A POLITICAL ECONOMY APPROACH TO BANKING REGULATION

Transcription

1 A POLITICAL ECONOMY APPROACH TO BANKING REGULATION José Alonso Olmedo Master Thesis CEMFI No December 2012 CEMFI Casado del Alisal 5; Madrid Tel. (34) Fax (34) Internet: This paper is a revised version of the Master's Thesis presented in partial fulfilment of the Master in Economics and Finance at the Centro de Estudios Monetarios y Financieros (CEMFI). I would like to thank my advisor Javier Suárez for his patience, guidance and for initiating me in the research activity. This Master's Thesis would not have been possible without him. I am also grateful to all the CEMFI faculty and staff and to my classmates and colleagues for all their helpful comments, support and all what they have taught me in these two years. Errors and omissions are exclusively my own responsibility.

2 Master Thesis CEMFI No December 2012 A POLITICAL ECONOMY APPROACH TO BANKING REGULATION Abstract This paper presents a theoretical framework to analyze the regulation of the banking sector from a political economy perspective. To do it, we combine a pressure group model in which different lobbies can make political contributions to influence the decision of the regulator and a model of banking regulation in which the regulator uses a capital requirement as instrument. We also analyze quantitatively the extent to which the pressure of the different interest groups can lead policy outcomes to deviate from the social optimum. José Alonso Olmedo NERA josealonsoolmedo@gmail.com

3 1 Introduction The truth is that globally and nationally we should have been regulating banks more - Gordon Brown (2010). The financial crisis which has plunged us in the current economic scenario has shown that prudential regulation in the banking system was flawed. This issue has become very relevant and it has attracted the interest of the general public. The controversy around the financial system and its companies has been one of the most discussed topics in the media all around the world during the crisis. The influence that the banking sector can exert in this debate has found the response of a part of the public opinion which demands a tougher regulation on banks. Different positions have been supported in this debate and politicians in many countries have had to take a stance on it. The crisis has opened a process of regulatory reforms. The design of the global standard of Basel III and the approval of the Dodd-Frank Act in the US are examples of this process. Academics have participated actively in this debate with proposals and analysis. This explains the rich literature which has focused on the possible improvements which can be done in the area of banking regulation and the very enriching debate generated with it. In spite of the abundant literature about banking regulation, there is little work about the influence of politics on the regulatory process. A political economy approach to this subject is missing. To follow this approach, we will focus on the analysis of the systemic risk generated by the banking activity. This will be our main connection with the political economy literature. As it is explained by Perotti and Suarez (2011), systemic risk can be viewed as a negative externality which establishes a natural link to normative economics and the economics of regulation. 3

4 Models of regulatory capture, like the one of Laffont and Tirole (1993), or those with pressure groups, as the one developed by Grossman and Helpman (1994), evaluate in a simple way how the political process needed for regulation can have an important impact on the equilibrium outcomes in social welfare terms. These models provide us with a tool to understand the reasons why actual regulation and its outcomes may differ from the optimal one and the social welfare maximizing outcomes. Aidt (1998), Groosman and Helpman (1994) and Fredriksson (1999) presented political economy models of regulation in the areas of trade and environmental policy where international coordination is an important issue. When we talk about trade, the economic agents are affected in very different ways by the policies. This creates incentives for these groups to try to influence the decision of the regulator. This situation is the same when we talk about environmental policy. Like in these cases, banking regulation has an important international dimension, an impact in the wellbeing of different agents in a way which can create important conflicts of interests and, as in the analysis of environmental policy, it includes the existence of externalities which makes the study richer. The objective of this Master thesis is to develop a theoretical model to analyze the regulation of the banking system from a political economy perspective. The analysis would serve to better understand the regulation existing before the crisis, its possible evolution after it and the forces which lead to the shift. The analysis might also deliver policy implications regarding the potential gains from creating an independent regulatory agency or intensifying the vigilance of the relationship between politicians and the banking lobbies. The political economy part of the analysis will follow Grossman and Helpman (1994). We will focus in the case in which different lobbies can exert their influence in the regulatory process. This approach makes sense if we think that different interest groups can be affected by banking regulation and that the regulator will deal with the pressure exerted by them. In the Appendix we explain how this framework can be implemented in this context. 4

