Impact of Market Concentration on Effective Sales Tax Rate: Evidence from Insurance Industry

Transcription

1 Impact of Market Concentration on Effective Sales Tax Rate: Evidence from Insurance Industry Ghanshyam Sharma June 8, 2018 Abstract This paper examines the impact of market concentration in the insurance industry on the effective insurance premium sales tax. I use unexpected disasters normalized by the population of the state as a source of exogenous variation in the market concentration. Unexpected disasters lead to the exit of insurance firms from the state leading to higher market concentration. I find empirical evidence that a ten percent increase in market concentration leads to five percent reduction in the effective insurance premium tax rate. This implies a reduction of $84 mn in tax revenues. I also show a correlation between market concentration and campaign contributions. This suggests that higher market concentration increases the likelihood of firms coming together as an interest group and lobby for a favorable (tax) policy. JEL Classification: Market Concentration, Lobbying, Sales Tax, Regulation. *Assistant Professor of Economics at Seton Hall University; gsharma@clemson.edu

2 1 Introduction Several studies have examined the impact of taxes on the behavior of individuals and on other aspects of the economy (Feldstein 2008; Harju Kosonen 2012). However, there exists substantial variation in tax structures and tax rates across jurisdictions and sectors 1. Various interesting explanations have been provided for this variation in taxation. Some of these include tax competition (Devereux & Loretz 2012; Devereux & Lockwood 2008; Pal & Sharma 2013), political affiliations (Warren 2013; Besley & Case 2005) and political institutions (Swank & Steinmo 2002; Bolton & Roland 1997). This paper adds to this literature by showing that the market structure of an industry (i.e. market concentration) can have a substantial impact on the effective tax rates on it. In this paper, I focus on the sales tax levied on the premiums written by the industry 2. I also show that high market concentration in the insurance industry is positively associated with higher campaign contributions by the industry. This paper, therefore, also contributes to the literature on rent-seeking, regulatory capture and privileges granted by the government (Nava & Klunover 2014; Murphy, Shleifer & Vishny 1993). One of the main highlights is that the welfare costs of monopolies are larger than the traditional dead-weight costs. This is in line with the arguments made by Tullock (1967). I utilize the data available from the insurance industry in the United States to examine the causal impact of market concentration on the effective premium sales tax rate. I utilize two-stage least squares estimation strategy. I use the unexpected disasters in a state normalized by a state s population as a possible source of exogenous variation in market concentration. I find that a ten percent increase in market concentration causes the effective tax rate to go down by five percent. This translates into about $84 million less in premium tax collections annually. The results are robust to state and year fixed effects. The errors are robust to heteroscedasticity and clustered at the state level. I also include 1 For example, see the report of the Tax Foundation on Local and Sales Tax Rates across the United States. Also, see the report on corporate income taxes across the world by Organization for Economic Co-operation and Development (OECD). 2 value of policies sold 1

3 a variety of controls like regulatory features, political affiliation of the Governor, campaign contributions, and the size of the economy. As the market concentration in the industry goes up, it becomes easier for the existing firms to organize themselves as an interest group. In highly concentrated industries, firms face lower costs of organization and limited free rider problem (Stigler 1971, 1974; Pittman 1977, 1984). As it becomes easier for the firms to come together as an interest group, they are more likely to lobby for a favorable policy. Therefore, the states in which the market concentration is higher, the firms are more likely to organize themselves as an interest group and lobby for lower effective premium tax rates. The first stage results provide evidence of increased lobbying activity associated with higher market concentration. The results show that there is a positive and statistically significant association between campaign contributions (made by the insurance industry to finance state elections) and market concentration. This is consistent with Pittman (1976) who showed a positive association between market concentration in an industry and campaign contributions by the industry. This is also consistent with the substantial economic literature examining the influence of interest groups and lobbying on regulation. This includes environmental regulation (Maloney and McCormick 1988), the banning of the importation of slaves into the United States (Anderson et al., 1988), immigration restrictions (Shugart et al., 1986) and apostolic decrees by the Roman Catholic Church (Ault et al., 1987) among others. Richter, Samphantharak, and Timmons (2009) have shown that there is an inverse relationship between campaign contributions and taxes paid by firms. Media outlets often report that interest groups belonging to different sectors of the industry exert tremendous influence on tax policy. For example, Huffington Post reported that Nissan got US$ 1.3 billion in tax benefits from the state government of Mississippi 3. The Daily Beast covered a story on race track owners associated with National Association for Stock Car Auto Racing (NASCAR) getting tax breaks worth about US$ 40 million a 3 Mississippi Cuts $1.3 Billion From Schools, Gives $1.3 Billion to Nissan, 23rd May 2014, The Huffington Post. 2

