Chapter 7: Asymmetric information, agency costs and capital structure

Essential reading
Grinblatt, M. and S. Titman Financial Markets and Corporate Strategy. (Boston, Mass.; London: McGraw-Hill, 2002) second edition [ISBN 0072294337] Chapters 16, 17,18, and 19.

Further reading
Brealey, R. and S. Myers Principles of Corporate Finance. (Boston; Mass., London: McGraw Hill, 2003) seventh edition [ISBN 0071151451] Chapter 18. Copeland, T. and J. Weston Financial Theory and Corporate Policy. (Reading, Mass.; Wokingham: Addison-Wesley, 1988) third edition [ISBN 0201106485] Chapter 14. Jensen, M. and W. Meckling Theory of the firm: managerial behaviour, agency costs and capital structure', Journal of Financial Economics (1976) 3, pp. 305-60. Masulis, R. The impact of capital structure change on firm value: some estimates', Journal of Finance (1983) 38, pp. 107-26. Myers, S. 'Determinants of corporate borrowing', Journal of Financial Economics (1977) 5, pp. 147-75. Myers, S. and N. Majluf 'Corporate financing and investment decisions when firms have information that investors do not have', Journal of Financial Economics (1984) 13, pp. 187-221. Ross, S. The determination of financial structure: the incentive signaling approach', Bell Journal of Economics (1977) 8, pp. 23-40.

Overview
In the previous chapter we introduced the capital irrelevance proposition first put forward by Miller and Modigliani (1958). We also explored cases in which the capital structure of a firm did matter in its valuation due to relaxations,of the MM assumptions (e.g. the introduction of corporation tax and bankruptcy costs). In this chapter we will focus on two classes of problem in which MMl does not hold. In the first, firms are subject to agency problems between outside stakeholders and insiders (managers). The second class of problem allows the possibility that insiders to the firm are better informed about its quality than the market or potential external investors.

Capital structure and agency costs

In most Western corporations, ownership and control are separate, in that the owners of a firm (the firm's security-holders) entrust the day-to-day running of the firm to managers. In general, although owners may have an idea of what the optimal strategy for the firm is, it is impossible to force managers to follow this plan. Managers may then behave opportunistically, taking inflated salaries, investing in pet projects and enjoying other perquisites (perks). Hence, in such scenarios, managers can corporate policy to maximise their own utility rather than setting the policy which would maximise shareholder wealth. This is the agency problem that arises in modern corporations and was first talked about in relation to capital structure by Jensen and Meckling (1976).
75

Agency costs of outside equity and debt

Jensen and Meckling (1976) argue that understanding of two types of agency cost is important in understanding why firm value is not invariant to capital structure. The first of these is an agency cost associated with outside equity. Assume a firm that is financed solely by equity. A proportion of the equity is held by the management of the firm, whereas the rest is held by outsiders to the firm. Jensen and Meckling argue that such a situation leads to firm values which are lower than that which would obtain if the manager was the sole owner of the firm. To see why this is the case, consider the rewards and costs facing the manager/equity-holder. The manager is the agent who undertakes activities that add value to the firm. Let's call these activities 'effort'. Increased effort supply leads to greater firm value and vice versa. However, supplying effort is also costly to the manager (it takes up his time and tires him or her mentally and physically, foj example). In situations where a proportion a of the firm's equity is held by outsiders, the manager bears the entire cost of effort supply but reaps only a portion (1 -a) of the benefit. Hence, the outside equity-holders gain from the manager increasing effort but don't bear any costs. This induces the manager to supply lower levels of effort for higher values of a (i.e. when the proportion of profits the manager appropriates is low, his or her incentive is to supply little amounts of effort). Hence, firm value is decreased when the proportion of equity held by outsiders is increased, and MM1 does not hold. This is the agency cost of outside equity. Jensen and Meckling argue that the agency cost of outside equity is decreasing in the leverage ratio of the firm (where leverage is the ratio of debt to equity values). The argument runs as follows: assume that the firm repurchases some of the equity held by outsiders, funding this with a debt issue - hence, leverage is increased. Also, the proportion of outstanding equity held by the manager is now increased. Thus, as his share of the residual value of the firm is increased, the manager chooses to supply more effort, leading to increased firm value. Then, as leverage rises, agency costs of outside equity fall. The second agency cost highlighted by Jensen and Meckling is that associated with debt finance. It is also known as the asset substitution or risk-shifting problem associated with debt finance. To illustrate the problem, consider the following example:
Example

Assume a firm that is financed by both debt and equity. A manager runs the firm in the interest of current equity-holders (i.e. the manager sets investment policy in order to maximise the expected shareholder pay-off). The manager is faced with the choice between two investment projects, A and B. These projects are assumed to have zero cost and are mutually exclusive. The cash flows to projects A and 8 are given in Table 7,1. State 1
Probabilities ..Cash flow A Cash flow B