5 We will combine this pressure group framework, in which different lobbies can make political contributions to influence the decision of the regulator over the capital requirement, with a banking regulation model. In the banking model we will have different groups. A continuum of firms will choose among two types of financing: banking or market funding. This decision will depend on the easiness of access of each firm to the financial market and will divide firms in two groups with opposing regulatory interests. The picture of possible lobbies is completed by the consumers in the market in which these firms compete and the bank managers. All these potential lobbies will have different policy interests, we will analyze how the interaction among them can lead to outcomes which are different from the social optimum and the forces behind this process. Then, we will use a baseline calibration to analyze quantitatively those deviations from the optimum. One first insight from the analysis is that what is important to determine the size of the influence of a group in the political economy interaction is not the weight of its own welfare in the total social welfare but the size of the changes that policy variables can cause on the group s welfare.. What matters is the importance of this policy for the lobby and this determines its willingness to pay (put pressure on) the regulator to move policy in its favor. Under the baseline calibration, the two lobbies of the model whose interaction has the greatest influence on the final policy outcome are the two groups of firms. market-financed firms push for a high value for the capital requirement to obtain more rents in the final market thanks to the lower competiton coming from the bank-financed firms. Bank-financed firms push in the opposite direction. The rest of the paper is organized as follows. Section 2 presents a model of banking which tries to build a rich theoretical framework to show how different forces can have an influence in the regulation of the banking system. Section 3 includes several quantitative results which can shed light on the importance of these influences in the final outcome. Finally, Section 4 discusses implications and possible extensions of this work and concludes the paper. 5

6 2 The model To pursue the goal of this paper we will try to combine a model of banking regulation with the political economy approach developed by Grossman and Helpman (1994). We will have a regulator which uses capital requirement as regulatory instrument in this market and different interest groups which make contributions to influence the decision of the regulator in the direction which benefits them. We can interpret these political contributions as campaign contributions but also as offers for future jobs or other type of incentives that may have an impact on the final choice of the regulator. A first approach could be to develop a simple model in which we have a banking environment characterized by deposit insurance, a representative bank and systemic risk as a social cost generated by bank failure. In such a framework we can introduce a moral hazard problem through the decision made by the bank about the risk undertaken in its search for shareholder value maximization. In Appendix B we show in detail the structure of such a model. The deposit insurance paid by the taxpayers, the distortions generated by these taxes and the systemic risk associated to the banking activity make optimal from a social welfare point of view to introduce a positive capital requirement. However, the bank is better-off if no regulation is set. If we assume that the unique lobby with enough power in this model to make political contributions to the regulator is the financial one, represented here by the bank, it is clear that the final policy outcome will be between the social optimum and the one desired by the bank. This leads to a capital requirement in equilibrium which is lower than the one which maximizes the social welfare. This result is very interesting. In the last years, with the dawn of the crisis, many voices have claimed that the banking system had a inadequate regulation. If we attend to this simple model, we can think that this insufficient regulation could be the result of having only one important interest group exerting its influence, the banking lobby. A scenario for the period previous to the crisis in which only the 6

7 banks were using their power in the regulatory process sounds plausible. This lobby would have put pressure to obtain a lax regulation. According to this, in the last years the regulation could be changing for several reasons. On one hand, different interest groups could have appeared in this process due to their increased awareness about the possibility of having an influence on a process that, after the crisis, occupies a much more central space in the public debate. On the other hand, the weight that the regulator gives to total social welfare (vis-a-vis the potential contributions of the interested parties) might have changed due to the mobilization of the public opinion. According to this view, the crisis would have produced a shift in the mentality of the regulators and the governments which, now, would give less power to the banking lobby. This would give them more social support or votes in return for moving regulation closer to the social optimum. Although the previous model yields interesting results, it is obvious that the political economy analysis derived from it is very simple and easily predictable. In what follows, we try to take advantage of the possibilities offered by this approach through a richer framework that allows for more interested parties (lobbies) in the regulatory process. 2.1 A banking regulation model The starting point will be a model with two periods (t = 0, 1) in which we assume risk-neutral agents and a risk-free asset with return r. There is a continuum of firms characterized by a parameter θ [0, 1] which follows a distribution with density f(θ) and c.d.f. F (θ). Each entrepreneur can invest in a project with gross returns: R = { A with probability p 0 with probability (1 p) If the project is successful, the firm can sell a unit of product in the final market at a price A. There are two financing possibilities and the probability of success p 7

8 depends on the funding option chosen by the firm. If the entrepreneur decides to use market funding, p is equal to the value of its parameter θ while if the firm is funded by the representative bank of our economy p is equal to a value q which is the expected value of a random variable q [0, 1] independent from θ and distributed with a density g(q) and a c.d.f. G(q). This random variable determines the proportion (1 q) of failed loans given by the representative bank to the firms. To analyze the financing decision of the firm we have to evaluate both alternatives: Market funding: Market financiers are promised an amount B 0. Given that the required expected rate of return in the market is (1 + r), the repayment at t=1 should satisfy θb 0 = (1 + r). Thus, with this financing option the expected profits generated by a firm are: π M = θ(a B 0 ) = θa (1 + r), (1) where A is the expected price in the final market which, as we will see later, depends, among other elements, on the expected value q. Bank funding: In this case we assume that this financing implies a monitoring activity which has a cost m for the bank. This amount is paid to the bank manager to induce his proper behaviour and it is contingent on the success of the projects. The bank has two sources of funding for its activity: capital k and deposits 1 k. The difference between these two options is that capital is associated to an expected rate of return 1 + r + δ. In this case δ can be interpreted as an exogenous deadweight loss from a social welfare point of view or as an scarcity rent obtained by the providers of equity capital due to the existence of a fixed capital supply K. Under this assumption, capital owners can obtain higher profitability from their investments. This is reasonable if we believe that they have greater information or skills to operate in the financial markets. 8