4 year 4. Fox News reported that Senate Finance Committee chairman, Democratic Sen. Ron Wyden was influential in getting tax breaks of up to US$ 2,500 on the purchase of an electric motorcycle while pointing out that the two electric motorcycle manufacturers came from his state of Oregon 5. Section two provides a brief description of the regulatory environment in state insurance markets and insurance premium sales tax. Section three contains data sources and description. Section four has a discussion on the instrument and the estimation strategy while Section 5 provides a discussion of the results. 2 Insurance Premium Sales Tax and Background Information McCarran-Ferguson Act of 1945 gave states the sole power over regulation and taxation of insurance which makes the insurance industry exempt from federal anti-trust regulations. An insurance firm requires regulatory approval from each state it wants to operate in. This approval, in the form of a license, is granted by the state insurance regulator. It is also quite common for a parent organization to operate in multiple segments of the industry by floating different firms. Hence, different segments of the insurance industry are linked to each other because they are part of a strategic decision making by the parent organization. Further, the insurance firms cannot cross subsidize risks across different states. Grace (2008) points out that as states are not subject to the commerce clause, they discriminate against out-of-state companies through taxation. To counter this discriminatory taxation, states adopted a so-called retaliatory tax. That is, if state A would tax state Bs companies at a higher rate than its own companies, state B would tax State As companies at the higher of the two states tax rates. Retaliatory taxes are imposed to tax away any advantage an out-of-state company may have because of lower taxes imposed in the state 4 8 Ridiculous Tax Loopholes: How Companies Are Avoiding the Tax Man, 25th Feb 2012, The Daily Beast

5 in which the company is domiciled. Insurance companies also get tax exemptions and credits in different forms in different states. For example, the statutory premium tax rate in Georgia is 2 percent. However, for companies with at least 25 percent assets in the state, the rate is reduced to 1 percent. The rate is further reduced to half a percent for companies with at least 75 percent of their assets in the state. In Kansas, insurance companies are given a tax credit equal to 15 percent of Kansas-based employees salaries (not including commissions or fringe benefits), or up to percent of taxable premiums dependent on the company s affiliation. This implies an effective tax rate of 0.8 percent 6. These examples are not exhaustive as most states have some kind of tax incentives for the insurance industry. Hence, actual tax rate that insurance firms face, on an average in a state, is quite different than the statutory tax rate. Therefore, I use effective tax rate (which is a percent of the taxes collected on insurance premiums to total premiums written and annual considerations). States in the US levy a state premium tax on the gross premiums written (rather than on profits) by the insurance industry. The industry is exempt from corporate income taxes 7. $1.67 billion were collected in taxes on insurance premiums on $1.54 trillion worth of insurance premiums and annuity considerations in the United States for the year Hence, this translates into 1.1 percent in effective premium tax rate across the United States. Delaware had the lowest effective rate of 0.23% while Nevada had the highest effective tax rate of 2.37%. On an average, premium tax revenues constitute about 2 percent of the total tax revenues of the state. This proportion is the lowest for Alaska at 0.8% and highest for Tennessee at 5.3%. 3 Data and Descriptive Statistics Effective Tax Rate (τ) is computed by the taking the ratio of insurance premium tax revenues in a state to the total premium written by the industry in the state in a specific except Oregon which for this reason has been dropped from the analysis 4

