Reserved

FE07CH01-Myers ARI 6 November 2015 16:51

Finance, Theoretical and Applied Stewart C. Myers Sloan School of Management, Massachusetts Institute of Technology, Cambridge, Massachusetts 02142; email: [email protected]

Annu. Rev. Financ. Econ. 2015. 7:1–34

The Annual Review of Financial Economics is online at financial.annualreviews.org

This article’s doi: 10.1146/annurev-financial-111914-042056

Copyright c© 2015 by Annual Reviews. All rights reserved

JEL codes: G31, G32, G34, G35, G38

Keywords

corporate finance, capital structure, valuation, regulation, real options

Abstract

Like Caesar’s Gaul, corporate finance is divided into three parts: theoretical, empirical, and normative. Important advances in any one of these three typically generate good ideas for the other two. I have been fortunate not to specialize in one part only. This review covers the history of capital structure theories, including the trade-off and pecking-order theories, and takes a skeptical view of how those theories have been tested so far. I give roughly equal space to normative, practical applications, including adjusted present value (APV), the valuation of real options, and the application of modern finance to regulation, insurance, the valuation of R&D, and the role of risk capital in financial institutions. Looking back, I realize that the supply of intriguing financial questions is inexhaustible.

1

Click here to view this article's online features:

• Download figures as PPT slides • Navigate linked references • Download citations • Explore related articles • Search keywords

ANNUAL REVIEWS Further

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

1. INTRODUCTION

I got into finance because it was boring. I was taking an MBA finance class at the Stanford Business School: just one Harvard Business School case after another. No doubt the cases taught useful practical lessons, but the next case never built on the last one. There was not much there intellectually.

Alex Robichek called me into his office. He demanded to know why I had cut several classes. I told him that his class was boring. He probed and learned that I was taking a doctoral course in macroeconomics. He gave me a stack of research articles in corporate finance and told me to report back next week.

I found financial research interesting and surprisingly accessible, compared with macroeco- nomics. (I remain to this day macroeconomically challenged.) Several months later, in June 1964, I graduated from the Stanford MBA program and started in the doctoral program. Alex put me to work immediately as his research assistant helping with a manuscript on corporate financing. Our book, Optimal Financing Decisions, was published a year later (Robicheck & Myers 1965). Alex and I also collaborated on three journal articles, discussed below. I defended my dissertation and started at the MIT Sloan School of Management in 1966, two years after receiving my MBA.

One can read between these lines how much I owe Alex. He was the thoughtful sponsor, I the less-than-ideal protégé—a young man with brains and commitment but an instinct to disagree. Alex’s and my research tracks diverged after I left for MIT, but he took great pride in my career.

2. PREVIEW

This review proceeds in three acts. Act 1 starts at the beginning, reviewing my coauthored book and papers with Alex and our development of the trade-off theory of optimal capital structure. Then I turn to my dissertation and early articles, which among other things showed why corporate diversification adds no value for public firms in well-functioning financial markets.

I am probably best known for work on capital structure. Act 1 continues with that topic, covering the following:

1. My 1974 paper on taxes, the weighted average cost of capital (WACC), and the adjusted present value (APV) rule for capital-investment decisions (Myers 1974)

2. “Determinants of Corporate Borrowing,” my 1977 paper on debt overhang and conflicts of interest between stockholders and holders of risky debt (Myers 1977)

3. Information, security issues, and the pecking-order theory, based on Myers & Majluf (1984)

I comment on empirical work, mostly by others, and on the performance of the trade-off and pecking-order theories.

Act 2 covers applied and how-to-do-it topics, starting with the Brealey-Myers (now Brealey- Myers-Allen) corporate-finance text. Then I review my research on valuing real options, including investments in pharmaceutical R&D, and on the implications of real options for corporate strategy. I was among the first to apply modern finance to rate of return regulation, for traditional utilities; later for oil pipelines, where I introduced the trended original cost method; and for railroad tariffs based on stand-alone cost. I tackled valuation problems raised by litigation. At the end of Act 2, I review pricing and capital allocation in insurance and the role of risk capital in financial firms. Act 3 returns to theory, including my recent research on dynamic agency models and their implications for corporate governance and the classic big issues of corporate finance.

2 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

3. ACT 1: VALUATION AND CAPITAL STRUCTURE

3.1. Robichek-Myers and the Trade-Off Theory

The first task set for me as Alex’s research assistant was to figure out what was right and wrong with proposition 1 of Modigliani-Miller (MM), which states that the mix of debt and equity financing is value irrelevant (Modigliani & Miller 1958). This was another lucky break, for I was forced to master MM with no recourse to easy intuition, now expressed by analogies, such as “the value of a pie does not depend on how it is sliced.” I was forced to search for first principles at a time when most finance professors—and nearly all practitioners—believed that there must be something wrong with MM’s proofs, which were considered counterintuitive.

Why did professors and practitioners have such a hard time signing on to MM’s argument, at least in principle? A review of the difficulties they faced will help to understand the state of play in corporate finance in the mid-1960s:

1. Professors and practitioners did not always clearly distinguish maximizing market value from maximizing personal utility. A CEO’s personal risk aversion could determine optimal borrowing if the firm’s decisions depended only on the CEO’s utility function. Shareholders’ personal risk aversion could limit corporate borrowing if shareholders were locked in and unable to manage portfolio risk.1 Professors and practitioners should instead have considered risk from the viewpoint of a rational investor participating in financial markets. Robichek & Myers (1966c) advanced this viewpoint.

2. MM focused on the cost of capital, defined as a weighted average of the costs of debt and equity. Today, use of after-tax WACCs to discount cash flows from long-lived assets is routine. But the necessary conditions for accurate valuation using WACC were not well understood, and MM did not assist.2

3. MM’s reliance on “homemade leverage” by shareholders seemed artificial and unrealistic. In addition, lacking an asset pricing model, MM had to place firms in “risk classes,” which they defined in an extremely restrictive way. With time, it became clear that MM’s debt- irrelevance result holds in any equilibrium asset pricing model where investors are free to lend or borrow. Frictions or segmented markets can upset the MM result, of course, but it was not necessary to imagine mom-and-pop investors trading on personal account to undo changes in corporate leverage.

4. Professors and practitioners often confused financial risk with the risk of default. They assumed that the expected rate of return on common stock would not increase with leverage as fast as MM’s proposition 2 predicts, provided that leverage is moderate and default risk remote. Therefore, they thought that WACC should decline with moderate levels of debt. MM did not introduce default risk, which left the impression that they had skated by an important objection to their proofs.

5. MM recognized the value of interest tax shields but introduced no costs that would limit leverage. Thus, they appeared to predict runaway leverage.