9 In this financing possibility we introduce a source of negative externalities as a consequence of bank failure. This appears as a social cost equal to a fraction λ of the assets of the failed bank. There is a positive likelihood of bank failure given by the random variable q. Thus, as we have mentioned above, (1 q) is the proportion of failed loans which is determined according to the distribution with density g(q) and c.d.f. G(q) where E(q) = q. We will asumme that deposits are uninsured and that the extra cost of the capital financing δ is given exogenously 1. In these circumstances, depositors will be promised a repayment D 1 which should satisfy that their expected return is equal to the one that they can obtain in the market (1+r). This implies: E [min {D 1, q(b 1 m)}] = (1 + r)(1 k). (2) This equation means that depositors will get D 1 if q D 1 B 1 because the bank m has sufficient returns not to fail. However, if the bank fails they will obtain the remaining quantity that the bank has after paying the manager. Following this reasoning we can rewrite (2) as: D 1 B 1 m 0 ( )] D1 q(b 1 m)g(q)dq + D 1 [1 G = (1 + r)(1 k). (3) B 1 m bank-financed firms should make a promised repayment B 1 which satisfies that the expected return obtained by the bank shareholders is equal to the one that they can obtain in the market as capital owners. This means that B 1 is defined as: E [max {q(b 1 m) D 1, 0}] = (1 + r + δ)k. (4) 1 There are different specifications of the model in which we include deposit insurance and an endogenously determined δ which show that many of these results are robust to these changes. 9

10 If we add (2) and (4) we obtain: E [q(b 1 m)] = (1 + r)(1 k) + (1 + r + δ)k q(b 1 m) = 1 + r + δk, (5) which leads to B 1 = m + 1+r+δk, and that the expected profits obtained by the q firms that use bank funding are: π B = E [q(a B 1 )] = E [Aq] (1 + r) δk mq. (6) Given these results we can check that the entrepreneurs will decide to use bank financing if and only if: π B π M E [Aq] (1 + r) δk mq θ(a B 0 ) = θa (1 + r) E [Aq] δk qm θ = θ(k). (7) A This means that only firms with low θ and, thus, with worse possibilities in the financial market would use bank funding. In addition, we can see that the threshold θ determining the demand of loans is decreasing in capital k due to the extra costs that should be paid to the capital owners. 2.2 Equilibrium and social welfare components The existence of externalities caused by bank failure introduces a motivation for capital regulation. To analyze the optimal decision of the regulator and, later, the outcomes obtained when we introduce the political economy framework we have to understand the payoffs obtained by each group. In this context, we assume that all the firms compete in a final market of products where the supply is determined by the expected price through the proportion of firms which decide to use each type of funding. The expression of the supply is: y S = 1 θ θf(θ)dθ + F ( θ)q, (8) 10

11 which is increasing in the expected price A. The higher expected profitability of this price increase makes the threshold θ to decrease more firms are capable of using market funding. This leads to a lower proportion of firms using bank funding. The supply is reduced because the probability of success under this type of financing is higher than the one under market funding as we will explain later. Finally the demand is assumed to be of the form: P (y) = A = P 0 P 1 y y D = P 0 A P 1. (9) Market competition is important because while it is clear that bank-financed firms will be affected by the capital requirement regulation, those firms which are marketfunded will be interested in the capital requirement set by the regulator only due to this competition in the final market. The regulation will affect the productivity of their competitors which use bank-financed and, thus, the outcomes in the final market and their profits. The next step is analyzing the objective function of each group. To do it, we need a an expression for for E [Aq] which can be developed from the definition of the demand and the supply in the final market given by (8) and (9): ( θ) E [Aq] = E [(P 0 P 1 y) q] = P 0 q P 1 E[q 2 F + q 1 θ ] θf(θ)dθ. (10) As we know that E[q 2 ] = q 2 +σ 2 q where σ 2 q = V ar(q) and A = P 0 P 1 ( 1 θ θf (θ) dθ + F we can express (10) as: E [Aq] = Aq P 1 σ 2 qf ( θ). (11) The term P 1 σ 2 qf ( θ) can be interpreted as a Jensen s inequality term for this expectation. With this result we can now express the expected profits obtained by bank-financed firms (6) as: ( θ) π B = Aq P 1 σqf 2 (1 + r) δk mq. (12) ( θ) ) q 11

