Jensen and Meckling _1976_ This paper deals with the effects of by keara


									Topic 8 – Agency theory - Copeland, Weston, Shastri – Chapter 12 (pages 439-461, and especially 456-460),
Chapter 18 (page 766)

      In this topic, we briefly discuss the topic of agency theory. Agents work for principals. Obviously, the principal

wants the agent to work in his or her best interest. Agents, on the other hand, have the incentive to work in their own

best interests. If the interests of these two parties are perfectly aligned, then there are no agency problems. When

they are not, agency problems exist.

      In corporate finance, there are two main types of agency problems: (1) agency problems associated with

stockholder-bondholder conflicts, and (2) agency problems associated with manager-stockholder conflicts. In the

first type, the agent is the manager who is assumed to be perfectly aligned with stockholders. The principal is the

bondholders. In the second, managers again are the agent, but work in their own best interests. The principal is the


      Myers, Stewart. “Determinants of corporate borrowing.” Journal of Financial Economics 5 (1977), 147-

175 addresses stockholder-bondholder conflicts and the problem of underinvestment for firms with risky debt.

      Assume that capital markets are perfect and complete and managers maximize shareholder wealth (i.e., no

manager-stockholder conflicts). (Since managers’ and stockholders’ interests are perfectly aligned, we will assume

firms decisions are made by stockholders.) In general:

      V A  VG  VD  VE  V                                                                                       (1)

where the value of the firm is the value of the debt plus equity and also the value of the existing assets and growth


      Consider the inception of the firm, i.e., no assets in place and only growth opportunities (to be funded in one

period t = 1) and also assume the firm is all equity and only equity financing is permitted. So, at time 0:

      0  VG  0  VE  V                                                                                          (2)

      This growth opportunity must be funded at t = 1. If not, it evaporates (and has no value). Also assume that the

growth opportunity is firm specific (it cannot be sold or its going concern value is greater than its liquidation value).

Are these reasonable assumptions?

      The firm will fund the project if its NPV is positive. This occurs if the PV of the project, as of t = 1, V(s), the

PV (at time 1) of the project in state s, is greater than the required investment, I. The value of the firm today (t = 0)

       V   q(s)[V (s)  I ]ds                                                                                      (3)

where sa is the lowest state in which the NPV is positive (states are ordered from lowest to highest) and q(s) is the

PV (at time 0) of $1 in state s. Why is the discount rate dependent on the state?

       Now consider debt issues prior to t = 1 (i.e., before state of nature is known), say at time 0. Since, at this time,

the firm has only growth opportunities, then it can only issue risky debt. For instance, if the realized state is less than

sa, then the firm will not invest in the project and the value of the firm will be zero (so the stock and bonds are


       0  VG  VD  VE  V                                                                                          (4)

       More specifically, assume that the firm (at t = 0) issues a zero-coupon risky bond (with face amount P). Also

assume that the debt matures after the state of nature is known, but before the investment is made in the project (i.e.,

right before t = 1). Then:

       (1) If V(s)  I + P the stockholders will payoff the bondholders and finance the project.

       (2) If V(s) < I + P and V(s)  I, then bondholders will undertake the project. (Why the bondholders?)

       (3) If V(s) < I the project will be abandoned.

       In this situation, the value of the firm is unaffected by the amount borrowed. Sometimes the bondholders

undertake the project, other times the stockholders, but in any case, the project will be undertaken if and only if V(s)

 I.

       Now, consider the case where the bond matures after the project is selected or rejected (say, just after t = 1).

This is the interesting case. Here, the stockholders will only undertake the project if V(s)  I + P. In other words, if

the project cannot recover the new investment required by the stockholders and the promised payment to

bondholders, then there is no net benefit to the stockholders. In fact, there is a net loss since stockholders invest an

additional $I and get zero back. The market value of the firm is now:

       V   q(s)[V (s)  I ]ds                                                                                      (5)

where sb is the minimum state such that V(s)  I + P. Notice that the higher P, the higher sb, which reduces V.

       Figure 3 in the paper shows the value of all equity financing (no loss) and the value given debt financing. Note

that if the firm selects the project, then the debt is risk-free (i.e., the value of the firm in states greater than b is more

than P), but is worthless if the project is dropped. Therefore:

     VD   Pq(s)ds                                                                                           (6)

     1) What is VD when P = 0?

     2) What happens when P is increased? Consider the impact on P and sb.

     3) What happens as P approaches infinity?

     This implies a unique maximum amount of value of the debt, and therefore a maximum amount that the firm

can raise through a debt issue.

     This also implies that the optimal capital structure is all equity. However, as discussed in Jensen and Meckling

(1976), there are agency costs associated with equity, so the optimal capital structure could involve minimizing the

combined agency costs. (In the appendix, Myers explores the tax argument for the use of some debt. We will skip

this part.)