6. MM’s conclusions defied everyday experience. The value of an actual pie does depend on how it is sliced. The sum of the slices sells for more than the uncut whole. David Durand

1 Schwartz (1959) relies on shareholder risk aversion to determine optimal corporate borrowing. Note the indifference curves in his figures IIIB and IIIC. 2 MM’s costs of capital were really cap rates, which they defined as ratios of expected average future cash flow or income to present value. Cap rates equal expected rates of return only for level perpetuities. MM’s cost of capital formulas are easiest to interpret for level perpetuities, but errors creep in when taxes are introduced and assets have irregular cash flows and limited lives (see my discussion below of Myers 1974).

www.annualreviews.org • Finance, Theoretical and Applied 3

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

pointed out that you pay more for skim milk plus cream than for the equivalent volume of whole milk.3

Difficulty 6 has an easy response if one thinks through grocery-store equilibria. There are two necessary conditions for the sum of the pieces to sell for more than the whole. On the demand side, customers must be willing to pay extra for separated products, for slices versus the whole pizza or for skim milk and cream versus whole milk. But there must also be a cost of supplying the sliced or separated products. Otherwise, competition would drive the price of the separated products back to an MM equilibrium.

These supply-demand conditions are one way to understand MM’s argument. There is clearly demand for different corporate securities, for example, for debt and levered equity instead of plain- vanilla unlevered equity. But the marginal cost of supplying the debt and equity separately is a small fraction of the market value of the firm, at least for public corporations operating in tolerably efficient financial markets. Therefore, if corporations could generate extra value by issuing a mix of debt and levered equity, the supply of the mix most demanded by investors would increase rapidly until the value premium is eliminated.

The history of mortgage-backed securities (MBSs) provides a dramatic example. Banks had a choice of selling the cash flows from pools of mortgages as simple pass-through securities or as collateralized mortgage obligations (CMOs). A pass-through MBS amounts to all-equity financing of the pool. A CMO allocates the cash flows to tranches, for example, AAA tranches backed up by tranches with lower ratings and a final residual claim of “toxic waste.” The tranches amount to several classes of debt and equity. The ability to make money issuing CMOs versus pass-through MBSs violated MM’s proposition that capital structure is irrelevant to value. What then was the supply response? Massive. Perhaps it overshot the MM equilibrium. Of course, the manufacture of CMOs will continue, because there is an equilibrium supply. The manufacturing profits will cover manufacturing costs. If the manufacturing costs are small, MM’s proposition 1 will hold to a good approximation.

Robichek & Myers (1966b) did not make this supply-side argument, but they did answer, or at least clarify, several of the other difficulties listed above. The first step was to set the objective as maximizing market value. We introduced a valuation framework based on Arrow-Debreu state prices, following Arrow (1964) and Hirshleifer (1965, 1966). This state-preference valuation model was imported from my dissertation (Myers 1967). MM’s risk classes were discarded.

MM’s proposition 1 followed easily from the model, provided that financial markets are suf- ficiently complete.4 Leverage is irrelevant to value because a dollar delivered in a future state of nature has an equilibrium present value (PV), which does not depend on whether the dollar is delivered as debt service or as a payout to shareholders. We showed that proposition 1 holds for risky or safe debt and for any mix of debt instruments, including a mix of maturities.

Previous critics of MM had quarreled with their assumption of perfect financial markets. We focused instead on MM’s implicit assumption of no feedback from leverage to the cash returns on the firm’s assets. We pointed out three feedback channels: (a) costs of bankruptcy or reorganization; (b) distortion of capital investment; and (c) the possibility that high leverage would force the firm to accept future financing transactions with negative net present value (NPV), for example, because of monitoring costs incurred by future lenders and charged by them back to the firm.

3 Oral history reports that when Modigliani & Miller (1958) was published, MIT statistician David Durand immediately went to the local Stop & Shop supermarket to verify that whole milk sold for less than the equivalent mix of skim milk and cream. 4 Markets are sufficiently complete if investors can buy any possible slice of the payoffs from the firm’s assets. Today we can point out that the ability to write and trade derivatives on asset payoffs assures sufficient completeness.

4 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

The trade-off of interest tax shields against these three costs of financial distress gives the trade-off theory of capital structure. We were the first to set out this theory, although the label “trade-off ” did not arrive until Myers (1984b). The figure we used to explain the theory seems ubiquitous today (see Robichek & Myers 1966b, figure 1b, p. 21).

3.2. Risk-Adjusted Discount Rates

Robichek & Myers (1966a) clarified what happens in discounted cash flow (DCF) valuations when expected future cash flows are discounted at a constant risk-adjusted discount rate. DCF gives each future cash flow a haircut for risk. The haircut for the cash flow at date t is the ratio of its certainty equivalent to its expected value. DCF assumes that the certainty-equivalent haircut declines at a constant rate with cash-flow maturity. This makes sense only if the proportional risk of the cash-flow stream is the same in each future period, so that cumulative risk increases at a constant rate as one looks further out in future time.

Suppose, for example, that an oil field generates cash flows for 20 years. If the capital asset pricing model (CAPM) holds and the field’s beta will be the same in each of the 20 years, then the market value of the cash flows can be calculated by discounting at the expected CAPM return.5

Though simple, Robichek & Myers (1966a) has important implications. First, it corrects casual intuition, which is inclined to say that “long-lived cash-flow streams are riskier and should be discounted at a higher rate than short-lived streams.” The standard DCF setup automatically gives distant cash flows greater haircuts for risk.

Second, fudge factors should never be added to risk-adjusted discount rates. Too often this is done without any thought of the implied risk haircuts on distant cash flows (see the cautionary example in Brealey, Myers & Allen 2013, p. 231).

Third, what if the certainty-equivalent haircut does not increase at a constant rate? Examples are easy to find. Suppose that the oil field’s beta depends on the period-by-period covariances of unexpected changes in spot oil prices with the market portfolio. If spot oil prices follow a random walk with constant drift, all is well. But oil prices are probably mean reverting. If so, a risk-adjusted discount rate based on the spot beta will value next period’s production correctly, but it will understate the PV of 20 years’ production. With mean reversion, cumulative risk does not increase at a constant rate as one looks further out in future time (see Bhattacharya 1978, Myers & Turnbull 1977).

The assumptions required for DCF valuations (with the same risk-adjusted discount rate for all future periods) imply that cash-flow distributions must be right skewed and that the skewness must increase for more distant cash flows (see Fama 1977, 1996). For example, if the time series of cash flows follows a geometric random walk, then cash-flow distributions will be lognormal. Yet practitioners often try to discount most likely (modal) cash flows, implicitly assuming symmetrical distributions. If the true distributions are, say, lognormal, then the practitioners are understating value, because expected cash flows exceed modal cash flows. If the true distributions are indeed symmetrical, then discounting at a constant rate is in principle incorrect. This problem awaits a practical solution.