12 of: Thus, the market-financed firms would obtain an expected total aggregate amount Π M = 1 θ [ ] Aθf(θ)dθ (1 + r) 1 F ( θ), (13) while the bank-funded firms would receive as expected aggregate profits: [ ] Π B = F ( θ) Aq P 1 σqf 2 ( θ) (1 + r) δk mq, (14) where σ 2 q is the variance of the random variable q. The main differences among both quantities are the different costs associated to each of them and the different probabilities of success q and θ. From the definition of θ(γ) we can see that the probability of success under bank financing is, in the margin, higher than the one under market financing due to the existence of extra costs (manager payment and the additional cost of capital δ) which is compensated through a higher probability q. This means that firms are more productive under bank financing. In addition, we know that the bank manager receives an expected amount: U BM = F ( θ)mq, (15) which is determined by the proportion of firms using bank funding and the expected probability of success of the loans. Finally, given the form of the demand in (9), the consumer surplus will be CS = y2 (P 0 A) 2 = P 1 2 y 2. Using this, we can express the expected consumer surplus as: U C = P 1 2 E [ y 2] = P 1 [ ] y 2 + σy 2, (16) 2 where σy 2 is the variance of y which can be obtained from its definition y = 1 θ θf(θ)dθ+ F ( θ)q. Given that in this expression q is the unique random variable, we can obtain ) 2. σy 2 = σqf ( θ 2 With this result, an expression for the expected consumer surplus is: U C = P 1 2 E [ y 2] = P [ ( θ) ] 2 1 y 2 + σ 2 2 qf. 12

13 This implies that consumers will obtain an expected surplus: U C = 1 2 P 1 [ ( 1 θ ) 2 ] ) 2 θf(θ)dθ + F ( θ)q + σqf ( θ 2. (17) The last term is reflecting the extra surplus obtained by the consumers due to the existence of the bank financing which leads to a higher expected productivity through those firms using this kind of funding. We should also consider the social cost generated by the externalities. We have seen above that the bank will fail when q < D 1 B 1 = q. Knowing this and that the m proportion of firms using bank funds is given by F ( θ) we postulate that the welfare loss generated by the bank failure is λf ( θ)g( q) 2 - where λ denotes the proportion of assets lost with the bank failure. In this context we have to notice that due to the fact that capital funds are more expensive than deposits, the bank would choose to use only deposits in the financing of its activities. However, the regulator can impose a capital requirement γ which due to the previous reasoning will be binding. A benevolent regulator will choose the optimal policy γ which maximizes the total welfare which is obtained aggregating the five elements explained above: W (γ) = Π M (γ) + Π B (γ) + U BM (γ) + U C (γ) λf ( θ)g( q) 2. Before analyzing the capital requirement chosen by the regulator, it is interesting to check how such a policy can affect the different groups. To do it we can evaluate how their payoffs change with γ. 2.3 The effect of the capital requirement on social welfare and its components The first thing that we should understand is the effect of the capital requirement γ on the price in the final market A and on the threshold value θ. These two ef- 13

14 fects are important because, as we have mentioned before, we could think that the market-financed firms would be indifferent about the policy but the competition existing between them and the bank-financed firms can make those entrepreneurs to desire higher capital requirements, increasing the cost of bank funding and reducing the competitiveness of their competitors. We know that a change in the capital requirement will produce changes in the relative cost of each type of funding. This will lead to changes in the proportion of firms using each option, the productivity and the price in the final market, and the aggregate payoffs of each group. As all the interest groups will be affected by changes in the capital requirement, we will have potential lobbies that can interact with the introduction of the political contributions generating conflict among them. This will give us the rich framework that we are looking for to analyze this regulatory process from a political economy approach. To analyze the effect of the capital requirement we start by identifying how the expected price in the final market A and the proportion of firms using each type of funding, given by the threshold θ, change with γ. Using (8) and (9) we can obtain the expected equilibrium condition for the final market: 1 θ θf(θ)dθ + F ( θ)q = P 0 A P 1. (18) With the implicit definition for θ obtained in Appendix C.1, ) ( θ) A ( θ q + P 1 σqf 2 + δk + mq = 0, (19) we can use the implicit function theorem to obtain δa δ θ > 0 and < 0 (see Appendix δγ δγ C.3 for details). Thus, an increase in the capital requirement reduces the proportion of firms which use bank funding and increases the proportion of firms which use market funding. The increase in γ makes more expensive for banks to obtain financing and this increase in costs is transferred to the entrepreneurs which demand loans. Because of this, some firms change their financing strategy and they start to use market financing which is now relatively cheaper. 14

15 This gives us the opportunity to understand better the effect of the capital requirement γ on the price of the final products. An increase in γ reduces the proportion of firms which use bank funds F ( θ(γ)). As firms are more productive under bank financing, the reduction in θ produced by a higher capital requirement reduces the production of the final good because a higher proportion of firms use market financing. This fall in production raises the price. To analyze the interest of each lobby in the capital requirement and in a possible change of it we should also know how the payment B 1 promised to the bank by those firms which use bank loans and the repayment promised to the depositors D 1 by the bank change with γ. From the expression B 1 = m + 1+r+δγ obtained from (5) we can q see that δb 1 > 0. This means that the higher is the capital requirement, the higher δγ will be the repayment promised by the firms to the bank. This has sense as higher requirements imply a higher financing cost for the bank which should be compensated through higher revenues. In other words, as more capital is used by the bank, it has to compensate a higher proportion of shareholders for the opportunity cost δ that they can obtain in the market and these additional revenues needed by the bank are asked to the firms which pay them through B 1. We could also explain why from (3) we obtain δd 1 < 0. Due to the reduction in δγ risk failure generated by the increase in the capital requirement the repayment asked by depositors is lower. Once we have all these effects that show us how the different variables of the model change with γ, we can try to evaluate how changes in the capital requirement affect the different groups of our economy. This is important because the policy interest of each group shows the cue to understand in which direction they will push with their contributions in the political economy framework. An easy starting point is to think about the bank manager. Given his expected payoff expressed in (15), a change in γ would produce the effect: 15