6 year. Mathematically, τ = ( T st P st ) 100 (1) where, P st = P lst + P hst + P p&cst ; P lst, P hst and P p&cst are the premiums written by the life, health, and property & casualty segments of the insurance industry respectively. T st is the insurance premium sales tax revenues as reported by United States Census Bureau (USCB). Data on premium tax revenues is not available by different segments of the industry. Hence, it is not possible to estimate the effect of market concentration on the effective tax rate for each segment. Market concentration in the insurance industry is the weighted average of Herndahl index (HHI) of life, health, and p&c segments, the weights being the proportion of the insurance premiums written by each segment. Data used for constructing the degree of concentration index in different segments of the insurance industry is available from SNL Financial database. This data is available for all 50 states and Washington D.C. I use data over the period The Federal Emergency Management Agency (FEMA) collects data on major disasters by state and year. To define an event as a disaster, FEMA uses a statutory definition provided by the US Congress. Hence, a disaster is defined as any natural catastrophe (including any hurricane, tornado, storm, high water, wind-driven water, tidal wave, tsunami, earthquake, volcanic eruption, landslide, mudslide, snowstorm, or drought) or, regardless of cause, any fire, flood, or explosion, in any part of the United States, which, in the determination of the President, causes damage of sufficient severity and magnitude to warrant major disaster assistance under the Stafford Act to supplement the efforts and available resources of States, local governments, and disaster relief organizations in alleviating the damage, loss, hardship, or suffering caused thereby. FEMA also provides information on the disaster declaration process. In the event of the 8 Prior to 2001, data on health segment of the insurance industry is not available. 5

7 occurrence of a disaster, state and federal officials conduct a joint assessment to determine the extent of the disaster and its impact on individuals and public facilities. The Governor has to furnish information on the nature and amount of state and local resources that have been or will be committed to alleviating the results of the disaster and provide an estimate of the type and amount of assistance needed under the Stafford Act. A Disaster is declared if the President determines that federal assistance is needed 9. Some studies have shown a statistically significant effect of political party affiliations of the Governor on the state tax policy (Warren 2013; Besley & Case 1995). Therefore, I have added controls for political affiliation of the Governor. I use a dummy variable which takes the value of 1 if the Governor is from the Democratic party, 0 otherwise. The insurance regulator is selected either through appointment (by the Governor) or election (by ballot). The selection method of the regulator can have some influence on the insurance industry (Besley & Coate 2003). The states where the insurance regulator is an elected position is Washington, California, Montana, North Dakota, Louisiana, Mississippi, Oklahoma, Kansas, Georgia, North Carolina and Delaware. This information is available from National Association of Insurance Commissioners (NAIC). Historical data on Gross Domestic Product (state-wise) for all states in the United States is available at Bureau of Economic Analysis. Data on campaign contributions by the insurance industry to fund state-level elections is available from National Institute for Money in State Politics. All states equip an insurance commissioner with the power to regulate insurance premiums, particularly in the Property & Casualty segment of the industry. This regulatory power varies from state to state. In several states, insurance firms cannot change premiums without prior approval from the insurance regulator. This can have a substantial impact on the premiums and the performance of firms. This regulatory power varies from state to 9 FEMA website. 6

8 state. This regulatory powers are Prior Approval 10, File and Use 11, Use and File 12 and Flexible Rate Regulation 13. The data on rate regulation is available from National Association of Insurance Commissioners (NAIC). 4 Estimation Framework 4.1 Instrument Variable I employ two stage least squares estimation strategy to show the causal effect of market concentration on effective premium tax rate. I use unexpected disasters normalized by the population of the state as a source of exogenous variation in the market concentration. Unexpected Disasters directly affect the market concentration of the insurance industry. This is because being exposed to catastrophic events pose unique problems for the insurers. Small insurers may not have sufficient resources to cover the losses when exposed to risks created by low probability and high-risk events like hurricanes, earthquakes, etc. Born and Viscusi (2006) show that an unexpected catastrophe leads to the exit of insurance firms from the state and firms with low levels of homeowners premiums are most adversely affected. Under the expectation of insurance firms being rational, expected disasters would not lead to the exit of firms and subsequent higher market concentration. This is because the firms will incorporate the damage from expected disasters in the premiums. However, the insurance firms will be adversely affected if the unexpected disasters are greater than the expected disasters. I define Unexpected Disasters as the difference between the actual number of disasters and the average number of disasters that occurred in the state in the previous five years. I normalize the unexpected disasters by the population of the state because in the event 10 firms need prior approval from the insurance regulator to change rates 11 where an insurance firm needs to file an application with the insurance commissioner to change the premium it charges on a particular automobile insurance policy. After filing, the firm can switch to new premiums and continue to use them unless the commissioner objects to the change. 12 insurance firms can start using the new premiums they want to charge on a policy and inform the insurance commissioner of this change within a stipulated time period (usually three months) 13 where insurance firms are allowed to change premiums on a policy within a range in an year and need approval from the insurance commissioner if they intend to change their premiums beyond those limits. 7