3.3. Security Valuation and Capital-Investment Decisions

Robichek & Myers (1966c) argued that corporate finance should focus on how risk affects equilib- rium asset values in financial markets. This focus is essential in corporate finance if the objective

5 Of course, such CAPM applications carry additional baggage, including the assumption of a constant risk-free interest rate and potential errors from noisy estimates of the expected market risk premium and beta.

www.annualreviews.org • Finance, Theoretical and Applied 5

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

is to maximize the market value of the firm. Our paper was exploratory and mostly qualitative but generally on target. It also contains the first statement of the CAPM formula r = rf + β(rM − rf ), where r is the expected return for a security with risk β, rM is the expected return on the efficient (market) portfolio, and rf is the risk-free rate. Oddly enough, this formula does not appear in Sharpe (1965).6

The financial implications of the state-reference framework used in Robichek & Myers (1966b,c) were worked out more comprehensively in my dissertation (Myers 1967) and in two journal articles.

Myers (1968a) derived implications for security valuation. In addition to equilibrium pricing in complete markets, I also covered short-selling restrictions, incomplete markets, and linkages to conventional valuation frameworks, principally DCF.

Myers (1968b) applied the valuation framework to capital-investment decisions. I set out the conditions for investment projects to be risk-independent, that is, the conditions for project values to add up, regardless of whether or how project returns are correlated. (I should have chosen the label “value additivity” instead of “risk independence.”) The argument for risk independence was not just a mechanical application of Arrow-Debreu prices in complete markets. For example, I presented a proof assuming that investors had access to lists of risk-equivalent securities, that is, securities that were not perfectly correlated but nevertheless perfect substitutes in diversified portfolios. I also gave a general proof that a merger of two firms A and B cannot add market value just by creating the more diversified portfolio AB—an obvious point, perhaps, but worth making at the time, when conglomerates were popular, partly because of assumed benefits of diversification. This proof does not depend on any stringent assumption of complete markets; the market is always complete enough, because A and B are traded separately premerger.

Value additivity, like MM’s proposition 1, seems obvious today. Financial managers value projects one by one, usually by DCF, with no thought or worry that a portfolio of projects might be worth more than the sum of individual project values.7 People say casually, “Corporate diversification is redundant, because investors can diversify on their own.” But this intuition was not seen in the 1960s. Capital budgeting was then viewed as a problem of selecting mean-variance efficient portfolios. For example, Lintner (1965, p. 65, emphasis added) asserted that “the problem of determining the best capital budget . . . is formally identical to the solution of a security portfolio analysis.”8 I believe that I was the first to show why this view was wrong.

Myers (1968b) did not have the impact it deserved. First, the CAPM made it easy to see that diversifiable risk does not matter for valuation. Second, I was naive about publication strategy. I agreed to publish the paper in Industrial Management Review, a good journal that was, however, not widely read by financial economists. Paul Samuelson had recently published in that journal, why not me? Of course, Paul’s papers were noticed regardless of where they were published. That was not true for an assistant professor.

3.4. Adjusted Present Value

Myers (1974) works out the implications of MM’s theory for valuing capital-investment projects under different assumptions about financing. Before I get to that paper, it’s worth reviewing the issues raised by MM’s assumptions about debt and taxes. The issues are often not understood today.

6 Robichek and Myers (1966c), pp. 218–219, note 10. 7 Assuming, of course, that there are no synergies across projects and that investment in project A does not change cash flows from any other project B. 8 Van Horne (1966) is another example.

6 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

MM recognized the value of corporate interest tax shields9 but took five years before finally deciding how the tax shields should be valued. First, they valued the tax shields at the same rate used to value after-tax operating cash flows. But they assumed fixed, permanent debt. Therefore, in the MM “correction” (Modigliani & Miller 1963), they decided that the tax shields should likewise be assumed fixed and safe and should be valued at the debt rate. (They should have relaxed the assumption of fixed, perpetual debt.) Their “corrected” present value of interest tax shields (PVTS) on debt D is rDTC D/rD = TC D, where rD is the rate of return on debt and TC is the corporate tax rate.

Suppose that the firm’s assets are expected to generate a perpetual stream of expected cash flows C = E(Ct ) with PV C /rA, where rA is the risk-adjusted discount rate, defined as the opportunity cost of capital assuming all-equity financing.10 Then MM’s market-value balance sheet for the firm is:

V(all-equity) = C/r A D

PVTS = T C D E

V V

Now suppose that the firm has a new project requiring investment I to generate an additional cash flow of �C, financed at the same debt ratio λ = D/V as for the firm as a whole. In Modigliani & Miller (1963), NPV equals �C discounted by an after-tax WACC:

WACC = rD(1 − T C )λ + rE (1 − λ). (1) Also, WACC = rA(1 − λTC ). WACC declines with financial leverage, as one expects, given valuable interest tax shields. MM’s proposition 2 (Modigliani & Miller 1963, equation 12.c, p. 439) becomes

rE = r A + (1 − T C )(r A − rD)D/E. (2) Equations 1 and 2 look like a complete, neatly wrapped package. But the package leaves three important open questions, even if business risk and rA are held constant.

3.4.1. What if project cash flows are not perpetuities, but instead annuities with limited lives? Myers (1974) showed that discounting at MM’s WACC (using Equation 2 for rE ) does not produce correct valuations, although the errors are probably not material, given the difficulty of determining expected cash flows.

3.4.2. What if the debt ratio (D/V or λ) is held constant over time? Then the future debt level must vary as the firm or project does well or poorly. MM’s assumption of a fixed, permanent debt level—not a constant debt ratio—simplified exposition of their theory, but it sowed confusion in practice. If the amount of debt is fixed, then the debt ratio must fluctuate. If there is a meaningful

9 Miller (1977) argued that higher personal taxes on investors’ debt versus equity income could offset the value of corporate interest tax shields. Here I assume that the same personal tax rates apply to income from both debt and equity. This assumption is sufficient for MM’s propositions. 10 Notice that I have substituted cash-flow perpetuities and expected rates of return for MM’s average cash flows and cap rates (see Footnote 2 above).

www.annualreviews.org • Finance, Theoretical and Applied 7

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

target debt ratio—as the trade-off theory predicts—then the level of debt supported by the firm or project must be rebalanced in response to fluctuations in future value to keep the debt ratio at least roughly constant.

If the firm will keep the future debt ratio constant, then the standard WACC formula from Equation 1 works for any project life or cash-flow pattern. MM’s alternative formula WACC = rA(1 − λTC ) does not value projects accurately, however. [The correct formula is derived in Miles & Ezzell (1980).] Also the tax rate drops out of Equation 2, restoring MM’s original proposition 2:11

rE = r A + (r A − rD)D/E. (3) WACC still declines with financial leverage, but more slowly than under MM’s assumptions in Modigliani & Miller (1963).