16 U BM γ = f( θ) θ mq < 0. γ As the expected payoff of the manager depends on the capital requirement through the proportion of firms that decide to use bank funding to finance their activities, the manager would like to reduce the capital as much as possible. An increase in γ reduces the proportion of firms which use bank-financed as we have explained above, this decreases the revenues that the bank can generate and, thus, the remuneration received by its manager. Because of this, we expect that the manager will try to induce the regulator to lower the capital requirement. The effect of the capital requirement on the negative externalities generated by the failure of the bank is given by how the costs λf ( θ)g( q) decrease with γ: U E γ = λ [ f( θ) θ G( q) + 2F ( θ)g( q)g( q) γ D 1 (B ] B γ 1 m) D 1 1 γ (B 1 m) 2 > 0. In this case we can see clearly two effects. First, if the capital requirement is increased, the proportion of firms which use bank-financed decreases and, as banking activity is reduced, the systemic risk generated by it is lowered. Second, we can see in the last element of the second term how that increase in γ decreases the threshold q. We have seen that for all the values of q < q the bank fails, which means that a lower value for this critical value implies less probability of failure. Given this, to reduce the negative effect of the externality the regulator should try to increase the capital requirement. This is the main motivation to introduce this policy. We can also study how this capital requirement affects market-financed firms from: Π 1 M γ = A θ [ ] θf(θ)dθ f( θ) A θ (1 + r) > 0. θ γ γ In this case, the interest group will try to pressure to increase the capital requirement γ due to two circumstances. First, we have seen that a higher γ implies a higher price A in the final market. Because of this, all the firms would like a higher capital requirement to take advantage from that higher prices. In addition, an increase in γ implies a higher proportion of market-financed firms which means that the aggregate 16

17 profits of this group would increase. Π B γ The case of bank-funded firms is more complex: [ ] θ [ = f( θ) Aq (1 + r) δγ qm P1 σ 2 A γ qf ( θ)]+f ( θ) γ q δ P 1σqf( θ) 2 θ. γ The sign of this expression is ambiguous. On the one hand, these firms would like a low capital requirement because this would imply a cheaper repayment to the bank shareholders and, thus, cheaper financing. In addition, we have again the effect through θ which affects the proportion of firms using each type of funding and the aggregate profits. For bank-financed firms this effect goes in the opposite direction to that of the market-financed firms reinforcing the desire for a lower capital requirement. To have an interest in reducing the capital requirement these two forces should compensate the positive price effect explained previously. As we are assuming perfect competition in the banking sector we can think that this lobby is representing all the rents obtained, not only by these firms, but also by the banking activity. In this case it is reasonable to think that the price effect obtained by these firms, that are those firms with lower possibilities in the financial markets or those small firms, is not enough to compensate the interests of the bank shareholders. Following this reasoning we can also interpret the effect through the threshold θ as an effect on the surplus of the banking sector because this threshold is determining the importance of this sector as the proportion of firms using their services. Finally, we focus on the surplus of the consumers and we observe that the change in γ leads to: [ U 1 ] [ C γ = P θ ( 1 θf(θ)dθ + F ( θ)q θ γ f( θ) q θ ) ] θ + P 1 γ f( θ)f ( θ)σ q 2 < 0. Consumers would like a low capital requirement because this would decrease product prices, which would increase their surplus. 17

18 3 Quantitative results To obtain quantitative results we should give values to the different parameters. Table 1 shows the values assigned to the different parameters in this analysis. Table 1: Baseline parameter calibration Risk-free asset return r 0.04 Excess cost of bank capital δ 0.04 Bank manager remuneration m 0.01 Distribution of firms success probability under market funding θ u[0.78, 1] Expected success probability under bank funding q 0.9 Distribution of the aggregate loan default rate q u[0.8, 1] Final market demand parameters P 0, P 1 3.4, 2.35 Cost of bank failure λ 0.15 Weight assigned to social welfare by regulator β 0.2 The idea behind this calibration is to illustrate the working of this model. To do it, we assign values to the parameters that lead to results which are in a range which can be viewed as compatible with what we can see in reality. For example, we use values for m, r, q and δ which can be considered as reasonable and consistent with what we can find in other studies (see Repullo and Suárez (2004) and (2012)). The rest of the parameters are calibrated to obtain several results in equilibrium. For the demand in the final market we set values to reach two goals. First, the calibration for P 0 and P 1 makes the bank-financed firms to obtain a 5 per cent expected profitability which is consistent with the working of the model as these firms have a higher expected profitability than the one they can obtain with investments in the risk-free asset. Second, with these values we have in equilibrium an elasticity of the demand which is around 0.5. The value of λ is set to obtain a social optimum capital requirement of 10 percent. This leads to a value which can be considered a bit low but it is, again, in a range which is realistic and useful as a benchmark. The distribution of the parameter θ 18

