 otherwise. Table 13(a) shows the results of this alternative explanation. Results reveal that TYPHOON_ CITY * LIGHT, a crossover item between the dummy variables of typhoon destructiveness and light-asset industry, is significantly positive at 5% compared with heavy-asset enterprises, indicating that light-asset industries have a stronger desire to reduce tax avoidance after the disaster to relieve the government fiscal pressure. This finding contradicts the expectation of the alternative hypothesis. Therefore, it is more likely to support our main hypothesis. Second, we use Baron and Kenny’s (1986) mediation effect test. We first test whether typhoon disasters affect the tax return subsidy policy received by companies in the current year or the overall level of government subsidies in the current year. Then, we examine 496 L. SIYI ET AL. Table 12. (b) Mediation effect test based on tax supervision intensity. Panel A: TE TYPHOON_CITY –0.123 (0.222) LCITY_GDP 0.077 (0.327) CITYPOP –0.357** (0.015) GDP_GROW 0.030 (0.856) INC_POP 0.005 (0.712) PAY_POP 0.013 (0.242) CONS 3.270*** (0.006) YEAR FE Yes CITY FE Yes N 283 R 0.962 Panel B: detr TYPHOON_CITY 0.095* (0.06) TE –0.083 (0.14) SIZE –0.004 (0.36) ROA 0.224*** (0.00) LEV –0.055*** (0.00) BM 0.103*** (0.00) TANGIBLE –0.060** (0.01) INTANGIBLE –0.006 (0.90) FOREIGN 0.009 (0.20) LCITY_GDP 0.017 (0.80) MARKET 0.003 (0.37) CONSTANT –0.022 (0.91) YEAR FE Yes INDUSTRY FE Yes CITY FE Yes N 5,567 R 0.058 whether tax rebates or government subsidies show the intermediary effect of disasters on corporate avoidance. If we fail to find that these two factors show the intermediary effect (affecting tax avoidance behaviour), then we will exclude this alternative explanation. In the test model shown below, TAXRE is a dummy variable for tax return, which takes the value of 1 if enterprises obtain the government subsidy of tax return, and 0 otherwise. SUB is calcu- lated as the difference between government subsidy and tax return divided by pre-tax profit. CHINA JOURNAL OF ACCOUNTING STUDIES 497 Table 13. (a) f urther test on the companies of different asset types. (1) (2) DETR DETR TYPHOON_CITY 0.110** 0.087 (0.04) (0.12) LIGHT 0.005 0.002 (0.59) (0.85) TYPHOON_CITY* LIGHT 0.160** (0.03) SIZE –0.004 –0.004 (0.45) (0.46) ROA 0.251*** 0.251*** (0.00) (0.00) LEV –0.058*** –0.058*** (0.00) (0.00) BM 0.119*** 0.119*** (0.00) (0.00) TANGIBLE –0.058*** –0.058*** (0.01) (0.01) INTANGIBLE –0.028 –0.028 (0.35) (0.36) FOREIGN 0.013* 0.013* (0.08) (0.08) LCITY_GDP –0.004 –0.003 (0.93) (0.94) MARKET 0.002 0.002 (0.39) (0.39) CONS –0.133 –0.136 (0.47) (0.46) YEAR FE Yes Yes INDUSTRY FE Yes Yes CITY FE Yes Yes N 6,999 6,999 R 0.057 0.057 Table 13(b) shows the results of these alternative interpretations based on the mediating effect. The results of columns (1) and (2) in Table 13(b) demonstrate that the level of typhoon destructive power has no significant effect on corporate tax returns and government subsi- dies in the current year. The results of columns (3) and (4) also indicate that the effect of tax subsidies and government subsidies on the level of tax avoidance is insignificant. The above results show that tax rebates and government subsidies are not intermediary mechanisms between typhoon damage and corporate tax avoidance. To sum up, enterprises do not reduce their motivation for tax avoidance because of tax rebates and government subsidies in the year when the disaster occurs. We thus exclude this possible alternative explanation. 5.5. Robustness test 5.5.1. Placebo test The above results show that the tax avoidance behaviour of coastal city enterprises increases when the severity of typhoon destruction decreases. However, is this a unique phenomenon in typhoon-affected areas? In other words, if typhoon disasters occur in other cities, will we still find that the typhoon’s destructive power and the level of corporate tax avoidance are linked? If not, this would clearly indicate that typhoons are a major factor in reducing the tax rate of local enterprises. Moreover, it would mean that the reduction of corporate tax 498 L. SIYI ET AL. Table 13. (b) Mediation effect tests on tax rebates and government subsidies. Tax return mechanism Government subsidy mechanism (1) (2) (3) (4) TAXRE DETR SUB DETR TYPHOON_CITY 1.442 0.111** 0.227 0.109** (0.43) (0.04) (0.82) (0.04) TAXRE –0.017 (0.10) SUB 0.012 (0.14) SIZE –0.016 –0.004 0.321** –0.003 (0.84) (0.45) (0.05) (0.49) ROA 1.399* 0.252*** 1.219 0.255*** (0.06) (0.00) (0.36) (0.00) LEV –0.344 –0.059*** –1.073*** –0.060*** (0.34) (0.00) (0.00) (0.00) BM 1.120*** 0.119*** 0.494 0.117*** (0.00) (0.00) (0.28) (0.00) TANGIBLE –0.348 –0.058*** –0.531 –0.062*** (0.53) (0.01) (0.51) (0.01) INTANGIBLE –0.480 –0.028 –0.357 –0.021 (0.70) (0.34) (0.78) (0.51) FOREIGN –0.043 0.012* 0.948*** 0.012* (0.83) (0.08) (0.00) (0.08) LCITY_GDP –0.911 –0.004 0.058 –0.004 (0.68) (0.93) (0.95) (0.94) MARKET 0.115* 0.002 0.102* 0.002 (0.05) (0.38) (0.06) (0.41) CONS –28.195*** –0.134 –5.017 –0.143 (0.00) (0.46) (0.19) (0.44) Year FE Yes Yes Yes Yes Ind. FE Yes Yes Yes Yes City FE Yes Yes Yes Yes N 6,999 6,999 6,999 6,999 R 0.195 0.056 0.334 0.061 0.3 0.25 0.2 0.15 0.1 0.05 TYPHOON-CITY TYPHOON-PLACEBO Figure 3. t he distribution of actual typhoon destruction index and placebo typhoon destructive index in 2008. note: Created by authors from data available in the regional Specialized Meteorological Centre (RSMC) Best t rack data. SanYa ZhanJiang FangChengGang QinZhou ZhongShan JiangMen DongGuan GuangZhou JieYang XiaMen QuanZhou FuZhou WenZhou NingBo HangZhou ShangHai WuXi NanTong YanCheng RiZhao WeiFang DongYing YanTai DaLian TangShan DanDong HuLuDao PanJin CHINA JOURNAL OF ACCOUNTING STUDIES 499 Figure 4. t he distribution of t-values of t YPhoon-Pla CeBo in placebo test of 100 and 500 times. avoidance might also be found in non-disaster areas. Hence, a typhoon’s destructive power may not cause the change in corporate tax avoidance behaviour. Therefore, we conduct a placebo test by (1) randomly assigning the actual typhoon destructive power on each city and (2) regressing the random typhoon destruction index TYPHOON-CITY(Random) on the corresponding city’s corporate tax avoidance index 100 and 500 times. We take 2008 as an example. Figure 3 shows the differences between the actual destruction index of coastal cities and the index in the placebo test. Actual typhoon destruction is concentrated in southern coastal cities, while the placebo typhoon destruction index is randomly distributed. Figure 4 shows that the proportion of significantly positive and negative indexes is small, which means that the virtual processing effect does not exist. It also suggests that the typhoon’s destructive power in the city is the factor that causes the decrease in corporate tax avoidance rather than other factors or noise. This test verifies the robustness of the main conclusions of the paper. 5.5.2. Endogeneity test Method 1: instrumental variable method. Although the effect of typhoon disasters is relatively exogenous, the regression still shows latent endogenous problems. For different reasons, such as topography distribution and climate characteristics, Chinese coastal regions face different probabilities of being affected by typhoons. Aec ff ted regions have historically suf- fered from previous typhoon disasters. As a result, these areas may be ‘path dependent’ on tax regulation. Regional governments habitually reinforce tax collection, which leads to higher effective tax rates in local enterprises. However, this regional government character - istic is difficult to quantify, resulting in endogenous problems caused by missing variables. To alleviate the possibility of endogenous problems, we adopt the two-stage regression method of instrumental variables. The instrumental variable (IV) used in this paper is the latitude of the city where the listed company is located (°N), as typhoons often come from tropical oceans (equator) with temperatures above 26°C. Formed by the combination of internal force, the rotation of the earth, and the air motion of the subtropical high air pressure, typhoons travel northwest to finally reach the mainland and gradually weaken until they dissipate. Therefore, typhoons more frequently affect low latitudes, and the economic losses caused by typhoons are often more severe. At the same time, the latitude of the city where the enterprise is located is a physical variable that objectively exists and does not directly affect corporate tax avoidance behaviour. As a result, it is a relatively ideal variable. The 500 L. SIYI ET AL. Table 14. endogeneity test. First stage Second stage Exclusive test (1) (2) (3) TYPHOON_CITY DETR DETR LAT –0.003*** –0.000 (0.00) (0.62) INSTRU_TYPHOON_CITY 0.268* (0.08) TYPHOON_CITY 0.119** (0.03) SIZE 0.000 –0.004 –0.005* (0.98) (0.44) (0.08) BM –0.001 0.124*** 0.127*** (0.77) (0.00) (0.00) ROA 0.003 0.266*** 0.267*** (0.42) (0.00) (0.00) LEV 0.003** –0.057*** –0.056*** (0.02) (0.00) (0.00) TANGIBLE 0.000 –0.068*** –0.065*** (0.85) (0.00) (0.00) INTANGIBLE –0.008 –0.027 –0.038 (0.46) (0.34) (0.31) FOREIGN 0.003** 0.015** 0.016*** (0.02) (0.04) (0.00) LCITY_GDP –0.001 –0.007* –0.006* (0.60) (0.07) (0.05) MARKET –0.002*** 0.003** 0.002*** (0.00) (0.02) (0.00) CONS 0.145*** –0.066 –0.038 (0.00) (0.50) (0.52) YEAR