In practice, WACC is used as a discount rate for projects with different, finite lives and different cash-flow patterns. The implicit assumption is therefore rebalancing and a constant debt ratio. Equation 3 must be used when calculating WACC at different debt ratios, even in the presence of valuable interest tax shields. It is therefore disturbing to find some practitioners still using Equation 2 in combination with WACC.12 See Brealey, Myers & Allen (2013, ch. 19) for a more complete explanation of the proper use of WACC and proposition 2.

3.4.3. What if the future debt ratio is not fixed, but instead changes over the life of the project? In this case, discounting at WACC does not work. Myers (1974) derived APV as a practical alternative:13

Project APV = NPV(assuming all-equity financing) + PV(interest tax shields). If the future debt ratio is fixed, then discounting at WACC generates APV in one step.

APV is useful in project financing, where debt is issued specifically for the project and paid down on a predetermined schedule. The project may be refinanced, but the refinancing date is distant. In this case, the interest tax shields would be fixed and discounted at rD. The APV rule is also required for leveraged buyouts (LBOs), where debt ratios are set high at first but paid down if business plans pan out (see Kaplan & Ruback 1995).

The APV rule is also useful because it discourages attempts to value projects with complex financing by adjustments to the discount rate. A simple example: Suppose a large project would have to be partly financed by a stock issue, which would incur transaction costs. Some are tempted to calculate a transaction-cost-adjusted cost of equity, which is both arbitrary and messy. Instead, just subtract the dollar value of incremental transaction costs from APV.

Another example: Long-term equipment leases can add value if the lessee pays a lower marginal tax rate than the lessor. Many commercial airplanes are lease financed. Some analysts are tempted to try to adjust WACC to account for the lower cost of debt implicit in the lease terms. APV says, First calculate NPV assuming normal financing; then add the NPV of substituting the lease for debt. The lease adds value only if its NPV is positive.

11 The tax term drops out because, with rebalancing, future interest tax shields are proportional to future value and just as risky. The exception is the first interest tax shield, which is fixed by borrowing at date zero. The adjustment for the safe first tax shield is usually trivial, however. Therefore, the risk of the portfolio of all the firm’s debt and equity securities does not depend on the debt ratio. 12 A similarly incorrect, tax-adjusted formula for levering or unlevering beta also circulates, for example, in Rosenbaum & Pearl (2013), p. 157. 13 APV can also be derived from the conditions for an optimal solution of a linear program; see Myers & Pogue (1974).

8 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

3.5. Determinants of Corporate Borrowing

Myers (1977) was motivated by my empirical observation—casual but true—that profitable growth firms usually operate at low debt ratios. Often they operate at negative debt ratios as net lenders. Yet they could easily swap debt for equity and save taxes. I offered an explanation, the debt overhang problem.

Debt overhang means that existing holders of risky debt capture some of the value added when the firm invests and expands, reducing the net NPV realized by equity investors. The gain to debt investors acts like a tax on future equity investment. The tax is potentially most burdensome on growth firms, whose current market value depends on the vigor of its future investment decisions. The growth opportunities are real options, which the firm can decide not to exercise.14 Pointing out the importance of real options was a simple but important contribution.

The debt overhang problem is easy to see from option-pricing formulas. Assume a firm has one risky debt claim outstanding. The debt matures with a bullet payment at date H. Then equity is a call option on the firm’s assets, with maturity H and an exercise price equal to the face value of debt plus accrued interest. The value of the call depends on firm value V, the exercise price EX, variance σ 2, maturity H, and the interest rate rf . All but rf are potentially under financial managers’ control. If managers act in shareholders’ interest, the following moral hazard problems arise:

� Debt overhang: With risky debt, δD/δV > 0 and δE/δV < 1. The marginal PV to equity investors from investment I is reduced to δV/δI(1 − δD/δV ), which may still be positive, but always less than marginal PV with safe debt or no debt at all. If marginal PV is negative, the firm is better off paying out cash to shareholders than investing. If cash is not available, the firm will not issue equity to invest.

� Risk shifting: δE/δσ 2 > 0. The firm is motivated to shift from safer to riskier assets. � Playing for time in financial distress: δE/δH > 0. When default threatens, shareholders

gain when the day of reckoning is postponed. Managers are tempted to cover up problems, keeping the equity call alive in the hope of recovery.

� Additional borrowing: The value of existing debt falls when more debt is added, holding assets constant. The value of existing debt also falls when new investment is 100% debt financed.

Notice that the debt overhang problem might be cancelled out by additional borrowing at the same debt ratio. Suppose all investment is partly debt financed and the debt-to-value ratio pre- and postinvestment is held constant. Then, shareholders benefit from all positive-NPV investments (σ 2 constant). This may be another rationale for a target debt ratio. On the other hand, the debt overhang problem becomes severe only when the risk of default is great and the market-value debt ratio is already dangerously high. These are exactly the circumstances in which lenders hope they had insisted on covenants that limit additional borrowing.

Myers (1977) explained in detail why the debt overhang problem cannot be solved by ex-ante contracting. A contract that requires managers always to do the right thing, exercising all maturing options with positive NPV, would be ideal. But contracts that depend on future NPV, which is subjective and nonverifiable, cannot work. Other hurdles include limited liability, which means that shareholders cannot be forced to contribute equity to finance investment. Restrictions on dividends or repurchases can help, however, if the firm has cash on hand at the optimal time for investment.

I concluded that growth firms might be better off avoiding the debt overhang problem than by trying to solve it. Avoidance means equity rather than debt. Any debt should be close enough

14 Miller & Modigliani (1961) distinguished the PV of growth from the value of assets in place but treated future investment as mandatory.

www.annualreviews.org • Finance, Theoretical and Applied 9

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

to default risk free.15 Thus, the trade-off theory got a helping hand in its attempts to explain why growth firms operate at low debt ratios.

3.6. Corporate Financing and Investment Decisions When Firms Have Information that Investors Do Not Have

Myers & Majluf (1984) was the first paper to investigate the implications of information asymmetry for corporate financing and investment.16 The investigation was launched in the late 1970s in discussions with Nicholas Majluf, then an MIT doctoral student in corporate strategy. Nicholas proposed that financial slack, defined as cash or unused debt capacity, was valuable because it allowed strategic investments without recourse to equity markets. “What’s wrong with equity financing?” I asked. “What market imperfection do you have in mind?” “Firms don’t want to issue undervalued stock,” he replied. I had not thought seriously about misvaluation, but suddenly realized that there might be an equilibrium in which optimistic managers would pass up a positive- NPV investment rather than issue shares to finance it. Nicholas and I worked out an example. The example is in Myers & Majluf (1984, pp. 192–94) exactly as it was written out on an MIT blackboard.