19 which characterized the firms is set to match the proportion of firms which uses bank financing in the United States - approximately a 50 per cent. The parameter β is explained in the Appendix A. In the political economy framework, the regulator maximizes a objective function which is equal to a weighted sum of a social welfare measure and the political contributions of the lobbies. This parameter is the weight given by the regulator to the social welfare. β is set to obtain a benchmark result which will be explained later. The distribution of the parameter q is determined to be consistent with the value assigned to the expected probability of loan failure (1 q). With this calibration, the social welfare and the payoffs of the different lobbies are of the form showed in various panels of Figure 1. 19

20 Figure 1: Social Welfare and its components 20

21 We can see that, with this calibration, the different parts of the social welfare behave as we have explained in the previous section. We can observe different policy interests for each group which will be important to explain the interaction that we will see in the political economy framework. What we can see is that, in our social welfare measure, consumers surplus is the most important part. On the other side, bank manager is the group with the smallest share in the welfare. We can also see that both firms groups have surplus of the same order of magnitude but they change in opposite directions with the capital requiremnt. However, we will show later that these weights in absolute terms will not be the important thing in the political economy interaction. Finally we can observe that the externality term defines the form of the social welfare. With the quadratic form of it we obtain a function which is monotone, concave and single-peaked. The shape of this payoff is smoother than in the case in which it enters linearly. If we solve the problem of the regulator in an environment in which the lobbies can not make political contributions we obtain the social optimum results which are given in the first column of table 2. Table 2: Results in various political economy equilibria β = β = 0.2 Capital requirement γ Expected price in the final market A Expected production in the final market y Promised payment to depositors D Payment of bank-financed firms if success B Financing decision threshold θ Proportion of bank-financed firms F ( θ) 51.2 % 51.4 % Expected profitability for bank-financed firms π B 5.27 % 5.24 % These equilibrium values satisfy our different calibration benchmarks. For example, with this θ we have approximately half of the operating firms using bank 21

22 financing and half of them using market financing. We can also see that with this expected equilibrium price we obtain the 5 per cent expected profitability for bankfinanced firms. The next step is to analyze how the introduction of political contributions changes the results. To do it the first thing is to obtain a value for the parameter β which determines the weight given to the social welfare in the new objective function of the regulator. We can think that a realistic scenario is the one in which only three lobbies have the resources to make contributions: both types of firms and the bank manager. Given this, a potential value for β could be the one that, in these circumstances, leads to an optimal choice of the regulator of γ = 0.8% - as the case presented in the last column of table 2. This value for β is equal to 0.2. This means that in the new objective function of the regulator the weight of the groups which make political contributions is six times higher than that of those which do not make them. Now, we can analyze the impact of each group s contribution in the final outcome. In Figure 2 we can see the different objective functions of the regulator when only one of the groups makes a contribution. These functions are standardized in such a way that they take a value of 1 when they achieve their maxima. What we can see is that the lowest choice of the regulator would be in the case when only bank-financed firms contribute. In this case, the capital requirement chosen by the regulator would be On the other hand, the highest choice would be in the case when market-financed firms make the contribution. In this situation, the choice would be γ = 1. These two extremes are the reflection of the two main forces which interact in the social welfare. We could think that, for example, consumer surplus would have more importance in this interaction as, in absolute terms, it is a substantial part of the welfare function. However, as we can see in figure 1, this surplus does not change a lot with the capital requirement. This shows that what is important in the political economy interaction is not the weight of a group in the social welfare but the size of the change in its payoff with the different policy out- 22

23 Figure 2: Political Economy Results with One Lobby at a Time comes. What matters is the importance of this policy for the lobby and this, as we have seen in Appendix A, determines its willingness to pay to the regulator. Because of this, market and bank-financed firms are the groups with power enough to change in a considerable way the decision of the regulator as their payoffs are very affected by the capital requirement. More results under this framework are shown in the first column of Table 3 where we can see how the optimal capital requirement γ changes with different combinations of lobbies - market-financed firms (MF), bank-financed firms (BF), Bank manager (BM), Consumers (C) and Financial externalities (FE) - contributions. In the first row of this column we can see that, due to our calibration benchmark, the policy chosen by the regulator when the firms and the bank manager are the only lobbies which make political contributions is γ = 8. The lowest capital requirement set by the regulator is obtained when only bank-financed firms and the bank manager make contributions. As it was mentioned in the previous section, we can also see that the representative bank does not appear as a lobby. This is because, with the different payments 23

