A Simple Model of Law and Corporate Governance Baptiste Massenot∗ Feruary 2012

Abstract The paper introduces limited enforcement in a model of the political economy of corporate governance. In general, legal protection (the laws) and judicial protection (the quality of enforcement) both offer protection against wrongdoing and are thus substitutes. In the case of corporate governance, however, investor and employment protection draw their resources from the same pool and it is thus not always feasible to increase both of them at the same time when judicial protection decreases. The empirical evidence is consistent with the main implication of the model, namely, that the quality of enforcement is negatively correlated with investor protection and positively correlated with employment protection. These results shed light on the literature on legal origins and corporate governance. Keywords: Corporate governance; Legal origins; Employment protection; Investor protection; Contract enforcement JEL Classification Numbers: G38; K42

∗

University of Lausanne; [email protected]. I am grateful to Giacomo Ponzetto, Raghuram Rajan, Gilles Saint-Paul, and Luigi Zingales for very helpful discussions. I gratefully acknowledge financial support from the National Centre of Competence in Research “Financial Valuation and Risk Management” (NCCR FINRISK).

1

1

Introduction

Political economics generally assumes that policies are perfectly enforceable in court. Although suitable for most purposes, this approach cannot account for the recent empirical evidence suggesting that the characteristics of the judiciary are systematically associated with a large number of policies. More specifically, the objective of this paper is to provide a theoretical basis for the following evidence:1 Fact 1 The empirical evidence on the regulation of corporate governance and legal origins can be summarized as follows: • Employment protection is higher in civil law (CIV) than in common law countries (COM). • Investor protection is lower in CIV than in COM. This paper focuses on the regulation of corporate governance because this area has received a lot of attention from the literature and leaves for future research the links between judiciaries and other areas of law. In particular, Fact 1 has contributed to generate a large empirical literature that uses legal origins either as an instrument or as a control variable.2 The theoretical underpinnings of Fact 1 remain, however, controversial. The legal origins literature has argued that the greater liberty given to common law judges to make new precedents makes the common law more adaptable to a changing environment.3 They further argue that this adaptability is efficient, which in the context of corporate governance means both a lower worker protection and a higher investor protection.4 However, as

1

See Botero et al. (2004) for evidence on worker protection and La Porta et al. (1998); Djankov et al. (2008) for evidence on investor protection. See La Porta et al. (2008) for an excellent survey of the legal origins literature. 2 Legal origins are a convenient instrument because they can be considered as exogenous. They are also an important control variable because they have been shown to have pervasive political and economic consequences. See La Porta et al. (2008) for arguments supporting the exogeneity of legal origins and for a survey of the research in which they are used. 3 See Beck et al. (2003) and La Porta et al. (2008). 4 See Rubin (1977) and Gennaioli and Shleifer (2007) for formal arguments. See Beck et al. (2003) for supportive empirical evidence.

2

noted by Roe (2006) and acknowledged by La Porta et al. (2008), this theory is unable to account for the fact that most of the laws that correlate with legal origins in Fact 1 are not judge-made. The contribution of this paper is to make the point that Fact 1 can be accounted for by a parsimonious political economy model in which the laws that regulate corporations are imperfectly enforceable. Because empirical evidence shows that the quality of enforcement is higher in COM than in CIV,5 the model can be used to predict how legal origins affect the regulation of corporate governance. At first sight, judicial protection (the quality of enforcement) and legal protection (the actual law) should be substitutes. Indeed, both of these institutions provide protection against wrongdoing. Then, lower judicial protection pushes a social planner or the political process to increase legal protection to implement his preferred level of de facto protection. For example, take a minimum wage Wmin that is imperfectly enforceable, meaning that if a firm pays a worker less than this minimum wage, a trial to enforce the payment of this minimum wage would cost the worker C or, alternatively, would only be succesful with some probability. Then, it is in the interest of the worker to only go to court when the firm pays less than Wmin − C as he would otherwise end up worse off. Knowing this, the firm can pay the worker only Wmin − C and by doing this make a larger profit without facing the risk of being sued. The substitutability between legal and judicial protection can well explain the first part of Fact 1 but not its second part. To make further progress, it is worth noting that investor and worker protections draw their resources from the same pool. Although it may be possible to increase both worker and investor protections at the same time, this should stop at the point where the firms are not willing to yield ground anymore. Starting from this point, investor and worker protections become substitutes. The social planner cannot adjust legal protection as he wishes because this affects the other group. Then, judicial protection moves worker and investor protections in opposite directions. The predictions of the model can thus be consistent with Fact 1 if the social planner prefers to compensate workers 5