     In summary, there is an underinvestment problem created when (1) the firm has debt outstanding which was

issued before the NPV of the project is known, and (2) the debt matures after the date required for project funding.

Stockholders have no incentive to fund the project if the NPV is less than the amount owed to the bondholders.

     Who is hurt by this underinvestment problem? Are bondholder’s hurt? What about stockholders?

     What is the dollar value of the agency cost in the Myer (1977)?

     Can’t we come up with a way to avoid this underinvestment problem? Here are some possibilities:

     (1) Include in the bond indenture that the firm (i.e., stockholders) must finance all positive NPV projects.

               a.   Limited liability problem

               b.   Free rider problem

               c.   NPV isn’t directly observable. It’s estimated. Who’s going to estimate project NPV?

               d.   Project NPV is a factor of market-wide and firm-specific conditions. Management could easily

                    manipulate firm-specific conditions to push the NPV below I (assuming V(s) is less than I + P).

     (2) Promise to take all future investments (with I in escrow). This takes managerial discretion out of the

          analysis. But, what is the problem with promising to take all projects? Any associated loss in value can be

          greater than the value lost by forgoing positive NPV investments. (Also, as described above, management

          has incentive to push NPV down if V(s) is less than I + P, with the hope of eventually getting the escrow

         money back.)

    (3) Renegotiate the debt contract or perhaps have one party buy out the other or a third party buy out the stock

         and bondholders.

             a.   Renegotiation is costly

             b.   Both parties have to be equally informed about the project NPV

             c.   With superior information, managers have the incentive to down play the value of the investment

                  in order to extract more. In addition, even if V(s)  I + P, the manager might say that V(s) < I + P

                  and extract value anyway.

             (Comment: In perfect capital markets, with no informational advantage, then costless renegotiation of

             the debt contract will eliminate the problem. In an imperfect capital market, this problem can occur in


    (4) Shorten the maturity date of the debt contract. As we saw, debt maturing before the investment is made

         does not cause a problem. Also, short term debt that is rolled over allows for continual renegotiation.

    (5) Mediation. Problem: when do you call in the mediator? Stockholders won't disclose that they are about to

         make a suboptimal decision, and they are not likely to permit mediation totally at the bondholders option.

         To know when to mediate, the bondholders need to monitor, which costs money.

    (6) Dividend restrictions will cause the firm to start storing up cash. Without the ability to pay them out, the

         manager will select the positive NPV investment rather than investing in the capital markets (at a zero or

         negative NPV). Having a lot of cash can cause problems as discussed in our next paper (Jensen, 1986), (b)

         if the capital markets provide a negative NPV and if the firm has no good investments, then it forces

         investment at a negative NPV. (Why would the capital markets provide a negative NPV for corporations?)

    (7) Honesty. Thus, while helped in the short term by taking a suboptimal strategy, the firm is hurt in the long

         term by higher interest rates or increased monitoring costs (which are borne by stockholders). Therefore, an

         on-going business will probably earn valuable reputational capital by taking all positive NPV investments,

         thus decreasing these costs.

    Multi-period setting: In the analysis above, we assume that the firm has zero-coupon debt outstanding (payment

= P) where the debt matures either right before or right after the investment decision is made. Assume now that debt

is long term. Similar to the previous case, investment in a positive NPV project increases firm value by ΔV(s). The

net present value of the investment is ΔV(s) - I(s). However, so long as the riskiness of the firm's cash flows do not

increase, the investment is likely to also increase the value of the bond, ΔVD(s). (Note that if the bond is risk free,

ΔVD(s) = 0.) The change in value of the firm's equity due to the investment is ΔVE(s) = ΔV(s) - ΔVD(s). The NPV of

the investment – with respect to the stockholders – is ΔVE(s) - I(s) = ΔV(s) - ΔVD(s) - I(s). Therefore, the firm will

only issue new equity to finance the project if ΔV(s)  ΔVD(s) + I(s), instead of the optimal investment rule which

would call for investment whenever ΔV(s)  I(s). Fig. 4 in the paper represents management's investment strategy.

    If the variance of the firm decreases with the investment, then ΔVD(s) is even higher and the problem increases.

However, if the variance increases, then it is possible that ΔVD(s) is negative and it could encourage firms to select

negative NPV projects (as discussed in Jensen and Meckling, 1976). An interesting conclusion is that if the firm has

risky debt, then the higher the risk of the investment opportunity, the more likely they will acquire funds to finance

the project. In addition, borrowed funds (without perfectly enforced me-first rules) will probably lower ΔVD, further

increasing the desire to borrow to fund risky projects.