9 of the occurrence of an unexpected disaster, the insurance industry will face greater losses in states where it has greater exposure to risks. Hence the insurance industry will be more adversely affected if a larger population is exposed to a disaster. Symbolically, t 1 i=t 5 D si 5 UD st = ( D st ) 100 (2) P op st where UD st, the instrument, is the number of unexpected disasters in state s and year t; D st is the number of disasters declared by the Federal Emergency Management Agency in a state s and year t; P op st is the population of the state s in year t. The instrument is valid subject to two conditions. First, the unexpected disasters affect the market concentration in the insurance industry. The first stage results (tables 2 and 4) corroborate this claim. The results are robust under different specifications. Born and Viscusi (2006) argue that in the absence of adequate reinsurance, the firm may go bankrupt or may choose to exit a state in which there is a substantial exposure to catastrophic risks. They provide examples. In response to losses incurred following Hurricane Katrina, which accounted for over $38 billion in insured losses, a major insurer, Allstate, exited several coastal states while another, State Farm, chose not to renew some policies in these areas 14. The fourth-largest insurer in Florida, Poe Financial, went bankrupt. Hurricane Andrew too caused 9 insolvencies (Born & Viscusi). The State of Florida s 2nd Annual Report on Property Insurance Market for Florida Legislature (January 2013) expresses similar concerns. According to this report, a high number of insurance companies are exiting the Florida market and there is a lack of companies entering to replace them; growing number of insurance companies are no longer writing new policies in Florida; there is a lack of insurance company formation in Florida and a slowdown in the growth of capital to support premiums written by primary insurers in Florida is slow. The regulatory environment too seems to increase the vulnerability of insurance firms to unexpected disasters. Grace and Klein (2002) highlight that insurers have sought to 14 The Price of Sunshine, The Economist, June 8, 2006, p. 76 as reported by Born and Viscusi (2006) 8

10 raise prices and decrease exposure to losses. However, the state legislature and insurance regulators have resisted insurers responses to increased risks in an attempt to preserve the availability and affordability of insurance. They also find that in New York and Florida, demand for catastrophe insurance is more price elastic than for non-catastrophic coverage, thereby putting pressure on the insurers. They also find that consumers prefer high-quality solvent insurers in Florida. Unexpected disasters can impact the firms selling health insurance and lead to subsequent increase in the market concentration in the health insurance markets. This is because disasters can adversely affect the local economy and lead to job losses 15. This would effect the sale of policies covering health through both individual and group based plans. As a report by Kaiser Foundation on Hurricane Katarina points out, most of those displaced by Katrina are now jobless, and many of those who lost their jobs will also lose their employer-sponsored health insurance coverage. Second, the exclusion restriction implied by the instrumental variable regression is that, conditional on other independent variables in the regression, the instrument has no effect on effective tax rate, other than their effect through market concentration. The instrument is a robust predictor of market concentration, yet as Figure 1 shows, is uncorrelated with the error terms of the model predicting the effective premium tax rate. The correlation between the residuals and the instrument is A concern is that the state regulatory agencies may provide assistance to the insurance firms to deal with unexpected natural disasters. This assistance may come in the form of tax relief for insurance firms. There is no study which has explored the relationship between unexpected natural disasters and premium tax rates. To explore the potential relationship between the instrument and effective premium tax rate, I examine the impact of Hurricane Katarina. After hurricane Katarina, the effective tax rates in the 4 states most affected by it was higher (1.6 percent) than in other states (1.3 percent). Besides, the effective tax rate 15 Please see Japan Institute for Labor Policy and Training s(jilpt) report on Economic development and the impacts of natural disasters. 9