Djankov et al. (2003) show that COM have shorter trials, less procedural courts as well as a higher perceived quality of their justice than CIV.

3

rather than investors when judicial protection decreases. Finally, the paper shows some supportive emprirical evidence for the main implication of the model: the quality of enforcement is positively correlated with investor protection and negatively correlated with employment protection. Pagano and Volpin (2005) and Perotti and Von Thadden (2006) also provide theories of the political economy of corporate governance but they do not account for Fact 1. Pagano and Volpin (2005) find that whether political systems are proportional or majoritarian can explain a significant share of the variation in corporate governance regulation across countries, both in theory and in the data. Interestingly, they also control for legal origins in their empirical analysis and find them to be often significant. Perotti and Von Thadden (2006) find that the evolution of wealth concentration can well explain the time-series variation in corporate governance observed in the twentieth century. The paper is organized as follows. Section 2 introduces the model. Section 3 describes the legal system. Section 4 solves for the equilibrium regulation of corporate governance. Section 5 shows some empirical evidence. Section 6 concludes.

2

The Environment

There are three types of agents: workers, investors, and entrepreneurs. Workers supply labor to entrepreneurs in exchange for a wage W . Investors supply capital to entrepreneurs in exchange for dividends D. Everybody has an outside opportunity of 0. If investors, entrepreneurs, and workers decide to participate in the production process, they produce Y > 0. Then, workers derive utility Uw = W , investors Ui = D, and entrepreneurs Ue = Y − W − D. ¯ that entrepreneurs Regulation imposes a level of worker protection W ¯ that enhave to pay to workers and a level of investor protection D trepreneurs have to pay to investors. This gives the following feasibility constraint: ¯ +D ¯ ≤ Y. W (1)

4

Both the data and the theory show that higher worker protection leads to higher unemployment rates and poorer investor protection leads to lower financial development.6 I thus assume that a higher worker protection and a lower investor protection entail a social welfare loss given by the function: L(W, B) =

W 2 (Y − D)2 + . 2 2

(2)

This loss function represents, for example, the disutility of the workers who are unemployed or of the entrepreneurs who could not get financed. ¯ and D ¯ are endogenously deterThe rest of the paper analyzes how W mined by the political process. A complicating factor is that these regulations are only partially enforced by the legal system.

3

To Sue or Not to Sue

Although there exist regulations on corporate governance, entrepreneurs can decide not to respect these regulations because of limited enforcement. ¯ , the worker can sue him If an entrepreneur pays a worker less than W at some cost and with some probability of success. One can expect that entrepreneurs are willing to pay a little bit less than the minimum wage whenever they expect workers are not going to sue them because the trial would be too uncertain or too costly. Let W be the wage for which workers are indifferent between suing or not. Then, the more expensive the lawsuit or the lower its probability of success, the lower W . I formalize this idea by setting ¯, W = pW (3) with 0 < p < 1 and where a higher p implies better and cheaper enforcement. Furthermore, I assume that the litigation costs and the compensatory damages are such that it is not in the interest of the entrepreneur to pay W < W . Then, entrepreneurs decide not to respect the law and to pay W 6

See Botero et al. (2004) on the positive relationship between worker protection and unemployment. See La Porta et al. (1998) and Djankov et al. (2008) on the positive relationship between investor protection and financial development. The model developed by Pagano and Volpin (2005) has predictions consistent with these correlations.