24 received and done by the bank, we introduce perfect competition in the sector and this institution has not rents in our framework. The bank manager seems to be the unique representative of the banking activity but his weight in the regulator s decision is very small compared to the one of the other groups. This situation can be viewed as unrealistic. A possible interpretation could be obtained through the bank-financed firms. As we have seen, the firms which use this type of funding are those with low values of θ. We can think that these are the small firms which, usually, have not power enough in a regulatory process like the one described here. However, we can interpret that in our framework these firms receive the rents which in reality, due to the lack of perfect competition in the banking market, belong to the banking sector. In such a situation, where there is market power or a oligopolistic structure, if banks appropriate a part of the surplus of the bank-financed firms, they would push for a lower γ. With this explanation we could interpret the lobby of the bank-financed firms as the banking sector one. Table 3: Results with political contributions ( γ ) Lobbies β = 0.2 β = 0 β = 0.1 β = 1 MF, BF, BM MF BF BM C FE MF,BF BF,BM An interesting exercise is to check the robustness of these results. To do it, we can analyze how our results change when we modify the value of some parameters of interest. In Table 3 we can see how the decisions of the regulator change for different values for β. All positive values are admissable for β and, as we could expect, when the weight given to the social welfare increases, γ becomes closer to the social optimum for all the possible cases. We can also see that the final outcomes do not 24

25 differ too much for the different values of β except for the extreme case in which the regulator does not give any value to the social welfare ( β = 0). This could have important empirical implications. We can think that regulators and the weight that they attribute to social welfare are different in each economy. In a context of a natural experiment, with data about the final policy outcomes for different countries we could infer the weight that the regulators give to the social welfare with respect to the one that they give to the lobbies. This gives us the opportunity to compare regulators from different countries and the extent to which they are influenced by the interest groups. In Table 4 we show how the policy outcome changes with different values for the social cost generated by the bank failure λ. This parameter gives the importance of the financial externality in terms of the proportion of assets of the bank lost due to its failure. As we could expect, a higher value for this parameter implies a policy otucome closer to the social optimum. The explanation for this is that the higher is this social cost, the more important is the externality in the objective function of the regulator. Table 4: Robustness check of results for γ Lobbies λ = 0.15 λ = 0 λ = 0.5 λ = 1 MF, BF, BM Lobbies u[0.45, 1] m = 0.05 r = 0.01 δ = 0.1 MF, BF, BM A final interesting analysis would be to change the proportion of firms using each type of funding. If we change the distribution of θ to u[0.45, 1] we obtain that a fifth of the firms uses market funding while a 80 per cent of them uses bank financing. These figures are similar to the ones that we can observe in a country like Spain. In this case, we observe that the results are quite similar. The social optimum policy is the same and most of the outcomes with political contributions remain unchanged. The main difference is that within this scenario the lobby constituted by marketfinanced firms alone has lower power to influence the decision of the regulator. If 25

26 it interacts with other lobbies the outcomes are similar but if it is the only group making political contributions the decision of the regulator does not achieve extreme values and it is close to the social optimum. More robustness checks are included in the previous table. Surprisingly, an increase in the remuneration of the bank manager leads to a higher γ. The explanation is that this increase in m leads to a higher payoff for the manager but, also, to a remuneration which is proportionally less affected than previously by the change in the capital requirement. As we have explained before, this implies less incentives to influence the decision of the regulator. 4 Conclusions The financial crisis questioned the design of prudential banking regulation. This has motivated the regulatory response which has taken place in the last years. In this process, we have seen that banking activity and the regulation of this sector have very important implications. Because of this, the process has become a general interest issue and it is at the center of the political agenda. Many academics have been involved in the regulatory process and an abundant literature in banking regulation has appeared. However, there is no economic analysis of the political process which is behind. With this crisis the society has realized again that the working and regulation of the banking sector have welfare implications. Politicians have reacted to the complaints of the public opinion and the defects of the regulation. However, this response affects the financial or banking lobby which is considered as one of the most powerful. The fact that this regulation is very important for many economic agents creates a very interesting conflict of interests. These circumstances make a political economy approach very attractive because it is obvious that the political procees behind is, in this case, of vital importance. 26

27 Given this, Grossman and Helpman (1994) give us a very interesting framework which can be applied to this context. In their analysis, different interest groups make political contributions to the regulator to influence its decision and the latter chooses the policy to maximize a weighted sum of the social welfare and those contributions. Combining this with models of banking regulation we can shed light on this complex process. With this framework, we have shown that if there exists a unique lobby which makes contributions the policy outcome chosen by the regulator will be between the social optimum and the one desired by this interest group. In the last years, with the dawn of the crisis, many voices have claimed that the banking system had a inadequate regulation. If we think, that what we could observe was an insufficient regulation, we could explain it as the result of having only one important group exerting its influence, the financial or banking lobby, which put pressure to obtain a lax regulation. In this context, the regulatory response that we are witnessing can be explained through different channels. We can think that different interest groups could have appeared in this process due to the increase of their power with the crisis which gives them the opportunity to have an influence. On the other hand, this could also be explained if the weight that the regulator gives to the social welfare and to the political contributions has changed due to the movement and mobilization of the public opinion or after understanding aspects of this regulation with the financial crisis. This interpretation is very interesting. However, the main contribution of this paper is the development of a useful theoretical framework to analyze the regulatory proccess of the banking sector from a political economy point of view. In the model built here, we include the possibility of having different lobbies trying to influence the decision of the regulator. In this case the final policy outcome is the result of the conflict of interests among all of them. This gives a very rich tool which can explain the deviations of the regulation from the social optimum as the result of the interaction of all these groups. 27