11 in these four states 16 remain stable in the range of 1.4 percent to 1.6 percent in the years before and after hurricane Katarina. Tax relief in the face of natural disasters may not be an adequate policy response. As Born and Viscusi (2006) point out, the firm may suffer losses well in excess of value of premiums that it charged for the coverage. Effective premium tax rates are only about 1 percent to 2 percent of premiums written by the industry. The tax relief in the event of a natural disaster is primarily aimed at individuals who can avail these benefits while filing taxes, rather than insurance firms. The insurance industry is regulated extensively and state regulatory agencies have substantial powers to help the struggling insurance firms 17. For example, the state insurance commissioners have the regulatory authority to relax solvency requirements for insurance firms operating in a state. There is substantial interstate variation in solvency requirements for insurance firms across states 18. The regulators can also allow the insurance companies to pay fewer dollars on the claims filed by policyholders. State insurance commissioners have the power to adopt and modify claim settlement laws 19. If unexpected disasters have an impact on the market concentration in the insurance industry, it shows that several insurance firm do not have proper insurance cover. In other words, they do not have adequate reinsurance. So, the primary policy response would be to provide incentives to the industry to have adequate reinsurance, rather than tax benefits. State insurance regulators have the powers to determine the reinsurance laws in the state (Rogan 2013). To help the insurance industry to deal with losses due to natural disasters like hurricanes, Florida Legislature created a Florida Hurricane Catastrophe Fund (FHCF) to to provide additional insurance capacity and help stabilize the property insurance market in Florida (Fla. Stat. s (1)). The Fund provides reimbursements for a portion of property insurers hurricane losses above the amount retained by the insurers. Insurers enter 16 Florida, Louisiana, Mississippi and Alabama. 17 See the report on State Insurance Regulation by National Association of Insurance Commissioners. Also see Insurance Regulation Report Card by R Street 18 NAIC report on Statutory Minimum Capital And Surplus Requirements

12 into contracts with the FHCF and pay a premium 20. Hence, I argue that in the presence of so many policy alternatives, the state insurance regulators are less likely to use tax benefits as a policy to help insurance industry cope with natural disasters. In presence of these alternatives, tax benefits seem to be a weak and less effective policy alternative to help the industry deal with catastrophic natural disasters. 4.2 Econometric Model I employ two stage least squares estimation strategy to show the causal effect of market concentration on effective premium tax rate. I use the logs of key variables of interest (Market Concentration and Effective Premium Tax Rate) for convenient interpretation of the results. The results are reported in Table two and three. I have done regressions without logs (Table four) and the general empirical predictions of the paper still hold. I consider a two period model. In the first period unexpected disasters lead to higher market concentration. In the next period (or next year), it becomes marginally easier for the firms to come together and lobby for lower taxes. The first stage estimates the impact of the instrument on market concentration in the insurance industry. I employ a variety of controls like state GDP, political affiliation of the governor, selection method of the insurance commissioner and rate regulation. I also employ state and year fixed effects to address the concerns arising out of omitted variable bias. The errors are clustered at the state level and robust to heteroscedasticity. The first stage regression equation is ˆ ln(marketconcentration) s,t 1 = β 0 +β 1 (UnexpectedDisasters) s,t 1 +γφ+β 3 v t +β 4 f s +e st (3) where ˆ ln(m arketconcentration) is the predicted value of market concentration in log values; Unexpected Disasters is the Unexpected disasters as defined in equation (3); φ is the set of 20 The State of Florida s 2nd Annual Report on Property Insurance Market for Florida Legislature. 11