    Borrowing against a portfolio of assets. Conventional wisdom says that a conglomerate has higher debt capacity

than a single business firm (due to the conglomerate's diversification benefits). Indeed, merger of two firms

combines the projects of each firm into one portfolio. This decreases the variance if the returns from the two assets

are not perfectly correlated. But the problem with the merger is that the combined firm has the debt of each of the

individual firms. Before merger, firm i would invest only if Vi(s)-Ii  Pi and firm j would invest only if Vj(s)-Ij  Pj.

Now the condition is that Vi(s)-Ii + Vj(s)-Ij  Pi + Pj (and both projects, individually have a positive NPV) then the

firm will issue to finance both.

    Assume that firm i and firm j have positive NPV projects and prior to the merger, firm i would not have issued

(because Pi is large), but firm j would have issued. There are a couple of possibilities. One is that the additional debt

from firm i in the merged firm would prevent the issuing and financing of either project (bad). The other possibility

is that the additional NPV from project j would be enough to allow project i to be financed (good). (Note that the

merged firm will not finance only one project (which is possible before the firms were merged) unless only one of

the project had a positive NPV.)

    In example A in the paper, merger of the two firms allows positive NPV project i to be selected:

    NPV(i) = 50, P(i) = 100                    NPV(j) = 80, P(j) = 20

So, $130 > $120 (when it would not have been in a separate company, $50 < $100).

    In case B, the merger takes away the incentive for the merged firm to undertake project j.

    NPV(i) = 50, P(i) = 100                     NPV(j) = 60, P(j) = 20

In this case, $110 < $120, so neither project is accepted. Obviously, the merger would not occur in case B.

    We continue the discussion of agency costs with Jensen, Michael. “Agency costs of free cash flow, corporate

finance, and takeovers.” American Economic Review 76 (1986), 323-329.

    As the title suggests, this paper explores manager-stockholder conflicts and the agency costs of free cash flow.

Free cash flow is the amount of the firm’s cash flow (i.e., cash inflows less cash outflows) in excess of that needed

to fund all positive NPV projects. Free cash flow appears to include not only excess “cash flow,” but also existing

cash (and marketable securities) and borrowing capacity.

    Agency costs are caused when free cash flow is invested in projects with negative NPVs or wasted in other


    Why would management have an incentive to invest free cash flow in a negative NPV project?

         Increase in resources under their control

         Increase in compensation

         Rewards to middle-level managers through promotion

         Investment in negative NPV projects creates agency costs only if there are manager-stockholder conflicts

                 (or conflicts with some other “principal”). Why?

    Free cash flow agency problems are most likely in firms with

         Large free cash flow

         Few positive NPV projects

    One way to reduce these agency problems is to disgorge the excess current cash, and promise to pay out any

future free cash flow.

    A current dividend can get rid of the current excess in cash.

    What about future free cash flow? What is the problem with “promising” to maintain dividends at a high level

in the future?

    A capital structure composed of a large amount of debt can reduce the agency problems associated with free

cash flow.

    For instance, instead of a promise to pay future dividends, the firm could issue debt in exchange for stock

         Since this is an exchange, there is no new cash brought into the firm (i.e., no increase in free cash flow)

         Since debt payments are mandatory, the promise to make increased future payments is believable.

    Notice that reducing free-cash flow agency costs is a benefit of the debt-for-equity exchange (plus, there are

some tax benefits). However, there are problems with additional debt in the capital structure. In review, what are

some these problems?

Empirical evidence

    Leverage increasing transactions results in an increase in stock prices (on average) and leverage decreasing

transactions cause a decrease in stock prices (on average)

         What else could cause these stock price reactions?

         Note – the announcement of the sale of new debt (i.e., not an exchange) is associated with an insignificant

         stock price increase. How does this fit with the free cash flow hypothesis?

    Leverage buyout / going-private transactions have large free cash flow agency cost reducing benefits

         LBOs significantly increase the leverage in the firm.

         Typically LBO firms are large firms, with stable business histories, with substantial free cash flow

              That is, they are low-growth, high free-cash flow firms

         LBOs often use strip financing (i.e., roughly proportional ownership in debt and equity securities) to reduce

         agency problems of debt. (Why would this reduce debt agency costs? Think agency costs associated with

         transfers of wealth between bondholders and stockholders.)

              See note about IRS regulations and interest deductibility for strictly proportional ownership

    Evidence from the oil industry

         In the late 1970s to early 1980s, the oil industry had large increases in free cash flow. Yet evidence shows

         that they continued to invest in exploration and development even though the rate of return probably had a

         negative NPV (see McConnell and Muscarella (1986)), rather than payout the excess cash to stockholders.

Free cash flow and takeovers

    Firms with large amounts of free cash flow have incentives to use the cash on mergers and takeovers that have

    little benefit (or that actually reduce value)

One positive aspect of a cash tender offer (as opposed to a stock) is that cash is disgorged to the target firm’s

stockholders, rather than retained in the firm and invested in negative NPV projects.


To top