State 2
0.5

State 3
0.25

0.25

40 20

50 40

60 80

Table 7.1

76

Clearly, both projects have positive expected NPV. Project A has the lowest risk and the higher expected NPV (50), whereas Project B is the riskier and its expected NPV is 45.1

We assume that debt-holders have a claim of 50 that must be repaid out of the cash flow to the chosen project. Given this debt claim, which project will the manager choose? If we first analyse Project A, it is obvious that, with a debt obligation of 50, only in state 3 will equity-holders get any pay-off, this pay-off being 10. This implies that the expected pay-off from Project A to shareholders is 10 x 0.25 = 2.5. The expected pay-off to debt-holders from A is equal to (0.25 x 40) 4(0.5 x 50) 4- (0.25 x 50) = 47.5. Moving on to the analysis of Project B, again equity-holders only get some cash in state 3 and their expected pay-off is 0.25 x 30 = 7.5. The pay-off to debt-holders from Project B is (0.25 x 20) 4- (0.5 x 40) 4- (0.25 x 50) = 37.5. Hence, from the equity-holders point of view, Project B maximises expected pay-off and, as a result, this will be the project chosen by the manager. Note that the choice of this project implies that debt-holders are worse off and firm value lower than in the case where Project A is chosen. When the face value of debt is 50, the firm invests in the project with the lower expected NPV and higher risk, as this project maximises the expected return to equity. What would happen if the debt repayment outstanding were 30 instead of 50? In this case the expected pay-offs to equity-holders are 20 from Project A and 17.5 from Project B. Therefore, the manager will choose Project A. This choice also implies that debt-holders are happy as Project A maximises their expected pay-off (they get 30 rather than the 27.5 that they would expect to receive if Project B were chosen). Note that, when the face value of debt is lower, the manager switches and chooses the low-risk, highexpected-return project. This in turn implies that, when face value of debt is lower, firm value is higher. Jensen and Meckling argue that the agency costs of debt are increasing in the level of debt outstanding and hence in corporate leverage. Combining the two agency costs then allows us to argue that an optimal (in the sense of maximising firm value) capital structure might exist. We contended that the agency cost of outside equity was decreasing in leverage, wh\reas the agency cost of debt increased with leverage. Firm value would be maximised where 'total agency costs are minimised, and this leads to the optimal leverage ratio shown on Figure 7.1.
Cost

' When we say that Project B is riskier, we mean that it has higher cash-flow variance than Project A.

^nun

r> ,

Agency cost equity

DjE'
Figure 7.1

DIE

The Myers (1977) debt overhang problem Another agency cost of debt was pointed out by Myers (1977). Rather than arguing that debt obligations induce managers to invest in excessively risky projects, Myers argues that the management of firms with large levels of debt outstanding will choose to reject some positive NPV projects. As a result, heavily indebted firms will see reductions in corporate value, and MM1 is violated. This is known as the debt overhang problem. To illustrate the previous argument consider the situation depicted in Table 7.2. A given firm is presented with the opportunity to invest in a certain project at the current time. The pay-off of this investment is $20,000 at time t+1 regardless of the state of nature, and the cost at time t is $10,000. We assume, for simplicity, that interest rates are zero such that the investment has a positive NPV. Further, the firm receives cash flow at time t, which reflects its past investments. This cash flow is uncertain. As depicted in Table 7.2, with probability 0.25 it will be^50,000; it will be $80,000 with probability 0.5 and, finally, with probability 0.25 it will be $120,000. The firm is run by a manager who acts in the interest of current shareholders. In the past, the firm issued debt with a face value of $100,000. This debt must be repaid out of the cash-flow to the firm, after the investment decision has been made and any pay-offs realised. Note that, if the project is accepted by the manager, its cost must be met out of the pockets of equity-holders. State 1 Probabilities Cash flow existing assets Cost New Project Return New Project
0.25
50 10 20
State 2
0.5

State 3
0.25
120
10 20

80 10 20

Table 7.2 When the face value of debt is $100,000, the manager will reject the new project. Why is this? Note that, in states 1 and 2, the new project pays $20,000, but this simply goes straight into the pockets of debt-holders through the required payment of $100,000. It is only in state 3 that the $20,000 pay-off of the new project accrues to equity-holders. Hence, in this case the expected net payoff to equity-holders is: (0.25x20) -10 = -5 As this is negative, the manager rejects the new project. The implication of this is that, when debt levels are high, a firm may reject a project with positive NPV, as little of that project's pay-off accrues to equity-holders. To confirm this, consider the case in which the required debt payment is $80,000 rather than $100,000. In this case, the pay-off from existing assets is sufficient to service the debt in both states 2 and 3. Hence, in both these states the equity-holders reap all of the rewards from the new project, whereas the new project pay-off goes to debt-holders in state 1. Hence, the expected net return to equity-holders from the new project is:
(0.5 x 20) +- (0.25 x 20) - 10 - 5

As. this is positive, the manager will accept the project as it increases expected shareholder wealth.
%
Activity

Compute the expected pay-off to equity holders if the required debt repayment is 90. Will the manager accept or reject the project?