13 independent variables discussed above; v and f are state and year dummies. tax rate. The second stage estimates the impact of market concentration on effective premium ln(effectivep remiumt axrate) s,t = β 0 +β 1 ln(marketconcentration) ˆ s,t 1 +γφ+β 3 v t +β 4 f s +e st (4) 5 Results In table two, I provide results using the OLS model. The coefficient on market concentration in the insurance industry is negative and significant at 1 percent level. This coefficient is also stable across different specifications and adding a variety of controls. The model includes state and year fixed effects and the errors are clustered at the state level. Table three presents the first stage results. The instrument is a robust predictor of market concentration. The first specification is without any controls and the results are statistically significant. One unexpected disaster (normalized by the state s population) leads to an increase in market concentration by 2 percent. I also add a variety of controls like rate regulation system, campaign contributions by the insurance industry to fund state elections and the size of the economy as measured by the log of state s Gross Domestic Product. selection method of state s insurance commissioner, and political affiliation of the Governor. The relationship between the instrument and market concentration is stable and statistically significant across all specifications. There is also a positive and statistically significant relationship between market concentration and campaign contributions made by the insurance industry to fund statewide elections. This result is consistent with Pittman (1976) who suggested that concentrated industries are able to avoid the free rider problem and therefore make higher campaign contributions. The results are robust to state and year fixed effects. The errors are clustered at the state level and robust to heteroscedasticity. The first stage F-statistic for the significance of the instrument fares well for all specifications, implying no weak IV problem (Stock & Yogo 2005). 12

14 Table four provides the second stage results, indicating a casual impact of market concentration on effective premium tax rates. The impact of market concentration on the effective premium tax rate is stable and consistent across all specifications. Amongst all the variables in the regression, only the market concentration has an impact on effective premium tax rate. The results suggest that as market concentration in the insurance industry goes up by 10 percent, the effective insurance premium tax rate goes down by about 5 percent. To check whether the results are sensitive to different variables, I test for different specifications for the econometric model in equation (3) and equation (4). The results are reported in Table three and Table four respectively. To provide for additional robustness checks, I have done regression without logs (Table five) and the general empirical predictions of the paper still hold. The F-statistic rules out the weak IV problem. 6 Conclusion In this paper, I show how high market concentration makes it easier for the firms to capture regulation. Using robust empirical evidence, this paper provides a novel explanation for the determination of and variation in the effective tax rates faced by the insurance industry. Using the evidence from the insurance industry in the United States, I show that market concentration in an industry can have a substantial impact on the effective tax rates. I show that market concentration have a casual impact on the effective tax rates. I also discuss and provide evidence about the mechanism of the casual impact of market concentration on effective tax rates. As market concentration in the industry goes up, it becomes easier for the firms to organize themselves as an interest group. This increases the likelihood of firms lobbying for lower effective tax rates (or other favorable policies). This is reflected in the higher campaign contributions associated with higher market concentration. The paper has important implications for studies analyzing tax policies. It will also be interesting to test the theory of the impact of market concentration on effective tax rates in 13

15 other industries and other countries. The implications of this paper can also be generalized into other aspects of regulatory capture. 14