FE Yes Yes Yes INDUSTRY FE Yes Yes Yes N 6,999 6,999 6,999 R 0.215 0.042 0.043 IV F TEST 842.221 – – regression results are shown in Table 14. They reveal that the latitude of the city is negatively correlated with the typhoon’s destructive power, indicating that the higher the latitude, the lower the typhoon effect, which is consistent with the above expectations. Furthermore, the latitude of the city passes the weak IV test, which indicates that IV variables have a strong explanatory power for independent variables. In the second stage of the regression, the explanatory variable is positively correlated with the DETR index at the 10% level. The above results show that the typhoon’s destructive power causes the reduction of corporate tax avoidance, which is consistent with the main test results. Method 2: PSM and difference in differences test. To further verify the robustness of the results, the paper takes coastal city enterprises as the research sample and adopts the PSM method to match the sample according to the enterprise characteristics, with similar matches performed between TYPHOON-CITY>0 and TYPHOON-CITY=0 regions. In the matching sam- ple, there are 1,667 observations of enterprises in the TYPHOON-CITY>0 regions, and 1,606 enterprises in the TYPHOON-CITY=0 regions. Panel B of Table 15 shows that the difference between each control index is not greater than 5%, which demonstrates the rationality of our matching result. Panel A shows the PSM sample regression results, and the coefficient of typhoon damage and DETR are still significantly positive at the 5% or 10% level, which further supports hypothesis H1. CHINA JOURNAL OF ACCOUNTING STUDIES 501 Table 15. t est of PSM paired samples. Panel A: (1) (2) (3) DETR DETR DETR TYPHOON_CITY 0.105* 0.101** 0.097* (0.054) (0.048) (0.051) SIZE –0.001 –0.001 (0.907) (0.885) ROA 0.244*** 0.245*** (0.000) (0.000) LEV –0.060*** –0.060*** (0.000) (0.000) BM 0.135*** 0.135*** (0.002) (0.002) TANGIBLE –0.077** –0.076** (0.011) (0.011) INTANGIBLE –0.021 –0.022 (0.635) (0.632) FOREIGN 0.014 0.014 (0.108) (0.106) LCITY_GDP 0.059 (0.466) MARKET 0.002 (0.500) CONS –0.247*** –0.203* –0.416 (0.000) (0.068) (0.186) YEAR FE Yes Yes Yes INDUSTRY FE Yes Yes Yes CITY FE Yes Yes Yes N 3,273 3,273 3,273 R 0.050 0.081 0.081 Panel B: Typhoon-City >0 Typhoon-City =0 dif(%) t -values t t SIZE 21.747 21.617 9.8 0.64 LEV 0.074 0.071 2.5 0.88 CASH 0.456 0.465 –3.4 –1.03 GROWTH 0.506 0.492 5.9 1.15 ROA 0.245 0.247 –0.7 –0.27 SOE 0.052 0.059 –9.6 –1.03 AGE 0.571 0.603 –6.4 –1.28 CITYGDP 6.331 6.361 –3.3 –1.13 N 1,667 1,606 – – Furthermore, to avoid the regional bias during the selection of matching sample, we use the method used by Dessaint and Matray (2017). We clearly separate disaster areas and surrounding areas and separately match the sample. In the paper, the sample of affected areas is retained, and the listed companies in these geographical locations are selected as the matching sample. As shown in column (1) in Table 16, the typhoon’s destructive power is positively correlated with DETR at the 5% level, which is consistent with the above conclusions. In addition, to separate more clearly the effect of typhoon damage and urban character - istics on corporate tax avoidance, we use the difference-in-differences method to explore the above conclusions with the matching sample. This specific test method consists of select - ing as the observation object a regional company affected by a typhoon in one year but not in the previous year, and keeping the observations of the year and previous year. At the same 502 L. SIYI ET AL. Table 16. difference in difference test. (1) (2) detr detr TYPHOON_CITY 4.988** (0.010) TYPHOON_DUM 0.004 (0.003) TYPHOON_DUM*TYPHOON_YEAR 0.034*** (0.001) TYPHOON_YEAR –0.007 (0.798) SIZE –0.013* 0.009 (0.084) (0.178) ROA 0.281*** 0.248*** (0.000) (0.000) LEV –0.080*** –0.063*** (0.000) (0.004) MB 0.189*** 0.101** (0.000) (0.026) TANGIBLE –0.079*** –0.020 (0.001) (0.489) INTANGIBLE 0.049 –0.147 (0.743) (0.194) FOREIGN 0.002 0.011 (0.850) (0.424) LCITY_GDP –0.250 0.112 (0.138) (0.317) MARKET 0.016 –0.011** (0.108) (0.023) CONSTANT 1.064 –0.905** (0.144) (0.041) YEAR FE Yes Yes IND FE Yes Yes CITY FE Yes Yes N 1,431 471 R 0.073 0.149 time, this method allows us to select a matching sample in adjacent regions not affected by typhoons. TYPHOON_DUM is a dummy variable for the area affected by typhoons, and the value of the affected area is 1 (including the year and year before the typhoon occurs). TYPHOON_YEAR is a dummy variable evaluated at 1 in the year the typhoon occurs (including the typhoon area and the adjacent area). As shown in column (2) in Table 16, the cross-term between TYPHOON_DUM and TYPHOON_YEAR is significantly positively correlated with DETR at the 1% level, which indicates that corporate tax avoidance significantly decreases in typhoon-affected areas. 