5

¯ . This leads to the distinction between the de jure to workers instead of W ¯ – the legal rule – and the de facto legal rule W – how level of protection W much workers are actually paid. A similar logic applies to investor protection. By paying D, entrepreneurs are not sued, where: ¯ (4) D = pD. The impact of a poorer quality of enforcement (a lower p) on each group of individuals for given regulations is as follows. Workers and investors end up worse off because they are paid less. By contrast, entrepreneurs become better off because they can steal more. The next section solves for the equilibrium regulations.

4

Corporate Governance

The de jure levels of worker and investor protections are chosen by a social planner who maximizes social welfare.7 Importantly, this social welfare function uses the de facto levels of protection as arguments. The social planner applies a weight αw to workers, αi to investors, and αe to entrepreneurs. He maximizes the following objective function: max ¯ ,D ¯ W

X

αj Uj − L(W , D),

(5)

j

¯ ≥ 0, D ¯ ≥ 0, and (1). under the constraints W The solution to this problem is: ¯ Result 1 Assuming αw > αi > αe − Y and αw > αe , worker protection W is given by ¯ = αw − αe , W (6) p ¯ is given by and investor protection D ( ¯= D

Y +αi −αe p

if D < Y − W ,

Y − W if D ≥ Y − W .

7

(7)

Pagano and Volpin (2005) use a more sophisticated model of political economy to determine the equilibrium corporate governance. This is because they need a more micro-founded approach to be able to compare political systems.

6

Equation (6) gives the optimal degree of employment protection. It has intuitive properties: It is increasing in the political weight of workers αw , decreasing in the political weight of entrepreneurs αe , and decreasing in the quality of enforcement p. Equation (7) gives the optimal degree of investor protection. For a small degree of worker protection, investor protection only depends on parameters. It is increasing in the political weight of investors αi , decreasing in the political weight of entrepreneurs αe , and decreasing in the quality of enforcement p. For a high degree of worker protection, however, the feasibility constraint (1) starts binding. Then, the social planner has to choose among the set of combinations of W and D that sum up to Y . Because Result 1 assumes that workers are given the highest weight, the ¯ unchanged and reduces D. ¯ social planner leaves W This result hinges on specific parametric restrictions. Assuming αi > αw , the result would be reversed in the case D ≥ Y − W , that is, the social planner would leave investor protection unchanged and would decrease worker protection. Assuming αe ≥ αw or αe − Y ≥ αi would give the ¯ = 0 or D ¯ = 0, respectively. solution W We can now summarize the effect of judicial protection p on corporate governance. As long as D < Y − W , both investor and worker protections increase. This is because judicial and legal protections are perfect substitutes. The social planner chooses the de facto levels of investor and worker protection that maximize social welfare. Then, if judicial protection changes, the social planner adjusts the de jure legal protection to obtain his preferred level of de facto protection. However, in the case D ≥ Y − W , the social planner cannot change freely legal protection anymore because of the feasibility constraint (1). In this case, a lower judicial protection still increases worker protection but decreases investor protection. How do legal origins affect corporate governance? Consider the two legal origins, COM and CIV, that differ only in their degree of judicial protection, pCOM and pCIV . Then, the following result holds: Result 2 Assuming αw > αi , pCOM > pCIV , and D ≥ Y − W , Result 1 is consistent with Fact 1. The first restriction αw > αi has already been discussed and implies that 7

the social planner prefers to impose the burden of lower judicial protection on investors rather than on workers. The second restriction D ≥ Y − W implies that worker and investor protections move in opposite directions. This is consistent with the empirical evidence and the theory presented in Pagano and Volpin (2005). If this restriction was not satisfied, worker and investor protections would be independent. Finally, there is supportive evidence for the restriction pCOM > pCIV : COM is associated with lower formalism of judicial procedures, with faster trials, and with a higher perceived quality of its justice than CIV (Djankov et al., 2003).