16 Anderson, G. M., Rowley, C. K., & Tollison, R. D. (1988). Rent seeking and the restriction of human exchange. J. Legal Stud., 17, 83. Ault, R.W., Ekelund R.B. Jr., & Tollison, R.D. (1987). The pope and the price of meat. Kyklos 40(fasc. 3): Besley, T., & Case, A. (1995). Does Electoral Accountability Affect Economic Policy Choices? Evidence from Gubernatorial Term Limits. The Quarterly Journal of Economics, Besley, T., & Coate, S. (2003). Elected Versus Appointed Regulators: Theory and Evidence. Journal of the European Economic Association, 1(5), Bolton, Patrick, and Gerard Roland. The Breakup of Nations: a Political Economy Analysis. The Quarterly Journal of Economics 112, no. 4 (1997): Born, P., & Viscusi, W. K. (2006). The catastrophic eects of natural disasters on insurance markets. Journal of Risk and Uncertainty, 33(1-2), Crain and McCormick, Regulators as all Interest Group, in Buchanan and Tollison(eds.), The Theory of Public Choice-II (Ann Arbor: University of Michigan Press, 1984). Posner, Richard A. The social costs of monopoly and regulation. Journal of Political Economy 83, no. 4 (1975): Devereux, Michael, and Simon Loretz. What do we know about corporate tax competition?. (2012). Devereux, M. P., Lockwood, B., & Redoano, M. (2008). Do countries compete over corporate tax rates?. Journal of Public Economics, 92(5), Feldstein, Martin S. Eects of taxes on economic behavior. No. w National Bureau of Economic Research, Fredriksson, P. G., Wang, L., & Warren, P. L. (2013). Party politics, governors, and economic policy. Southern Economic Journal, 80(1), Grace, M. F., Klein, R. W., & Kleindorfer, P. R. (2004). Homeowners insurance with bundled catastrophe coverage. Journal of Risk and Insurance, 71(3), Grace, M. F., Sjoquist, D.L., & Wheeler, Laura (2008). Eect of Insurance Premium Taxes on Employment. Fiscal Research Center, Georgia State University, FRC Report No. 184, September Grace, Martin F., & Robert W. Klein 2002). Natural disasters and the supply of home insurance. Harju, J., & Kosonen, T. (2012). The impact of tax incentives on the economic activity of entrepreneurs (No. w18442). National Bureau of Economic Research. Kahana, N., & Klunover, D. (2014). Rent seeking and the excess burden of taxation. European Journal of Political Economy, 35, Maloney, M.T. and McCormick, R.E. (1982). A positive theory of environmental quality regulation. Journal of Law and Economics, 25(April): Murphy, K. M., Shleifer, A., & Vishny, R. W. (1993). Why is rent-seeking so costly to growth?. The American Economic Review, 83(2), Pal, R., & Sharma, A. (2013). Endogenizing governments objectives in tax competition. Regional Science and Urban Economics, 43(4), Pittman, R. (1977). Market structure and campaign contributions, Public Choice, Springer, vol. 31(1), pages

17 Pittman, R. (1988). Rent-seeking and market structure: Comment, Public Choice, Springer, vol. 58(2), pages Pittman, R. (1976). The effects of industry concentration and regulation on contributions in three 1972 US Senate campaigns. Public Choice, 27(1), Richter, B. K., Samphantharak, K., & Timmons, J. F. (2009). Lobbying and taxes. American Journal of Political Science, 53(4), Rogan, P. (Ed.). (2013). The Insurance and Reinsurance Law Review. Law Business Research Limited. Shughart, W.F. II, Tollison, R.D., & Kimenyi, M.S. (1986). The political economy of immigration restrictions. Yale Journal on Regulation, 4(Fall): Stigler, G. J. (1971). The theory of economic regulation. The Bell journal of economics and management science, Stigler, G. J. (1974). Free riders and collective action: An appendix to theories of economic regulation. The Bell Journal of Economics and Management Science, Stock J, Yogo M. (2005). Testing for Weak Instruments in Linear IV Regression. In: Andrews DWK Identification and Inference for Econometric Models. New York: Cambridge University Press, Swank, D., & Steinmo, S. (2002). The new political economy of taxation in advanced capitalist democracies. American Journal of Political Science, Tullock, G. (1967). The welfare costs of tariffs, monopolies, and theft. Economic Inquiry, 5(3), The State of Florida s Property Insurance Market 2nd Annual Report (2013), The Florida Catastrophic Storm Risk Management Center. 16

18 Table 1: Summary statistics Variable Mean Std. Dev. Min. Max. N Market Concentration Effective Premium Tax Rate StateGDP (in $bn) Natural Disasters Unexpected Disasters Population (in mn) Campaign Contributions Appointed Regulator Democratic Governor Prior Approval File & Use Use & File

19 Figure 1: Scatterplot: The instrument and residuals from the model predicting effective premium tax rate 18

20 Table 2: OLS Results. Dependant Variable: ln (Premium Tax Rate) (1) (2) (3) (4) (5) (6) ln(marketconcentration t 1 ) (0.07) (0.07) (0.07) (0.07) (0.07) (0.07) Controls for regulatory environment Prior Approval (0.15) (0.15) (0.15) (0.15) (0.15) File & Use (0.21) (0.21) (0.21) (0.21) (0.21) Use & File (0.13) (0.13) (0.13) (0.13) (0.13) Campaign Contributions (0.00) (0.00) (0.00) (0.00) ln(gdp) (0.19) (0.19) (0.19) Appointed Regulator (0.03) (0.04) Democratic Governor 0.03 (0.02) constant (0.14) (0.18) (0.18) (0.91) (0.91) (0.90) N R Standard errors in parentheses are clustered at the state level and robust to heteroscedasticity p < 0.10, p < 0.05, p <