5.5.3. Test of other samples To verify the generalisability of our conclusions in terms of spatial and temporal dimensions, we test the sample of listed companies in China and the sample of coastal listed companies in the 2002–2015 period. For the spatial dimension, considering that typhoon damage mainly occurs in coastal areas, in the above test we limit the scope of the sample to listed companies in coastal areas. However, when a ‘super typhoon’ with high wind speed and long duration approaches, some non-coastal areas are likely to be affected based on the current model for the degree of typhoon damage. Therefore, to avoid the sample selection bias, we use the sample of listed companies at the national level and recalculate the damage degree of CHINA JOURNAL OF ACCOUNTING STUDIES 503 Table 17. t est of other samples Listed companies in China Coastal listed companies in the 2002–2015 (1) (2) DETR DETR TYPHOON_CITY 0.108* 0.154** (0.057) (0.01) SIZE –0.001 –0.005 (0.800) (0.47) ROA 0.252*** 0.085** (0.000) (0.01) LEV –0.060*** –0.029*** (0.000) (0.00) MB 0.102*** 0.077** (0.000) (0.02) TANGIBLE –0.047*** –0.040** (0.001) (0.03) INTANGIBLE –0.065** –0.059 (0.017) (0.24) FOREIGN 0.008* 0.015** (0.064) (0.03) LCITY_GDP –0.023 0.069 (0.494) (0.23) MARKET 0.001 0.000 (0.552) (0.95) CONS 0.035 –0.203 (0.730) (0.15) Year FE Yes Yes Ind. FE Yes Yes City FE Yes Yes N 15,549 10,116 R 0.067 0.022 all typhoons in cities all over the country during the sample period, including cities in non- coastal areas that may be affected. The results of column (1) in Table 17 show that in the national sample, TYPHOON_CITY is significantly positively correlated with DETR at the 10% level, which is consistent with the main results. For the temporal dimension, in the above tests we take into account the effects of income tax reform; therefore, the sample interval is set after 2008. However, after the 2002 tax-sharing reform, the phenomenon mentioned earlier existed in the years before 2008. Therefore, we examine the sample before 2008 to diminish the sample selection bias. The results of column (2) in Table 17 show a significant positive cor - relation at the 5% level between TYPHOON_CITY and DETR during the 2002–2015 period, which is consistent with the main results. In summary, the main conclusion of the paper is not affected by different sample ranges or intervals. 5.5.4. Other robustness tests Considering the possible problems caused by die ff rent variable measurements, data sources, and model settings, we conduct a series of other robustness tests. First, regarding the effect of deferred corporate tax on the calculation of the effective tax rate according to the related literature, we use the difference between corporate tax expense and deferred corporate tax expense divided by pre-tax profit to measure the degree of tax avoidance (Liu & Wu, 2014). The results of column (1) in Table 18 show that the coefficient of the degree of typhoon damage is still significantly positive at the 5% level. Second, considering that not all listed 504 L. SIYI ET AL. Table 18. o ther robustness tests. Parent company Y=ETR, And control Control previous losses and Y=DETR_DEF data X=Typhoon_Num STR investment gains Control individual fixed effects (1) (2) (3) (4) (5) (6) detr_def detr detr etr detr detr TYPHOON_CITY 0.096** 0.061* 0.091* 0.107** 0.104* (0.016) (0.08) (0.09) (0.05) (0.070) TYPHOON_NUM 0.007* (0.077) STR 0.325*** (0.00) INVRATE –0.023 t-1 (0.34) LOSS –0.044*** t-1 (0.00) SIZE 0.003 0.004 –0.005 –0.003 –0.005 0.021** (0.343) (0.16) (0.349) (0.46) (0.30) (0.015) ROA 0.263*** 0.126*** 0.267*** 0.232*** 0.241*** 0.211*** (0.000) (0.00) (0.000) (0.00) (0.00) (0.000) LEV –0.104*** –0.041*** –0.056*** –0.041*** –0.049*** –0.010 (0.000) (0.00) (0.000) (0.00) (0.00) (0.516) BM 0.055*** –0.047*** 0.126*** 0.132*** 0.117*** 0.091** (0.001) (0.01) (0.000) (0.00) (0.00) (0.035) TANGIBLE –0.086*** 0.142*** –0.066*** –0.051** –0.055*** –0.048* (0.000) (0.00) (0.001) (0.02) (0.01) (0.071) INTANGIBLE –0.070** 0.094 –0.038 –0.017 –0.021 –0.089 (0.041) (0.26) (0.228) (0.57) (0.47) (0.320) FOREIGN 0.009 0.007 0.016** 0.003 0.013* 0.023 (0.203) (0.15) (0.023) (0.62) (0.06) (0.130) LCITY_GDP 0.019 0.058 –0.005 –0.004 0.001 0.003 (0.757) (0.19) (0.124) (0.92) (0.98) (0.958) MARKET –0.001 0.003 0.002** 0.001 0.002 0.002 (0.720) (0.11) (0.022) (0.82) (0.38) (0.562) CONS –0.274 –0.294* –0.048 –0.035 –0.126 –0.581** (0.223) (0.08) (0.608) (0.85) (0.50) (0.034) Year FE Yes Yes Yes Yes Yes Yes Ind. FE Yes Yes Yes Yes Yes no City FE Yes Yes Yes Yes Yes no Firm FE no no no no no Yes N 6,925 6,372 6,999 6,999 6,996 6,999 R 0.087 0.103 0.042 0.087 0.059 0.266 CHINA JOURNAL OF ACCOUNTING STUDIES 505 companies pay corporate tax as a group, their subsidiaries and headquarters may not be in the same area. To alleviate the effect of inconsistent statutory tax rates on corporate tax avoidance indicators, we replace the data from consolidated fiscal statements with the data from parent companies. The results of column (2) in Table 18 show that after the data source is replaced with the parent company data, the coefficient of typhoon damage degree is still significantly positive at the 10% level. Third, we use the natural logarithm of the number of typhoons that cause damage (Typhoon_Num) as an indicator of typhoon damage degree, as the effect of frequent typhoons may have a more substantial effect on the economy than a single destructive typhoon. The results of column (3) in Table 18 show that the degree of typhoon damage is still significantly positively correlated with the degree of corporate tax avoidance with the new indicator. Fourth, following Atwood, Drake, Myers, and Myers (2012), we use the effective tax rate (ETR) as the independent variable and the statutory tax rate (STR) as the control variable to rebuild the testing model. The results of column (4) in Table 18 show that the coefficient of typhoon damage is still significantly positive at the 10% level. Fifth, we test whether the losses occurred during the previous period and whether the investment gains are likely to offset taxable income. To avoid the effect of the previous operating results, we further control these two factors. The results of column (5) in Table 18 show that the degree of typhoon damage is still positively related to tax avoidance at a 5% significance level. Sixth, there may be differences in the characteristics of companies in affected areas, and these die ff rences may ae ff ct the motivation of corporate tax avoidance. Therefore, we control the individual fixed effects at the firm level. The results of column (6) in Table 18 show that the coefficient of typhoon damage is still significant at the 10% level. In summary, the main conclusions of the paper are robust to different variable measurements, data sources, model settings, and other factors. 6. Conclusion This paper studies the effect of natural disasters on corporate tax avoidance behaviour at the micro level. Based on the data of A-share listed companies in China from 2008 to 2015, this paper chooses a specific disaster (typhoon) and examines the relationship between the degree of typhoon damage and the corporate tax avoidance behaviour in the current year. Results show that the greater the damage caused by typhoons, the lower the tax avoidance rate of local enterprises. Moreover, this relationship is more evident in state-owned enter- prises, political affiliates, and local leading enterprises. We also examine the differences in post-disaster tax avoidance behaviour in different regions and in different types of enter - prises from four perspectives: institutional environment, economic growth and fiscal pres- sure, market supervision mechanism, and tax avoidance margin. The main results are as follows. Typhoon disasters have a lower effect on corporate tax avoidance behaviour in areas with a higher degree of marketisation. The effect is more significant in areas where the pressure of economic growth and fiscal revenue is stronger. The effect of typhoon disasters on corporate tax avoidance mainly exists in enterprises with an effective tax rate lower than their nominal tax rate and with a tax avoidance margin. These effects are also lower in firms with more institutional holdings and analyst following. At the same time, we find that enter - prises that pay more taxes after a disaster in the current year can obtain more government 506 L. SIYI ET AL. subsidies and credit resources in the future, which creates a balance in the two-way exchange of benefits between local governments and enterprises. The findings of this paper reflect the interaction between local governments and enter - prises in tax distribution from the perspective of natural disasters, and provide a better understanding of the relationship between local governments and enterprises in the context of the economic transition in China. Taxes can be considered the residual claim of local governments on enterprises. To some extent, local governments become major corporate shareholders (Desai & Dharmapala, 