5

Some Empirical Evidence

One implication of the model is that the quality of enforcement should be positively correlated with investor protection but negatively correlated with other measures of legal protection, including employment protection. This section shows that the cross-country empirical evidence is consistent with these predictions. The objective of this section is not to provide a formal test of the model nor to test alternative hypotheses of the determinants of the regulation of corporate governance but more modestly to provide a descriptive account of the main relationships predicted by the model. To proxy for the quality of enforcement, I use the formalism of judicial procedures to collect a bounced check from Djankov et al. (2003). A higher formalism implies a lower quality of enforcement. To proxy for employment protection, I use three variables from Botero et al. (2004). First, the employment laws index that summarizes a number of labor regulations that protect workers and that are related to dismissal, overtime, part-time, etc. Second, a dummy variable indicating whether there exists a mandatory minimum wage; Third, the number of days of annual leave wih pay. We expect these three measures to be negatively correlated with the quality of enforcement as workers demand to be compensated with higher legal protection when judicial protection is low. Consistent with this hypothesis, Figures 1, 2, and 3 show a negative relationship 8

between these three measures of employment protection and the quality of enforcement. To further document the substitutability between legal and judicial protection, I also use the number of steps to start operating a business legally from Djankov et al. (2002) as an additional measure of legal protection. This measure gives us the extent to which incumbent firms are protected from new entrants. If legal and judicial protections are really substitutes, we should expect entrants to demand a higher level of legal protection when judicial protection is lower. Consistent with this hypothesis, Figure 4 shows a negative relationship between the protection of entry and the quality of enforcement. To proxy for investor protection, I use two variables. First, an index of creditor rights from Djankov et al. (2007) that measures the extent to which creditors are protected during bankruptcy; Second, a self-dealing index from Djankov et al. (2008) that measures the degree of protection from self-dealing by corporate insiders. If legal and judicial protections are substitute, we should expect as before a negative relationship between these measures of investor protection and the quality of enforcement. However, if as suggested by Result 1 investor and worker protections are themselves substitutes, we should instead expect a positive relationship. Consistent with the latter hypothesis, Figures 5 and 6 show a positive relationship between formalism and both creditor rights and the anti self-dealing index. As a robustness check, I also run OLS regressions of legal protection on judicial protection controlling for income as this was shown to be an important determinant of legal protection. To proxy for income, I use the logarithm of per capita GDP PPP in 1995 from the WDI. The coefficients on the quality of enforcement variable are of the expected sign and are all highly significant. Finally, these findings are robust to several alternative measures of employment protection, investor protection, and quality of enforcement. These additional specifications are not reported for conciseness.

9

6

Conclusion

This paper presents a simple model of the political economy of corporate governance and focuses on the role of legal systems. Under reasonable parametric restrictions, the model can account for the empirical evidence on legal origins and corporate governance. This paper thus contributes to our understanding of the political consequences of legal systems. One implication of the model is that the quality of enforcement should be positively correlated with investor protection but negatively correlated with other measures of legal protection, including employment protection. The data is consistent with these predictions. The focus of this paper is on corporate governance. The literature suggests, however, that legal origins have more pervasive political consequences, for example, on government ownership of banks and the media, entry regulations, conscription, etc. Future research should investigate whether these correlations can as well be reconciled with the theory of political economics. Finally, more research remains to be done to build a unified theory of the determinants of the regulation of corporate governance. In particular, recent work has pointed out the importance of trust for regulation (Aghion et al., 2010). It would be interesting to further investigate whether trust and enforcement are also related.

References Aghion, Philippe, Yann Algan, Pierre Cahuc, and Andrei Shleifer, “Regulation and Distrust,” The Quarterly Journal of Economics, 2010, 125 (3), 1015–1049. Beck, Thorsten, Asli Demirg¨ u¸c-Kunt, and Ross Levine, “Law and Finance: Why Does Legal Origin Matter?,” Journal of Comparative Economics, 2003, 31 (4), 653–675. Botero, Juan C., Simeon Djankov, Rafael La Porta, Florencio Lopez-De-Silanes, and Andrei Shleifer, “The Regulation of Labor,” Quarterly Journal of Economics, 2004, 119 (4), 1339–1382. 10