The setup of our example and model starts in the same place as Myers (1977). The firm has assets in place and a real option that requires a decision about exercise. Managers act in shareholders’ interest. But now managers have inside information about the value of assets in place and also about the NPV of the real option. The firm may have cash on hand (financial slack) but not enough to finance the investment. Investment requires fresh equity from shareholders. (The initial model did not allow debt issues.) Shareholders value the firm rationally, based on their prior information and what they learn from the firm’s decision whether to issue and invest.

A pooling equilibrium may result, in which all types of firms invest or all don’t. The separating equilibrium is more interesting. Managers with optimistic private information about the values of assets in place may sacrifice a positive-NPV investment rather than issue undervalued shares. Other managers with bad private news about the values of assets in place may issue and invest. Firms with more financial slack are more likely to invest because the required equity issue is less. Thus, Nicholas’s suggestion about the value of financial slack was confirmed.

A firm is more likely to issue in a separating equilibrium when its managers are more optimistic than are investors about the NPV of the new investment and less optimistic about the value of assets in place.17 Stock price falls at the issue announcement. The prediction that stock prices will fall at announcement holds for any joint probability distribution of the manager’s inside information about the value of assets in place and the investment’s NPV. The predicted stock-price fall was confirmed by event studies (see, e.g., Asquith & Mullins 1986).18

The Myers-Majluf model is widely cited, but the model has three attributes that are often missed. First, it is critical that managers have two tranches of inside information and that outside

15 Also the firm may be better off with short-term debt, that is, with debt that matures before the exercise date of the firm’s real option to invest. But the debt overhang problem can be worse with short- rather than with long-term debt if the debt is still outstanding on the exercise date (see Diamond & He 2014). 16 Our names are out of alphabetical order because I took the paper well beyond the basic model of equity issues and investment set out in Majluf (1978). 17 Notice that I did not say “pessimistic.” All managers are optimistic, at least in public. It is just a matter of degree. Business Week used to publish surveys asking CEOs whether their stock price was too low, too high, or just about right. The results never varied. As I recall, approximately 75% said “too low” and 20% said “just about right.” The remaining 5% didn’t understand the question. 18 Nicholas and I derived the prediction before seeing working-paper versions of these event studies.

10 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

investors see two sources of uncertainty. If the value of assets in place is known, then the firm always issues and invests when NPV is positive, even if investors are pessimistic and the manager views the stock price as too low. It is better to capture some fraction of NPV than to let the investment go and realize nothing. If the firm has only assets in place of uncertain value and the NPV of investment is small, equity issues may be ruled out because the manager’s primary motive for issuing would be to get investors to pay too much for the overvalued assets. Investors, of course, refuse to play in such a game.

The impediment to issuing equity to invest comes mostly from investors’ uncertainty about the value of assets in place. Thus, growth firms with large, positive-NPV investment options will have an easier time issuing equity than mature firms valued mostly on assets in place.

Second, managers in the Myers-Majluf model are assumed to act in the interests of the existing, old shareholders. They refuse to issue undervalued shares if too much value would be transferred to new shareholders. This objective seems reasonable, particularly if managers also own shares and do not participate pro rata in a new issue. But the objective is not derived from any first principles. Value would be maximized for all, old and new shareholders alike, if the managers could commit ex ante to make all positive-NPV investments. Old shareholders would endorse such a commitment if it could be implemented. (The commitment cannot be enforced by contract, for the same reasons that contracts cannot solve the debt overhang problem.) On the other hand, the Myers-Majluf objective function seems realistic. For example, share price would not fall at the announcement of a stock issue if managers really acted in the old and new shareholders’ combined interest.

Third, the Myers-Majluf model assumes that old shareholders do not buy 100% of a new issue, either because that is their optimal portfolio choice or because some sit passively on the sidelines as the issue happens. If they did buy 100%, there would be no barrier to equity issues because managers would not have to worry about a transfer of value to new shareholders.

Why does financial slack help? Suppose that the firm retains sufficient cash flow to finance investment without issuing equity. Then the old shareholders provide 100% financing, as if they had purchased 100% of a stock issue to finance the same investment. There is no possibility of transfer of value to new stockholders.

This argument is fine so long as old stockholders can be assumed passive. If they are passive, then internal financing is a way of enforcing 100% participation. But suppose that old shareholders are not passive, but instead rebalance their portfolios in a scrupulously rational fashion when the firm invests. They end up with the same portfolio regardless of financing. Absent other frictions or imperfections, one gets to an MM equilibrium. The firm may still let positive-NPV investment options expire, not because of financing, but because shareholders’ optimal portfolio rebalancing opens the door to new shareholders.

The assumption of passive stockholders may be strong empirically even if weak logically. Later research building on Myers-Majluf has, as far as I know, accepted the assumption unanimously, though usually implicitly.

3.7. The Capital Structure Puzzle and the Pecking Order

Myers & Majluf (1984) extended the model of investment and equity financing to allow for debt. We showed that debt dominates equity when managers have inside information about the value of growth options and assets in place.19 Investors who understand their information disadvantage

19 The pecking order can be reversed for some joint probability distributions; see, for example, Fulghieri, Garcia & Hackbarth (2014).

www.annualreviews.org • Finance, Theoretical and Applied 11

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

will buy only the safest security that the firm can issue, the security where information asymmetry matters least.

Myers-Majluf therefore suggested a pecking-order theory of financing, which I fleshed out in “The Capital Structure Puzzle” (Myers 1984b), my presidential address to the American Finance Association. The theory proceeds as follows:

1. Capital investment responds to prospective NPVs. Any feedback of financing to investment is second order.

2. Payouts (dividends and repurchases) are smoothed and respond only gradually to changes in profitability and investment opportunities.

3. Firms turn first to internal financing to finance investment and payout. 4. If external financing is required, debt is issued. Equity issues are a last resort. 5. If there are no equity issues—and they should be rare for mature firms—then the observed

debt ratio depends on the firm’s cumulative requirements for external finance.

Step 1 is an assumption, which I justify by noting that capital investment is much more important to firm value than financing. Step 2 is at least roughly true. Steps 3 and 4 follow from Myers-Majluf. Step 5 follows from Steps 1–4.

The pecking-order theory has some immediate attractions. It explains manager’s preferences in mature corporations for internal financing and their aversion to stock issues (see Donaldson 1961, pp. 67–70). It explains the within-industry dispersion in debt ratios and the inverse correlation of profitability and debt ratios. More profitable firms have lower debt ratios simply because they need less external financing. It explains, based on Myers-Majluf, why mature public corporations rarely sell seasoned equity offerings.20

The pecking-order theory ignores taxes and costs of financial distress, the main elements of the trade-off theory. I agree that these elements can be important, but left them out in the pecking- order derivation in order to present a simple theory that is easily testable. I wanted to present a simple competitor to complex trade-off theories. “People [felt] comfortable with the static trade- off story because it sounds plausible and . . . rationalizes ‘moderate’ debt ratios. Well, the story may be moderate and plausible, but that does not make it right” (Myers 1984b, pp. 588–89). I doubted whether the theory had really been tested.