The preceding example illustrates the debt overhang argument. Managers that run heavily indebted corporations in the interest of equity-holders may reject positive NPV projects as the cash flows from such projects accrue mostly to debt-holders, whereas equity-holders bear the costs. The rejection of such projects implies that firm values are sub-optimal.

Firm value and asymmetric information

The preceding section emphasised the point that agency problems may lead to departures from MM1. An alternative reason for such departures is the presence of information asymmetries between corporate insiders and outsiders. The role played by asymmetric information is emphasised by Ross (1977) and Myers and Majluf (1984).

Ross (1977) signalling argument for debt

The crux of Ross' argument is as follows. Assume firms differ according to their future cash-flow prospects. High-quality firms expect large future cash flows, whereas low-quality firms expect cash flows to be small. Firm quality is not observable to outsiders to the firm. The managers of high-quality firms have an incentive to attempt to signal their quality to the market, as in the absence of signals the market can't distinguish high- and low-quality firms and will value them identically. One way the management can signal is through debt policy. High-quality firms choose large leverage ratios and lower quality firms choose low leverage ratios. The market can observe leverage and hence values firms accordingly (assigning firm values increasing in leverage.) Leverage is a credible signal, as it is assumed that firm managers are averse (in terms of their own utility) to bankruptcy. High levels of debt imply a higher probability of bankruptcy, and only managers in charge of high-quality firms are willing to expose themselves to this probability. The preceding intuition can be formalised with the following model, which is a simplified version of that contained in Ross (1977). Assume a population of firms, each of which has future cash flow that is uniformly distributed.2 Firm quality varies, as the upper bound of the cash flow distribution (call this parameter t) varies across firms (i.e. a high-quality firm may have cash flow distributed on [0,t ] and a low-quality firm might have cash flow distributed on [0,t ] where t exceeds t ). Manager of firms know the value of t for their own firms, but the market as a whole does not. Managerial utility is increasing in date 0 firm value and date 1 firm value,~but is decreasing in the expected cost of bankruptcy. In line with the prior argument, managers will try to use debt to signal their quality. However, non-zero debt levels imply that bankruptcy is possible. Hence, we can write the managerial optimisation problem as follows:
7.1

If cash flow is uniformly distributed on lajijit means that the probability density of cash flow is constant from a fo b and zero elsewhere. This implies that the probability distribution function of cash flow is F(x)=(xajffb-aj lot x between <? and A

where we have assumed firm quality of t, V(B) is date 0 firm value, L is a parameter reflecting the cost (in managerial utility terms) of bankruptcy and y is a weight parameter. Given that the manager knows the true distribution of firm cash flow, his assessment of date 1 firm value is t/2. Similarly, if a debt level of B is chosen, the manager knows the firm will be bankrupt with probability B/t and the expected utility cost of bankruptcy is hence LB/t. Assume that the market assigns a firm with debt level B a date 0 value of f(B). Substituting this into 7.1 gives:

79

7.2

To compute the optimal level of debt, we compute the first order condition of 7.2 with respect to B. After rearrangement this yields:
7.3

(1-

Finally, we assume that in equilibrium, the market's beliefs about firm quality (based on a firm's debt level) are correct. Hence, we have the condition f(B(t)) =y where we have also acknowledged the dependence of the debt level, B, on firm quality through managerial actions. Differentiating this condition yields:
7.4

Substituting/' from 7.4 into 7.3 yields the following differential equation:

7.5

B' (t) =

2yL

This differential equation has the following general solution:

7.6

B' (t) =

+c

where c is a constant term. The constant c can be assigned a value through noting that the lowest quality firm in the population has no incentive to signal and will hence elect not to have any debt. Denoting the lowest quality by t , use of this intuition in 7.6 gives:

\
7.6 c=
Substitution of 7.7 in 7.6 gives the final debt rule:

B'(t) =

Equation 7.8 gives us the key results from the Ross (1977) model. Debt level (B) is increasing in firm quality (t). Clearly then, firms with higher debt levels will have greater date 0 market values and MM1 is violated once more. In more loose terms, the arguments in Ross (1977) are that debt is a costly signal (due to the possibility of bankruptcy it entails), and hence its use implies higher-quality firms. From an empirical standpoint, evidence that supports this notion can be found in Masulis (1983). This paper demonstrates that firms which swap debt for equity (hence increasing leverage) experience positive stock price returns whereas firms swapping equity for debt experience negative stock returns. The stock price reactions are interpreted as implying that leverage increasing transactions are good news whereas leverage decreasing transactions are bad news, consistent with the asymmetric information story.