Djankov, Simeon, Caralee McLiesh, and Andrei Shleifer, “Private credit in 129 countries,” Journal of Financial Economics, 2007, 84 (2), 299–329. , Rafael La Porta, Florencio Lopez de Silanes, and Andrei Shleifer, “The Regulation of Entry,” Quarterly Journal of Economics, 2002, 117 (1), 1–37. , Rafael La Porta, Florencio Lopez-De-Silanes, and Andrei Shleifer, “Courts,” Quarterly Journal of Economics, 2003, 118 (2), 453– 517. , Rafael La Porta, Florencio Lopez de Silanes, and Andrei Shleifer, “The law and economics of self-dealing,” Journal of Financial Economics, 2008, 88 (3), 430–465. Gennaioli, Nicola and Andrei Shleifer, “The Evolution of Common Law,” Journal of Political Economy, 2007, 115 (1), 43–68. La Porta, Rafael, Florencio Lopez de Silanes, and Andrei Shleifer, “The Economic Consequences of Legal Origins,” Journal of Economic Literature, 2008. , Florencio Lopez-De-Silanes, Andrei Shleifer, and Robert W. Vishny, “Law and Finance,” Journal of Political Economy, 1998, 106 (6), 1113–1155. Pagano, Marco and Paolo F. Volpin, “The Political Economy of Corporate Governance,” American Economic Review, 2005, 95 (4), 1005– 1030. Perotti, E.C. and E.L. Von Thadden, “The political economy of corporate control and labor rents,” Journal of Political Economy, 2006, 114 (1), 145–175. Roe, Mark J., “Legal origins, politics, and modern stock markets,” Harv. L. Rev., 2006, 120, 460. Rubin, Paul H., “Why Is the Common Law Efficient?,” Journal of Legal Studies, 1977, 6 (1), 51–63. 11

.8

RUS

TUN PRT MOZ

GEO SWE FINFRA NLD

KAZ

SVN

ESP

LVA DEU JOR

NOR

TZA IDN UKR

ITA

VEN

POL LTU

.6

DOM DNK

PAN MEX

BRA VNM GRC CZE

BEL

BGR

AUT HRV LKA

TWN

CHE

.4

PHL PER

ECU HUN

KEN UGA IRL

ZAF

LBN

KOR IND CHN

THA

TUR

AUS

SEN

CHL

EGY PAK

BOL COL

ARG

ROM

SGP GHA GBR

ISR

CAN

URY MAR

ZWE

.2

USA MYS HKG

0

NZL

ZMB

JAM

.2

MWI JPN

NGA

.4 Formalism Empl. index

.6

.8

Fitted values

Figure 1: Employment laws index and formalism. Source: Botero et al. (2004) and Djankov et al. (2003).

12

1.5

CAN ZMB

JAM TWN TUR UGA USA GHA BEL

MWI JPN BRA NLD KEN THA NGA FRA VNM ISR IND KOR CHN HUN

MYS SGP DNK GBR IRL

NOR SWEGEO ZWE CHE FIN

HRV UKRPAK LKA TZA IDN LVA PRT GRC TUN URY CZE DOM COL POLSVN ROM LTU MOZ BGR MAR CHL MEX SEN KAZLBN ECU PHL

ESP ARG

PER BOL PAN VEN

.5

1

NZL

0

HKG

0

ZAF AUS

.2

DEU AUT JOR

.4 Formalism Min. wage

ITA

.6

.8

Fitted values

Figure 2: Mandatory minimum wage and formalism. Source: Botero et al. (2004) and Djankov et al. (2003).

13

30

FIN

AUT

MOZ

BOL

KOR DNK POL SWE ZMB

TWN TUR

FRA

HUN

ARG

GEO

UKR

PRTTUN

BRA

20

BEL

SEN

GRC CZE

NOR KEN AUS

LTU

ECU

ESP

PER

PAN

JOR

JPNNLD CHE ISR

TZALVAITA URY

UGA

HRV

MEX

DOM SVN ROM CHL BGR MAR

MYS NZLZAF

JAM

IRL

MWI

IND

DEU

SGP GHA

COL

KAZLBN

PAK

10

VNM HKG

CAN

LKA IDN

THA NGA

CHN PHL

0

GBR USA

0

.2

ZWE

.4 Formalism

Nb days paid leave

.6

.8

Fitted values

Figure 3: Number of days of annual leave with pay and formalism. Source: Botero et al. (2004) and Djankov et al. (2003).