What would a simple, testable version of the trade-off theory look like? It would predict that mature firms (with relatively few valuable real options) would stay close to their target debt ratios, which implies prompt adjustment of actual debt ratios toward targets. The theory would be specific about the business characteristics that determine the targets.

I do not know whether that simple, testable version was ever taken seriously. If so, its lure was fading when I wrote “The Capital Structure Puzzle.” Now we have a relaxed version of the trade-off theory, which (depending on the interests of the researcher) allows for transaction costs of issuing securities, other costs of adjusting capital structure, nondebt tax shields, risk-averse managers to explain low debt ratios, and various agency arguments. The relaxed version no longer says that firms stay tight to their targets, nor can it derive specific predictions on what the targets depend. Empirical research has identified some characteristics that correlate with debt ratios and could correlate with targets. The characteristics seem reasonable. However, it is too easy to rationalize the correlations as consistent with the relaxed theory. The two most dangerous words in empirical corporate finance are “consistent with.”

20 Fama & French (2005) show that such firms do issue equity in other ways, for example, in acquisitions and by exercise of stock options.

12 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

The relaxed version of the trade-off theory does not make specific predictions that could be rejected by data. (I exaggerate for emphasis.) It is a coatrack for researchers, who can hang up whatever empirical results they can find and rationalize. No one seems concerned if the coats on the rack clash. This coatrack version of the trade-off theory was clearly emerging in 1984. There was no point adding information asymmetry to the coatrack and giving the relaxed trade-off theory one more dimension of flexibility.

Recall that the trade-off theory was first set out in Robichek & Myers (1966b). We believed that both taxes and costs of financial distress can be important. Transaction costs and other elements of the relaxed trade-off theory can also be important. Also, we know certain empirical regularities (for an interesting survey, see Graham & Leary 2011). The capital structure puzzle arises when we try to find a compact theory to explain the regularities and predict new ones.

Wikipedia says that I was a “particularly fierce critic” of the trade-off theory in Myers (1984b). If so, it was to emphasize the points that I am repeating here.

3.8. Testing the Pecking-Order and Trade-Off Theories

What follows falls far short of a complete review of empirical work on capital structure theories. It is a historical view of my own research, with scattered comments on papers by others that struck me as interesting. I have not fully kept up with recent empirical research, so I am probably summarizing my views circa 2005.

Myers & Majluf (1984) says that it’s mostly managers’ inside information about assets in place that blocks equity issues and investment. This point is widely misunderstood in empirical papers, for example, in Frank & Goyal (2003, p. 219):

[The pecking-order] theory should perform best among small, high-growth firms with large informa- tion asymmetries. Contrary to this hypothesis, small high-growth firms do not behave according to the pecking-order theory. In fact, the pecking order works best in samples of large firms that continuously existed during the 1970s and 1980s. Large firms with long uninterrupted trading records are not usually considered to be the firms that suffer the most acute adverse selection problems.

Frank and Goyal’s hypothesis does not follow from Myers & Majluf (1984). I do not know where they got it from. The results they summarized in the quote are consistent with (dangerous words!) Myers-Majluf and the pecking order.

A horse race between the simple pecking-order theory and a simple trade-off theory can be informative. The race should be run on a sample of mature firms, not on a sample including firms trading at high price-earnings multiples because of valuable growth opportunities. The two theories make the same vague predictions for growth firms, so it becomes impossible to distinguish them.

For mature firms, where assets in place are the most important source of value, the predictions of the pecking-order and simple trade-off theories diverge clearly.21 The trade-off theory says that a firm’s change in debt will bring its debt ratio back toward a target. The pecking-order theory says that the change in debt will equal the firm’s net requirement for external finance.

Shyam-Sunder & Myers (1999) tested these predictions on an (admittedly small) sample of mature firms. Both theories seemed to work when fitted separately to the data. But the real insight

21 The tricky part may be picking the sample of mature firms in a way that does not give one horse a lead in the race. Researchers will have to find instruments for “maturity,” probably following Lemmon & Zender (2010).

www.annualreviews.org • Finance, Theoretical and Applied 13

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

of Shyam-Sunder & Myers (1999) was their investigation of the power of the statistical tests. We generated for each firm two simulated time series of changes in debt. The first assumed that the change in debt was always exactly equal to the requirement for external finance (the pecking- order theory). The second assumed a simple target adjustment model (the trade-off theory) with reasonable parameters. In each case, only the change in debt was simulated. Otherwise, we used each firm’s actual operating cash flow, capital investment, change in working capital, and payout. Then we ran the two theories on the simulated data. Of course, each theory worked perfectly on the data simulated according to that theory. But could each theory be rejected if it was wrong? The results were striking. The pecking-order theory failed totally when the simulated changes in debt followed a target adjustment process. But the trade-off specification still seemed to work on pecking-order data. The reason is that firm’s requirements for external finance are to some extent cyclical, for example, over the business cycle, or because periods of heavy capital investment alter- nate with periods of lighter investment and positive free cash flow.22 Thus, debt ratios generated by the simplest pecking-order model were mean reverting. The target-adjustment specification of the trade-off theory seemed to work when the target debt ratio was calculated as an in-sample average.

The standard time-series specification of the trade-off theory had no power to reject the theory if it was false and the pecking order was true. The pecking-order specification did have the power to reject—and the pecking-order theory was not rejected using actual data.

I do not understand why later papers have not used the Shyam-Sunder & Myers (1999) method- ology to check the power of their specifications. Sometimes I wonder whether authors have been afraid of what they might uncover.

I do not believe that the simple pecking-order theory is a complete explanation of debt policy, even for mature firms in normal times. It surely falls short in explaining financing across all public companies, which include growth, mature, and declining firms; small and large companies; and companies that have ready access to public debt markets versus companies dependent on bank financing. But clearer answers may be in reach from investigations of financing by mature firms that are free to choose between debt and equity. Lemmon & Zender (2010) conducted such a test, in which the pecking order works well.

Recall the joke about the drunk looking for a dollar bill under the lamppost—not because the drunk thought the dollar was really there, but because he or she would never find it in the dark. The joke does not apply here. There is money under the lamppost; we just have to figure out what kind of money it is. It is useful to shine the light on mature firms’ debt policies because these firm’s policies should be clear and easy to interpret. We can shine the light on growth firms, trying to distinguish competing theories of capital structure, and still not understand what’s going on.