The Myers-Majluf (1984) pecking order theory of finance Another study that generates departures from MM1 through information asymmetries is Myers and Majluf (1984). Although Ross focuses on the level of the debt-equity ratio as a signal of firm quality, Myers-Majluf concentrate on the information revealed by security issues. The intuition behind their arguments is as follows. We start by assuming a population of firms differing in both the quality (value) of their assets in place and the quality (net present value) of their investment projects. Any investment project has to be financed through an issue of equity. Assume also that the managers of any firm are better informed about both the quality of their firm's assets in place and the quality of their firm's investment project than are outsiders. Furthermore, assume that managers act in the interests of their firm's existing equity-holders. Only managers know whether the equity of their firm is over- or under-priced though, and this creates an opportunity for them to exploit the market in order for existing shareholders to profit. The existence of information asymmetries thus implies that the market can misprice corporate equity: some firms' equity may be over-priced and others will be under-priced. In this setting, managers may raise equity for two reasons: They may wish to invest in a positive net present value investment, which would result in an increase in the value of the firm's equity. Alternatively, they may wish to issue overpriced equity, which would result in a transfer of wealth from the new to the old equity-holders. Given rational expectations, the financial market correctly recognises both incentives to raise equity. In equilibrium, managers of low-quality firms (i.e. managers of firms with assets in place whose true worth is low enough - and are hence overvalued), raise equity in order to take projects with a small but negative net present value. The benefit to the existing equity-holders that results from issuing overvalued equity exceeds the cost resulting from taking the negative net present value project. Similarly, managers of high-quality firms (i.e. managers of firms with assets in place whose true worth is high enough - and are hence undervalued), abstain from raising equity and hence from taking projects with a small but positive net present value. The dilution to the existing equity-holders that results from issuing undervalued equity exceeds the benefit resulting from the positive net present value generated by taking the project. The presence of information asymmetries between managers and equity-holders hence leads to distortions in investments. Issue decisions affect prices as they reveal information on firm quality. Managers are more likely to issue equity when their firm's assets in place are overvalued, as opposed to undervalued. On average, equity issues thus lead to stock price drops. Furthermore, the highest quality firms avoid issues at all costs. Generalising the above somewhat, we can fit risk-less debt, risky debt and other securities into our pecking order. Obviously, issuing risk-less debt to finance investments conveys no information to the market, as there is no possibility of exploitation (as there is no risk). Thus, stock prices should not react to risk-less debt issues and the highest quality.firms will issue risk-less debt in order to finance any investments. Low-quality firms don't issue riskless debt, as they cannot exploit new investors through its issue. Risky debt comes with a possibility of default and hence could be overpriced if the market underestimates the probability of default. Issues of risky debt, therefore, convey some information, but clearly less than issues of equity.

 * Putting this all together leads to a model in which equity issues cause stock prices to drop a lot (as the market infers that firms that issue are very poor quality), risky debt issues cause small price decreases (as fairly low-quality firms issue risky debt) and risk-less debt issues cause no price impact (as only high-quality firms issue risk-less debt). Hence, in a dynamic sense, Myers-Majluf implies that capital structure decisions do affect firm values. This is the pecking order theory of finance. There is a fair amount of empirical evidence that supports the pecking order theory. First, the event study results on exchange offers detailed above are consistent with the pecking order theory. Second, event study evidence on new security issues confirms the theory too. Common stock issues lead to price impacts of around -3 per cent, for example, whereas risky debt issues cause small price drops, which are not statistically different from zero. Hence, the intuition that underlies the model is regarded by many as very plausible. * 1

Summary
In this chapter we have examined theoretical models (and examples), which imply that firm value does depend on the financing choices it makes and on capital structure choices in particular. First we examined arguments based on agency costs and then looked at a model of asymmetric information. The empirical evidence for these models is mixed. Evidence for agency problems can be found in the specification of corporate debt contracts, which contain clauses aimed specifically at preventing debt overhang and asset substitution problems. The previously discussed evidence on exchange offers is supportive of asymmetric information models (although it would contradict the implications of a debt overhang model). Research in these areas still . proceeds. The most recent strand of literature on capital structure builds on the agency cost approach and examines incomplete contracts as the source of violations of MM1.

'

J
82