14

VEN

3

RUS

BOL DOM

VNM FRA

MOZ

COL

SEN MEX

ECU

GRC

2.5

GEO TUR

MWI

UGA

KOR

2

TWN

THA

ZMB

HKG

JAM

BGR CHL

DEU AUT

PAK LKA

LTU

ESP LBN PHL

SVN LVA

PAN

NLD CHE

PER

ZWE ISR HUN

GBR

NOR SWE

DNK USA IRL

TUN

FIN

1

NZL

ARG

KAZ

PRTURY CZE POL

BEL SGP

MAR

HRV EGY TZA IDN ITA

IND CHN KENNGA

MYS

1.5

ROM

JOR

JPN BRA

GHA ZAF

VEN

UKR

CAN

.5

AUS

0

.2

.4 Formalism Log nb steps

.6

.8

Fitted values

Figure 4: Number of steps to start operating a business legally and formalism. Source: Djankov et al. (2002) and Djankov et al. (2003).

15

4

HKG

NZL

GBR

3

ZAF AUS

MYS SGP DNK

2

JAM TWN TUR UGABEL

CAN ZMB

1

KEN ZWE NGA

NLD

ISR KOR DEU AUTHRV

MWI NOR JPNGEO THABGD IND RUS CHN

USA IRL GHA

0

.2

PAK PRT GRC POL

CIV

.4 Formalism Creditor rights

KWT LVA URY CZE BWA SVN

UKRLKA EGY ARE NAM TZA IDN ITA DOM

SWE BRA CHE FIN VNM HUN JOR

FRA

0

LBN

TUN COL

PAN

SLV

VEN

LTU MOZ BGR KAZ HND CHL

ROM

MAR

MEX SEN

.6

PHL

ESP

BOL

ARG CRI

ECU

GTM

PRY

PER

.8

Fitted values

Figure 5: Creditor rights and formalism. Source: Djankov et al. (2007) and Djankov et al. (2003).

16

1

SGP HKG

NZL

MYS GBR

THA

.8

ZAF AUS

IRL

CHN

GHA ISR IDN CAN

BGR

USA

.6

CHL COL TWN

SLVMAR

IND

BEL NGA DNK

.4

TUR UGA

JPN FIN NOR ZWE

EGY PRT

RUS KOR

ARG PAK LKA LVA

SWE BRA CHE

ROM ITA

FRA JAM

KAZ

.2

ESP

CZE POL

DEU LUX HRV

KEN NLD

PER

LTU

GRC

ISL

PHL

AUT HUN MEX

TUN URY

JOR

PAN UKR BOL

0

ECU

0

.2

.4 Formalism Anti self−dealing

.6

.8

Fitted values

Figure 6: Anti self-dealing and formalism. Source: Djankov et al. (2008) and Djankov et al. (2003).

17

VEN

18

p-values in parentheses ∗ p < 0.10, ∗∗ p < 0.05,

∗∗∗

p < 0.01

70 0.306

89 0.048

1.958∗ (0.068)

0.869∗∗∗ (0.001)

Constant

Observations Adjusted R2

0.082 (0.451)

-0.005 (0.851)

Log GDP pc

-1.468∗∗ (0.030)

-0.782∗∗∗ (0.000)

Formalism

(1) Anti self-dealing

78 0.172

-0.011 (0.952)

0.029 (0.118)

0.499∗∗∗ (0.000)

76 0.272

76 0.102

3.024 (0.662)

1.003 (0.156)

-0.125∗∗∗ (0.002) 1.421∗∗∗ (0.000)

14.051∗∗∗ (0.002)

(5) Nb days paid leave

0.900∗∗∗ (0.000)

Table 1: Legal and judicial protection (2) (3) (4) Creditor rights Empl. index Min. wage

78 0.424

3.114∗∗∗ (0.000)

-0.192∗∗∗ (0.000)

1.437∗∗∗ (0.000)

(6) Log nb steps