The recent financial crisis has put a strain on public finances all over the world. In the US, the deficit reached over 12% of GDP in 2009 and 2010, the highest level in nearly 65 years, and public debt jumped from 64% of GDP in 2008 to 104.8% of GDP in 2014. In the EU, public deficits increased from close-to-equilibrium in 2007 to an average of 6.8% of GDP by 2009, and the average public debt increased from 57.6% to 88.6% of GDP by 2014.1 Part of the deterioration in public finances can be attributed to output and revenue losses due to the financial crisis. Laeven and Valencia (2013) have studied 147 episodes of banking crises between 1970 and 2011 and find that for the crises that started from 2007 onward, the median output loss reaches 25% of GDP. In addition the large public support to the financial sector contributed to debt increases in many countries. During the period 2008-2012, recapitalisation of financial institutions and asset relief measures taken by EU governments amounted to close to 5% of EU GDP, with figures up to 40% GDP in individual countries (e.g. Ireland).
Implementing policies to reduce the probability of crises, and to better manage them when they nonetheless occur, are therefore of paramount importance. Many regulatory initiatives such as the new requirements under the Basel III framework, the Dodd-Frank Act in the US or the Banking Union in the EU are policies meant precisely to address these issues.2 Strengthening the equity base of the financial sector is a key element of these regulatory initiatives. Recent studies suggest that tax policy could play a significant complementary role in this regard.
Corporate taxation leads to over-indebtedness in both the financial and non-financial sector
In most countries of the world, corporate tax systems usually favour debt over equity by allowing for tax deductibility of the cost of debt (i.e. interest paid), while the return to equity is not considered a deductible expense. This asymmetric tax treatment of different sources of finance at the corporate level – the ‘corporate debt bias’ – originates from historical conventions and does not have any economic rationale.
The effects of corporate taxation on the capital structure of firms are empirically established. In a meta-analysis study synthesizing evidence from about 50 previous studies, Feld et al (2013) predict that each one percentage-point increase in the corporate tax rate increases the debt-to-assets ratio by 0.27 percentage-points. The authors also find this effect to be large in comparison to other determinants of capital structure, reflecting large tax incentives of financing corporate projects by debt. Recently, the EC (2015) stressed that "differences in the tax treatment of debt and equity financing might increase the reliance of companies on debt and bank funding". In addition, there is empirical evidence of larger effects for intra-group debt (Desai et al 2004) and of the use of debt to shift profit (Huizinga et al 2008).
Despite regulations on capital requirements, the financial sector is not immune from incentives towards indebtedness. Heckemeyer and de Mooij (2013) review the existing literature and find that, with a few exceptions, empirical studies find on average a similar responsiveness of leverage to tax incentives for the financial sector as for non-banks.
The presence of a debt bias is clearly at odds with the regulatory objectives of strengthening the banking sector through higher equity capital. Considering that the financial sector is highly leveraged and that excessive indebtedness of banks generates externalities linked to systemic risk, ending the preferential tax treatment of debt would improve the stability of the financial sector and avoid potentially large welfare losses.
Removing the debt bias could lead to substantial public finance savings
De Mooij et al (2014) have made a first attempt to establish a link between the tax incentives encouraging indebtedness and the likelihood of financial crises. Using a two-step approach, they translate the effects of removing the debt bias on indebtedness and hence on the probability of a crisis. Their results show larger effects, the higher the initial leverage ratio. Translating the decrease in the likelihood of a financial crisis into potential GDP gains,3 eliminating the debt bias could lead to potential GDP gains of between 0.5 and 11.9%.
We estimate the benefits for financial stability of eliminating the corporate debt bias in two steps (Landgedijk et al 2015). First, we assess the long-run response of leverage to corporate income taxation for a large sample of EU banks. We find that a one percentage point increase in the statutory corporate income tax rate leads to an increase in the leverage ratio of 0.2 to 0.3 percentage points. This estimated range of the elasticity of leverage to taxation of 0.2 to 0.3 is consistent with previous literature. Second, we use these estimates in a Vasicek-based model with bank actual balance sheets to simulate the effects of removing the debt bias on the costs of systemic crises for six large EU Member States.
Our model is based on the Basel formula for the foundation internal-ratings based (FIRB) loss distribution.4 This formula is used by many banks to calculate their minimum capital requirements based on the probability of default, the loss given default (i.e., the average loss expected on a defaulted loan), the size and the maturity of individual loan portfolios. By inverting this formula and using actual balance sheet data of banks, we can obtain the average implied probability of default of the obligors of each bank, assuming that the banks’ risk models adequately reflect the credit risk in their portfolios. By Monte Carlo simulations using the estimated probability of default in the FIRB loss distribution we generate a sample of loan losses.
Figure 1. Loss distribution and default of banks
Losses are first absorbed by the bank provisions (Pr), then by the minimum required capital (MCR) and further losses could be absorbed by excess capital (EC) above the regulatory minimum, when available. A bank defaults when the portfolio losses cannot be absorbed by provisions and total capital. These losses in excess of capital constitute potential recapitalisation needs.5 Summing them over all banks in the system gives the estimated public finance costs.
To simulate the effects of eliminating the debt bias, we use long-term estimates from our regressions and empirical literature to compute the effect of this policy on the capital-to-assets ratio and plug this into our bank loss model. Our results suggest that even if the tax elasticity of bank leverage is taken at the lower end of the ranges estimated in empirical literature,6 eliminating the debt bias could lead to drastic reductions of public finance losses. Taking our lowest elasticity of leverage to taxation at 0.05, fully removing the debt bias reduces potential public finance losses by between 25 in Spain to 55% in France. For the full range of elasticities [0.05-0.20], potential public finance gains in a financial crisis of the magnitude of the 2008-2012 crisis range between 0.5 and 11.2% GDP, across countries.