3.9. Were Modigliani and Miller Right After All?

MM’s proposition 1 was a statement about the market value of the firm, not about the firm’s debt ratios and financing tactics. The trade-off theory accepts MM’s proposition and derives predictions about financing. But the proposition itself is not tested.

The magnitude of the contribution of interest tax shields to the value of the firm has likewise not been tested; it is just assumed to be big enough to matter. Fama & French (1998) tried to measure the contribution of interest tax shields to firm value, but they failed “to measure how (or

22 Recall that the test by Shyam-Sunder & Myers (1999) used their sample firms’ actual requirements for external finance. Only debt policies were simulated.

14 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

whether) the tax effects of financing decisions affect firm value. The relations between financing decisions and value we observe are unidentified mixes of tax effects and the factors that affect profitability” (Fama & French 1998, p. 821).23

What if MM’s proposition 1 holds in practice without any adjustment for interest tax shields? That is, what if the mix of debt and equity does not affect value, even when interest is tax deductible? There would still be plenty of room for financing tactics. One firm could find it convenient to target a single-A debt rating, and another to operate at a 35% book debt ratio. An especially profitable firm could pay down its debt and build up a cash mountain. Mature firms could follow the pecking order while growth firms issued equity. One firm could make large public debt issues and repurchase shares with the proceeds. Another could rely on bank debt and find it difficult to reduce its debt ratio in the short run. None of these tactics would have a first-order effect on firm values.

Suppose that tactics can add or subtract 2 or 3% of the value of public companies. If MM’s proposition 1 is correct within a band of error of ± 0.02 or ± 0.03, then their proposition is an empirical success and MM are right after all.

Corporate financing may not be driven by any deep theory, but rather by tactics. If so, we can relax and work to understand the tactics. But we must not be surprised when tactics change, as they will in response to changes in markets, financial technology, regulation, and the nature of firms’ operations.

4. ACT 2: APPLICATIONS

I came to research in financial economics by way of an MBA program, and I teach in a business school. That may explain my conviction that research in finance should help managers make better decisions. I understand that actual decision makers must cope with poor information and with circumstances and constraints that do not fit in general-purpose valuation models. I also understand that research in financial economics has shifted from normative to empirical research and to pure theory. Nevertheless, I am convinced that both empiricists and theorists in finance are better at their jobs when they take a serious interest in actual decision making.

I have already mentioned two papers with direct implications for decision making: Robichek & Myers (1966a) and Myers (1974). Now I turn to later applied publications, starting with Principles of Corporate Finance. I cover applied research on real options, including their implications for corporate strategy and R&D, and the valuation of options when taxes and financing are important. I comment on regulatory finance and research in insurance. Finally, I touch on recent work on the role of risk capital in financial firms.

4.1. Principles of Corporate Finance

The first edition of Dick Brealey’s and my Principles of Corporate Finance was published in 1981 (Brealey & Myers 1981; now Brealey, Myers and Allen 2014). I met Dick shortly after arrival at MIT in the 1960s. He was working as a portfolio manager in Boston and regularly attended MIT finance seminars (so did Fischer Black, who was a consultant at the time). Dick then joined the fledgling London Business School (LBS). In 1975, I spent six months at LBS, where Dick and

23 Miller & Modigliani (1966) did include the value of interest tax shields in their tests of the cost of capital for the electric utility industry in the 1950s. The tax shields did seem to add value in that setting, although utility regulation passed that value on to ratepayers.

www.annualreviews.org • Finance, Theoretical and Applied 15

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Harold Rose—the elders and intellectual leaders of LBS Finance—had assembled an excellent, then-young finance faculty. “Determinants of Corporate Borrowing” (Myers 1977) was mostly written at LBS, as were my papers on lease financing (Myers, Dill & Bautista 1976) and on capital budgeting and the CAPM (Myers & Turnbull 1977).

Dick had completed several chapters of a corporate-finance text in the mid-1970s. He asked me to join the project after a previous coauthor could not match Dick’s pace. Now I was the one who had to keep up. The drill never changes: We agree that he will work on chapters x and y and I will work on i and j. Dick is back with nice drafts of x and y before I have organized my thoughts for i and j. I have spent 35 years playing the tortoise to Dick’s hare. Another hare, Franklin Allen, joined the team for the ninth and subsequent editions. The same tortoise-hare relationship applies to my undergraduate text Fundamentals of Corporate Finance, with Dick and Alan Marcus.

Principles was the first corporate-finance text to apply modern finance theory in a consistent and practical way. By the late 1970s, modern finance theory included the objective of maximizing market value, portfolio theory, the CAPM, the efficient-market hypothesis, and the principles of option pricing. The theory included MM’s leverage and dividend-policy proofs as starting points for capital structure and payout policy. At a more applied level, it included DCF and APV valuation methods and the use of WACC as a discount rate.

I remember my frustration with the popular texts of the 1970s, which mentioned modern finance theory but did not exploit it. I think the authors of these texts shied away from the theory because they thought it would make finance more difficult and less practical. Dick and I believed on the contrary that good theory makes common sense and simplifies practice. (As Bob Merton says, “If the principle is right, the practice follows.”24) We must have been right, because Principles is still a best seller worldwide. I believe the book, along with later competitors using much the same template, has improved practice materially.

We are told that Principles is clear, engaging, and sometimes even entertaining, compared with many other textbooks. I trace this to the blend of Dick’s and my writing styles. I try to be compact and direct.25 His prose is more conversational and sophisticated, in an Oxbridge sort of way. The combination has hybrid vigor.

Writing the first edition of Principles was a risky investment. We took no guidance from potential publishers about what we ought to cover and how. The first reviews of the manuscript were mixed. One reviewer offered faint praise for several chapters but went out of the way to warn that American professors and students would be put off by Dick’s “English humor.” All the jokes cited by the reviewer were mine.

4.2. Real Options

I was the first to identify real options and to stress their importance for financing. But real options are everywhere in corporate finance, so applications multiplied rapidly. Saman Majd and I worked out two of the applications.

Myers & Majd (1990) analyzed the abandonment decision. For example, the decision to shrink or shut an elderly plant is a put. The exercise price is the plant’s scrap value or its value in its next- best use. (If the value from exercise is uncertain, the put option becomes an exchange option.)

24 Bob does not recall this wise saying, but I give him credit for it anyway. 25 I did not learn good writing in school or college. I learned it as a cub reporter in the Albany, NY, Times Union in 1959 and 1960. Barney Fowler, the City Editor, edited my copy ruthlessly, crossing out all unnecessary words and wandering sentences. It took a howitzer to get him to accept anything in the passive voice. I had to learn quickly or sit alone in a corner of the pressroom with nothing to do. Therefore, I learned quickly.