28 The analysis points out that what is important in the political economy interaction is not the weight of a group in the social welfare in absolute terms but the size of the change in its payoff with the different policy outcomes. What matters is the importance of this policy for the lobby and this determines its willingness to pay to the regulator. In our framework, the two lobbies of the model whose interaction determines the final policy outcome are the two groups of firms. market-financed firms push for a high value for the capital requirement to obtain more rents in the final market through a lower competitiveness of the bank-financed firms. These firms push in the opposite direction. As we hace perfect competition in the banking market, we can also interpret that these firms are the representatives of the banking sector. The model has important implications. On one hand, we can think about the policy ones. It is clear that it is important to understand well the different interests of the groups involved in this regulatory process. In addition, a good comprehension of how the lobbies influence the decision of the regulator is important to improve the outcomes. It seems clear that the independence of the regulator is a basic goal to obtain. The possibilities of this framework do not end here. It could be interesting to analyze the effects of deposit insurance. We can also think about the introduction of more lobbies - an endogenous opportunity cost for capital owners δ gives us the possibility of including them in the game. Although the interpretation of the bank-financed firms as the supporters of the interests of the banking sector seems reasonable, it would be a good exercise to evaluate the implications of eliminating perfect competition in the banking market and introducing a financial lobby. Finally, given the media impact of the process, we could try to use this framework to analyze the influence of the public opinion. To do it, we could introduce a new lobby but also use or endogenize the weight β given to the social welfare. 28

29 References [1] Aidt, Toke (1998): Political Internalization of Economic Externalities and Environmental Policy, Journal of Public Economics, 69, 116. [2] Bernheim, Douglas, and Whinston, Michael (1986): Menu Auctions, Resource Allocation and Economic Influence, Quarterly Journal of Economics, vol. 101(1), 131. [3] Diamond, Douglas (1991): Monitoring and Reputation: The Choice between Bank Loans and Directly Placed Debt, Journal of Political Economy, University of Chicago Press, vol. 99(4), [4] Fredriksson, Per (1999): The political economy of trade liberalization and environmental policy, Southern Economic Journal, vol. 65(3), [5] Grossman, Gene, and Helpman, Elhanan (1994): Protection for Sale, American Economic Review, Vol. 84(4), [6] Holmstrom, Bengt and Tirole, Jean (1997): Financial Intermediation, Loanable Funds, and the Real Sector, Quarterly Journal of Economics, MIT Press, vol. 112(3), [7] Laffont, Jean-Jaques, and Tirole, Jean (1993): A Theory of Incentives in Procurement and Regulation, MIT Press. [8] Perotti, Enrico, and Suárez, Javier(2011) : A Pigovian Approach to Liquidity Regulation, International Journal of Central Banking, 7, [9] Repullo, Rafael, and Suárez, Javier(2012) : The Procyclical Effects of Bank Capital Regulation, Review of Financial Studies. [10] Repullo, Rafael, and Suárez, Javier(2004) : Loan Pricing under Basel Capital Requirements, Journal of Financial Intermediation, 13(4),

30 A The Basic Political Economy Model One of the main issues of this paper is the interaction between the regulator and the lobbies. As we have mentioned, to describe it we will follow the work of Grossman and Helpman (1994) which, at the same time, is inspired in the menu-auction problem explained by Bernheim and Whinston (1986). We are interested in the equilibrium of a two-stage noncooperative game. Firstly, N lobbies - indexed by i = 1, 2,..., N - simultaneously choose their contribution schedules C i (p) 0. These political contributions are binding and contingent on the policy vector p finally implemented by the regulator. In the second stage, after the submission of these schedules, the regulator sets the policy p. The equilibrium is a set of contribution functions {Ci (p)} i=1,...,n, which maximizes for each lobby its welfare given the contributions of the remaining lobbies and the anticipated problem faced by the regulator, and a policy vector p that maximizes the regulator s objective function given the schedules submitted by the lobbies. To find the subgame-perfect Nash equilibrium of the policy game we can use backwards induction. In the second stage, the regulator maximizes its objective function which is a weighted sum of the total contributions from the lobbies and an aggregate measure of social welfare W (p): Max p [ N ] C i (p) + βw (p), (20) i=1 where β is the weight the regulator attaches to aggregate welfare, which can be interpreted as determined by electoral concerns, political premia or the regulator s own ethic. The aggregate welfare measure is: N M W (p) = U i (p) + V j (p) + G(p), (21) i=1 j=1 where U i (p) is the payoff obtained by lobby i, V j (p) is the utility of some agent or 30