2006). However, local governments must also fulfil their contract with enterprises during profit distribution. Enterprises have the incentive to alleviate the government fiscal pressure after a typhoon disaster by lowering their level of tax avoid- ance and paying more taxes, which changes the stability of the contract between local governments and enterprises. Our findings show that if natural disasters modify the original contract, they also induce an implicit contract of benefit exchange. To some extent, this process may affect the market equity mechanism. For instance, enterprises supporting local governments by paying more taxes can offset their past expenditure by obtaining fiscal subsidies and credit resources in the future. Conversely, enterprises that do not support the government tax revenue demand may be at a disadvantage in resource allocation. Such behaviour may thus distort the market’s allocation of resources and cause an unfair distri- bution of resources. Our findings also have some policy implications. First, why do enterprises have a strong motivation to help local governments and reduce tax avoidance to alleviate the government fiscal pressure after natural disasters? The root cause lies in the mismatch between financial power and social relief power of local governments. Local governments assume the respon- sibility of post-disaster relief and reconstruction; however, as they have insufficient funds for fiscal expenditure, they can only rely on enterprises to share their fiscal pressure. Therefore, it is important to find ways to mitigate the post-disaster fiscal pressure on local governments through institutional design. For example, local governments should establish and improve a disaster reserve system and distribute the proportion of fiscal allocations between different government levels in response to different levels of disaster. Correspondingly, the annual budget of governments should also consider possible natural disasters and transfer reserves for disaster-related expenditure. In addition, to ease the pressure on local government offi- cials for economic growth in the promotion championship and reduce the possibility of excessive investment in post-disaster reconstruction, non-controllable factors should be considered when higher-level departments examine the performance of local officials. Second, it is difficult to fully cover the adverse effects of natural disasters based solely on government fiscal expenditure. The market insurance mechanism should be established and marketed to build the disaster insurance scheme at the social level. Insurance institutions would be a powerful supplementary mechanism for disaster relief. Current research also shows that an effective social insurance system is an important factor influencing disaster response in different regions (Borensztein, Cavallo, & Valenzuela, 2009). Finally, local enterprises should stop treating corporate taxation as ‘political contributions’. Instead, ‘self-selection’ corporate charitable donations should be encouraged, as these meas- ures can effectively raise corporate public awareness, enhance brand awareness, and play an advertising role by accumulating reputation capital (Bennett, 1998). Hence, enterprises can choose their optimal level of charitable donations according to their own business con- ditions, industry characteristics, or other factors as part of the rational choice to maximise CHINA JOURNAL OF ACCOUNTING STUDIES 507 profit (Pan et al., 2017; Xu et al., 2011). As a result, enterprises can fulfil their social respon- sibilities and establish a good social reputation through charitable donations while helping to ease the fiscal pressure on local governments in a more direct manner to create a win- win-win situation for local governments, enterprises, and society. Funding This work was supported by National Natural Science Foundation of China [Grant Numbers 71790604; 71332008; 71702030]. References Atwood, T.J., Drake, M.S., Myers, J.N., & Myers, L.A. (2012). Home country tax system characteristics and corporate tax avoidance: International evidence. The Accounting Review, 87, 1831–1860. Bai, C., Li, D.D., Tao, Z., & Wang, Y. (2000). A multitask theory of state enterprise reform. Journal of Comparative Economics, 28, 716–738. Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Bennett, R. (1998). Market orientation among small to medium sized UK charitable organisations: Implications for fund-raising performance. Journal of Nonprofit & Public Sector Marketing, 6, 31–45. Bhojraj, S., & Sengupta, P. (2003). Effect of corporate governance on bond ratings and yields: The role of institutional investors and outside directors. The Journal of Business, 76, 455–475. Boose, E.R., Serrano, M.I., & Foster, D.R. (2004). Landscape and regional impacts of hurricanes in Puerto Rico. Ecological Monographs, 74, 335–352. Borensztein, E., Cavallo, E., & Valenzuela, P. (2009). Debt sustainability under catastrophic risk: The case for government budget insurance. Risk Management and Insurance Review, 12, 273–294. Cai, H., & Rao, P. (2015). Institutional investor tax enforcement and tax avoidance. Accounting Research, 10, 59–65 (In Chinese). Cao, S., Liu, X., & Zhang, W. (2009). Fiscal decentralization, local government competition and the actual tax burden of listed companies. The Journal of World Economy, 4, 69–83 (In Chinese). Cavallo, E. (2011). Natural disasters and the economy – A survey. International Review of Environmental and Resource Economics, 5, 63–102. Chen, D., Chen, Y., & Dong, Z. (2016). Policy uncertainty, tax collection strength and corporate tax avoidance. Management World, 5, 151–163 (In Chinese). Chen, D., Kong, M., & Wang, H. (2016). Economic cycle and state-owned companies’ tax avoidance. Management World, 5, 46–63 (In Chinese). Chen, H., Tang, S., Wu, D., & Yang, D. (2015). The political dynamics of corporate tax avoidance: The Chinese experience. Working paper. Available at: https://ssrn.com/abstract=2640111. Dai, Y., Pan, Y., & Feng, S. (2014). Are Chinese enterprises’ charitable donations ‘political contributions?’ – Evidence from the replacements of the municipal party secretaries. Economic Research Journal, 2, 74–86 (In Chinese). Desai, M.A., & Dharmapala, D. (2006). Corporate tax avoidance and high-powered incentives. Social Science Electronic Publishing, 79, 145–179. Dessaint, O., & Matray, A. (2017). Do managers overreact to salient risks? Evidence from hurricane strikes. Journal of Financial Economics, 126, 97–121. Elliott, R.J.R., Strobl, E., & Sun, P. (2015). The local impact of typhoons on economic activity in China: A view from outer space. Journal of Urban Economics, 88, 50–66. Emanuel, K. (2011). Global warming effects on US hurricane damage. Weather, Climate, and Society, 3, 261–268. 508 L. SIYI ET AL. Faccio, M., Masulis, R.W., & McConnell, J.J. (2006). Political connections and corporate bailouts. Journal of Finance, 61, 2597–2635. Fan, J.P., Wong, T.J., & Zhang, T. (2007). Politically connected CEOs, corporate governance, and post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics, 84, 330–357. Fan, J.P.H., Wong, T.J., & Zhang, T. (2013). Institutions and organizational structure: The case of state- owned corporate pyramids. Journal of Law Economics & Organization, 29, 1217–1252. Fan, Z., & Tian, B. (2013). Tax competition, tax enforcement and tax avoidance. Economic Research Journal, 9, 99–111 (In Chinese). Felbermayr, G., & Gröschl, J. (2014). Naturally negative: The growth effects of natural disasters. Journal of Development Economics, 111, 92–106. Gillan, S.L., & Starks, L.T. (2000). Corporate governance proposals and shareholder activism: The role of institutional investors. Journal of Financial Economics, 57, 275–305. Goldman, E., Rocholl, J., & So, J. (2013). Politically connected boards of directors and the allocation of procurement contracts. Review of Finance, 17, 1617–1648. Green, S. (2013). What makes 10 million local government officials tick. Research report, Standard Chartered Bank. Available at: http://www.cf40.org.cn/plus/view.php. Gu, Z., Tang, S., & Wu, D. (2016). The political economy of labor cost behavior: Evidence from China. Working Paper. Available at: https://ssrn.com/abstract=2786533. Guo, J., & Li, T. (2009). Inter-governmental tax competition in China: Based on the province-level panel data. Management World, 11, 54–64 (In Chinese). Hanlon, M., & Heitzman, S. (2010). A review of tax research. Journal of Accounting and Economics, 50, 127–178. Holland, G. (2008). A revised hurricane pressure wind model. Monthly Weather Review, 136, 3432–3445. Jia, J., & Ying, S. (2016). Fiscal decentralization and corporate tax incentives: An analysis from views of local government competition. China Industrial Economics, 10, 23–39 (In Chinese). Kim, Y., & Cannella, A.A. (2008). Social capital among corporate upper echelons and its impacts on executive promotion in Korea. Journal of World Business, 43, 85–96. Kubick, T.R., Lockhart, G.B., Mills, L.F., & Robinson, J.R. (2017). IRS and corporate taxpayer effects of geographic proximity. Journal of Accounting and Economics, 63, 428–453. Landry, P.F. (2012). Decentralized authoritarianism in China: The Communist Party’s control of local elites in the post-Mao Era. Cambridge, UK: Cambridge University