Figure 2. Losses as share of GDP reduction due to debt bias elimination
Our results hold for capital-tight banks and for an endogenous increase in portfolio risk
Not all financial institutions are similar in terms of their fulfilment of minimum capital requirements. ‘Capital-tight’ banks have very small buffers to the minimum regulatory requirements while ‘capital-abundant banks’ may optimise their debt at levels that easily fulfil these requirements. Keen and de Mooij (2012) find that the sensitivity of debt to tax is lower for the former than for the latter. To account for this, we rerun our simulations with low CIT effects (zero to 0.1) for capital-tight banks and larger CIT effects for capital-abundant banks (0.2 to 0.3). In countries where a large majority of bank assets are located in banks that have sizeable buffers above the minimum requirements (such as the UK and the Netherlands7), removing the debt bias could reduce the costs for public finance by 80% or more. Interestingly, in countries with a lot of capital-tight banks, such as Spain, removing the debt bias could still reduce potential public finance costs by 30%, even for a very low tax elasticity of 0.05 for capital-tight banks.
Another element that we take into account is the possibility for banks to increase the risk of their assets in conjunction with the increase in capital (Devereux 2014). We use estimates from the literature on the effects of CIT on Risk-Weighted Assets (zero to 0.2). Including this effect into our model does not dramatically alter our results with a range of loss reductions between 20 and 70% across countries using a conservative long-term CIT effect on bank leverage of 0.1.
The case for policies aiming at eliminating the corporate debt bias
Remarkable achievements have been recorded in improving the regulatory framework of the financial sector. The policy goals are to reduce the risk of new crises and limit their costs, should they unfortunately occur. Regulation alone could go a long way towards achieving those goals and reducing the burden of recapitalisations and asset relief programmes on public finances. There is, however, mounting evidence that tax policy can play in the same direction and provide additional incentives towards lower debt. The possible solutions have long been debated and include the Comprehensive Business Income Tax, the Allowance for Corporate Equity, the Allowance for Corporate Capital or even a move to Cash-Flow taxation. All have pros and cons in terms of tax collection, incentives for investment, or countering profit-shifting but they share the desirable characteristic that they can all drastically reduce the potential public finance costs of future crises and promote the creation of a vibrant Capital Market Union.
Authors' note: The views expressed herein are those of the authors and should not be attributed to the European Commission.
De Mooij, R, M Keen and M Orihara (2014) "Taxation, bank leverage, and financial crises", chapter 11 in Taxation and regulation of the financial sector, de Mooij and Nicodeme (eds), MIT Press.
Desai, M, F Foley and J Hines (2004) "A multinational perspective on capital structure choice and internal capital markets", Journal of Finance, 59: 2451–2487.
Devereux, M P (2014) "New bank taxes: Why and what will be their effects", chapter 2 in Taxation and regulation of the financial sector, de Mooij and Nicodeme (eds), MIT Press.
EC (2015) "Green paper – Building a capital market union", COM(2015)63.
Feld, L P, J H Heckemeyer and M Overesch (2013) "Capital structure choice and company taxation: A meta-study", Journal of Banking and Finance, 37: 2850-2866.
Heckemeyer, J and R de Mooij (2013) "Taxation and corporate debt: Are banks any different?", IMF Working Paper, 13/221.
Huizinga, H, L Laeven and G Nicodeme (2008) "Capital structure and international debt shifting", Journal of Financial Economics, 88(1): 80-118.
Keen, M and R de Mooij (2012) "Debt, taxes, and banks", IMF Working paper, 12/48.
Laeven, L and F Valencia (2013) "Systemic banking crises database", IMF Economic Review, 61: 225-270.
Langedijk, S, G Nicodeme, A Pagano and A Rossi (2015) "Debt bias in corporate income taxation and the costs of banking crises", CEPR Discussion Paper 10616.
Vacisek, O A (1991) Limiting Loan Loss Probability Distribution, KMV Corporation.
Vasicek, O A (2002) "The distribution of loan portfolio value”, Risk, December, 160-12.
1 Ameco database
2 Since 2008, the EC has proposed more than 40 legislative and non-legislative measures to strengthen the global financial system.
3 The authors take a mean output loss of a financial crisis of 23% for the period 1970 - 2009. This is measured as the cumulative loss in GDP relative to the pre-crisis trend over a four-year period.
4 Adjustments are introduced to take into account the impact of the introduction of Basel III on risk-weighted assets (RWA) and regulatory capital. These are implemented using the average EU results of the 2014 Comprehensive Quantitative Impact study (C-QIS) by the European Banking Authority. These adjustments imply increased RWA, and a more strict definition of regulatory capital.
5 In this analysis we do not consider bail-in of bondholders or the use of other safety nets such as resolution funds or deposit guarantee schemes. We do not consider direct interbank contagion either. Indirect contagion is considered by positively correlating portfolio losses across banks in the Monte Carlo simulations.
6 Across empirical studies, estimates of long-term effects of a one percentage point increase in corporate income taxations on the ratio of total bank capital over total assets ranges from a 0.08 to 0.34 percentage point decrease. For our simulations of the effects of eliminating the debt bias we report on a range of 0.05 to 0.20.
7 The terms ‘capital-abundant’ and ‘capital-tight’ refer only to the buffer above the risk-weighted minimum capital requirement to allow reflecting the findings in the empirical literature that banks with a small capital buffer to the regulatory minimum are less responsive to tax changes. While banks in the UK and in the Netherlands may have relatively large buffers vis-à-vis the risks-weighted minimum capital requirements, the unweighted total capital ratios may be low.