21 Table 3: First Stage Results. Dependant Variable: ln (Market Concentration) (1) (2) (3) (4) (5) (6) Unexpected Disasters (0.07) (0.01) (0.01) (0.02) (0.02) (0.02) Controls for regulatory environment Prior Approval (0.14) (0.14) (0.16) (0.16) (0.16) File & Use (0.12) (0.12) (0.13) (0.13) (0.13) Use & File (0.16) (0.16) (0.18) (0.18) (0.18) Campaign Contributions 1.6e e e e-08 (8.2e-09) (8.e-09) (7.8e-09) (7.9e-09) ln(gdp) (0.36) (0.36) (0.36) Appointed Regulator (0.05) (0.06) Democratic Governor (0.35) constant (0.82) (0.88) (0.86) (1.61) (1.63) (1.60) StateF ixedef f ects Y Y Y Y Y Y Y earf ixedef f ects Y Y Y Y Y Y N F-statistic Unexpected Disasters per capita is the unexpected disasters in the state normalized by the state s population; Campaign Contributions is the campaign contributions made by the insurance industry in statewide elections; ln(gdp) is the GDP of state in log values; Appointed Regulator is a dummy variable which takes the value of 1 if the Insurance Regulator in the state is appointed, 0 otherwise; Democratic Governor is a dummy variable which takes the value of 1 if the Governor of the state is affiliated to Democratic party, 0 otherwise; Prior Approval, File & Use, Use & File is a dummy variable which takes the value of 1 if a state has a Prior Approval, File & Use and Use & File system of insurance premium regulation respectively, 0 otherwise. Errors in parentheses are clustered at the state level and robust to heteroscedasticity p < 0.1, p < 0.05, p <

22 Table 4: Reduced Form estimates. Dependant Variable: ln (Effective Tax Rate) (1) (2) (3) (4) (5) (6) ln(marketconcentration t 1 ) (0.23) (0.27) (0.26) (0.28) (0.28) (0.27) Controls for Regulatory Environment Prior Approval (0.14) (0.14) (0.15) (0.15) (0.14) File & Use (0.20) (0.20) (0.20) (0.20) (0.20) Use & File (0.12) (0.12) (0.12) (0.12) (0.12) Campaign Contributions 1.1e e e e-08 (8.0e-09) (8.1e-09) (8.3e-09) (8.0e-09) ln(gdp) (0.24) (0.24) (0.24) Appointed Regulator (0.04) (0.04) Democratic Governor 0.02 (0.02) constant (0.82) (0.88) (0.86) (1.61) (1.63) (1.60) StateF ixedef f ects Y Y Y Y Y Y Y earf ixedef f ects Y Y Y Y Y Y N R ln(market Concentration is the log of market concentration in the insurance industry in a state; Campaign Contributions is the campaign contributions made by the insurance industry in statewide elections; ln(gdp) is the GDP of state in log values; Appointed Regulator is a dummy variable which takes the value of 1 if the Insurance Regulator in the state is appointed, 0 otherwise; Democratic Governor is a dummy variable which takes the value of 1 if the Governor of the state is affiliated to Democratic party, 0 otherwise; Prior Approval, File & Use, Use & File is a dummy variable which takes the value of 1 if a state has a Prior Approval, File & Use and Use & File system of insurance premium regulation respectively, 0 otherwise. Errors in parentheses are clustered at the state level and robust to heteroscedasticity p < 0.1, p < 0.05, p <

23 Table 5: Results without logs (Reduced Form) (First Stage) Tax Rate Market Concentration MarketConcentration t (0.02) - Unexpected Disasters (1.96) Other variables Y Y cons (0.52) (2.4) State&Y earf ixedef f ects Y Y N R F statistic 87 Errors are clustered at the state level and robust to heteroscedasticity p < 0.1, p < 0.05, p <