Press. Leiter, A.M., Oberhofer, H., & Raschky, P.A. (2009). Creative disasters? Flooding effects on capital, labour and productivity within European firms. Environmental and Resource Economics, 43, 333–350. Li, W., & Xu, Y. (2013). Tax avoidance effect of political status. Journal of Financial Research, 3, 114–129 (In Chinese). Liu, H., & Wu, L. (2014). Institution environment, state ownership and the corporate effective tax rate. Management World, 4, 42–52 (In Chinese). Liu, J., & Liu, F. (2014). Fiscal centralization, government control and corporate tax burden: Evidence from China. Accounting Research, 1, 21–27 (In Chinese). Luo, D., & Wei, Z. (2012). Political connection and the tax avoidance of enterprises: The evidence from Chinese private listing enterprises. South China Journal of Economics, 11, 29–39 (In Chinese). Luo, D., & Yang, Y. (2013). Property rights, political connections and corporate tax: Evidence from listed companies in China. World Economic Papers, 4, 1–19 (In Chinese). Muller, A., & Whiteman, G. (2009). Exploring the geography of corporate philanthropic disaster response: A study of Fortune Global 500 firms. Journal of Business Ethics, 84, 589–603. Noy, I. (2009). The macroeconomic consequences of disasters. Journal of Development Economics, 88, 221–231. Noy, I., & Nualsri, A. (2011). Fiscal storms: Public spending and revenues in the aftermath of natural disasters. Environment and Development Economics, 16, 113–128. Pan, Y., Weng, R., & Liu, S. (2017). The selfish goodwill: New evidence from corporate philanthropy in typhoon. China Industrial Economics, 5, 133–151 (In Chinese). Pelling, M., Özerdem, A., & Barakat, S. (2002). The macro-economic impact of disasters. Progress in Development Studies, 2, 283–305. CHINA JOURNAL OF ACCOUNTING STUDIES 509 Raddatz, C. (2007). Are external shocks responsible for the instability of output in low-income countries? Journal of Development Economics, 84, 155–187. Raschky, P.A. (2008). Institutions and the losses from natural disasters. Natural Hazards & Earth System Sciences, 8, 627–634. Shan, L., Gan, L., & Zheng, T. (2008). Corporate donations and economic incentives: An empirical study based on corporate donations following the 512 earthquakes in China. Economic Research Journal, 11, 51–61 (In Chinese). Toya, H., & Skidmore, M. (2007). Economic development and the impacts of natural disasters. Economics Letters, 94, 20–25. Vu, T.B., & Hammes, D. (2010). Dustbowls and high water, the economic impact of natural disasters in China. Asia-Pacific Journal of Social Sciences, 1, 122–132. Vu, T. B., & Noy, I. (2015). Regional effects of natural disasters in China: Investing in post-disaster recovery. Natural Hazards, 75, 111–126. Wang, Y., Wang, G., & Peng, Y. (2012). Does ownership nature affect tax sensitiveness? Evidence based on firm’s labor demand. Economic Research Journal, 4, 52–63 (In Chinese). Wong, T.J. (2016). Corporate governance research on listed firms in China: Institutions, governance and accountability. Foundations and Trends in Accounting, 9, 259–326. Wu, L., & Li, C. (2007). Tax refunds, the corporate tax burden and the effectiveness of tax policy. Social Sciences in China, 4, 61–73 (In Chinese). Xu, L., Xin, Y., & Zhu, J. (2011). Media coverage and the social responsibilities of listed company: Empirical study base on the donation in Wen Chuan Earthquake. Management World, 3, 135–143 (In Chinese). Ye, K., & Liu, X. (2011). Tax administration, income tax cost and earnings management. Management World, 5, 140–148 (In Chinese). Ye, S., & Ye, Z. (2012). Location choice of industrial firms in China: Market potential, resource endowment and tax burden level. Nankai Economic Studies, 5, 94–110 (In Chinese). Zeng, Y., & Zhang, J. (2009). Cantax collection play a corporate governance role? Management World, 3, 143–151 (In Chinese). Zhang, M., Ma, L., & Zhang, W. (2013). The bond effect of firms’ charitable donation: Empirical evidence from Chinese listed company. Management World, 7, 163–171 (In Chinese). Zhang, M., Zhang, S., Wang, C., & Shen, H. (2010). Political connection and the credit resources allocative eci ffi ency: Empirical evidence from Chinese private listed company. Management World, 11, 143–153 (In Chinese). Zhuo, Z., & Duan, S. (2012). Disaster prevention and mitigation investment, disaster control and economic growth: Economic analysis and empirical in China. Management World, 4, 1–8 (In Chinese).
China Journal of Accounting Studies – Taylor & Francis
Published: Oct 2, 2017
Keywords: Government-company relationship; natural disasters; tax avoidance; typhoon
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.