16 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

The implications for a company with aging assets are clear, but the implications for the valuation of new assets are potentially just as important. A typical DCF analysis of a capital investment (CAPEX) project forecasts cash flows to an assumed economic life and discounts to NPV—as if the financial analyst knew for sure how long the project will last. In fact, the project life may be much longer or shorter than expected, depending on performance. It is better to do DCF with a very long life, with declining expected cash flows as obsolescence accumulates, and then add the value of the put to close down early. Notice: another application of APV.

Majd & Myers (1987) investigated the US government’s call option to tax income. The Internal Revenue Service (IRS) is happy to tax positive income but limits deduction of losses. Losses can be carried back for two years. However, when carrybacks are exhausted, losses must be carried forward and are tax deductible only if and when cumulative taxable income returns to positive territory. We compared the PV of income taxes on a risky project in two settings. First, we assumed that the project is stand-alone, with no other assets and no previous income that could allow tax carrybacks. The project could carry tax losses forward, however. Second, we valued taxes on the same project held by a profitable company with plenty of other taxable income to absorb project losses. In the second case, taxation of income and losses from the project are symmetric. In the former, the IRS taxes mainly the upside. We found dramatic differences in the PV of taxes in the two settings. The stand-alone project was at a significant after-tax disadvantage.

Majd & Myers (1987) was a challenging paper technically,26 but its results still strike me as important. Yet it has gained hardly any traction in the world of taxation and public finance. Perhaps I should have marketed policy implications more aggressively.

4.3. Real Options and Strategy

“Short-termism” is defined as excessive preoccupation with quick financial results at the expense of long-term value. Short-termism is often paired with “tyranny of.” It is often blamed on an impatient stock market and on top managers more motivated to increasing a current year’s earnings per share than to add long-run value. Some say short-termism is built into modern corporate finance.

The stock market is not generally impatient and unwilling to value the long run. Those who complain about short-termism on Monday may on Tuesday complain about crazy-high price- earnings (P/E) ratios for hot growth stocks. But a high P/E ratio reveals investors’ beliefs that a company can keep investing at positive NPV into the distant future.

Myers (1984a) explained why corporate-finance theory, especially DCF valuation models, can nevertheless aid and abet short-termism. First, DCF is often misapplied. Suppose managers are encouraged to rank projects on percentage rate of return [internal rate of return (IRR)]. The place to look for high IRRs, not necessarily high NPVs, is projects with small investments and quick payoffs. The search for high IRRs gets worse if top management adds a fudge factor to the true cost of capital in a naive attempt to offset optimistic cash-flow forecasts.

Brealey, Myers, and Allen’s Second Law explains why such fudge factors never work. The law states that the proportion of projects having positive NPVs at the corporate hurdle rate is indepen- dent of the hurdle rate. We borrowed the law from Al Olenzak, who oversaw capital-investment analysis at Sunoco in the 1980s. Al observed that approximately 80% of project proposals had positive NPVs; the remaining 20% were motivated by safety, environmental, and other such re- quirements. One year, top management boosted the corporate hurdle rate by several percentage

26 The IRS’s tax call is complex and path dependent because of carry forwards. But tax rules are not discretionary, so we were able to value taxes using Monte Carlo simulation in a risk-neutral setting.

www.annualreviews.org • Finance, Theoretical and Applied 17

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

points in an attempt to impose financial discipline on the capital-investment process. Cash-flow forecasts adapted. The proportion of proposed projects with positive NPVs remained steady at 80%.27

Second, the internal capital market in many large firms is not a market but an annual budgeting exercise (thus the term “capital budgeting”). Budgets may be set tight in an attempt to impose financial discipline. Once set, a plant’s or division’s budget cannot be easily changed. A plant or division manager may therefore act as if facing a capital constraint and try to cope by investing in incremental, inexpensive, quick-payback projects rather than in projects designed to add value in the long run. I have seen such capital-constrained behavior firsthand in corporations that could raise $1 billion in cash overnight.

Third, the payoffs to strategic investment projects come from future exercise of real options. For example, the first model of a new aircraft from Boeing or Airbus is typically negative NPV, but if the new airplane succeeds, further investment options open up. Follow-on models can be stretched, modified for extended range or for air freight, re-engined, etc. These investments are options, not obligations, and therefore cannot be valued by ordinary DCF, which assumes that probability distributions of future cash flows are symmetric.

Therefore, DCF valuations of strategic investments are often much too low. The APV of a strategic investment should be calculated as the sum of the NPV of the initial investment plus the PV of real call options created by the initial investment.

The value of strategic options is easy to illustrate but difficult to quantify confidently. But it helps in practice to see them as options and to realize that the options are more valuable when uncertainty increases.

Short-termism can exist. It can come from misuse of DCF valuation tools, rigid internal capital markets, and failure to recognize the value of strategic options. But short-termism is not built into modern finance theory if the theory is applied correctly.

4.4. Pharmaceutical R&D

In the mid-1990s, MIT launched the Program on the Pharmaceutical Industry. At roughly the same time, Judy Lewent, then CFO of Merck, organized a series of workshops on the economics of her industry. I participated in both programs and was therefore forced to figure out how to value pharmaceutical R&D.

In those days, “big pharma” companies did most of their R&D in house,28 taking candidate drugs from discovery to preclinical trials and if successful through three stages of clinical trials: Phase 1 for safety, Phase 2 on small samples to show efficacy, and Phase 3 on large samples to prove efficacy. The final hurdle was application for approval from the Food and Drug Administration, required for commercial launch. Cumulative failure rates were high, so a cohort of hundreds of candidates yielded only a handful of drugs that actually reached their markets.

R&D for a specific drug candidate was an investment in a sequence of scientific lotteries. The prize in each lottery was a real call option to continue investment. The underlying for each call was the next lottery on the next call. It sounds complicated, but there is a simple approximation.

Think of a cohort of 100 drugs at the start of preclinical testing. Each drug has NPV of at least zero. Otherwise it would not join the cohort. Suppose that 20 drugs survive preclinical testing and

27 I do not remember the exact percentage, but 80% seems about right. 28 Now much R&D is farmed out to smaller companies. Big pharma companies acquire smaller companies’ successes by license agreements or by purchasing smaller companies outright.

18 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

that 10 survive Phase 1 trials. Most or all of the 10 that survive will have large positive NPVs, be- cause two gauntlets of uncertainty were run successfully. Conditioned on success through Phase 1, the real option is far in the money and almost always worth exercising. Thus, option valuation simplifies, because the value of a far-in-the-money call approaches the PV of the underlying minus the PV of the exercise price. The PV of the underlying is the PV of the net after-tax cash flows from sale of the drug over its commercial life cycle multiplied by the probability that the drug will pass all subsequent trials, receive FDA approval, and reach the market. The PV of the exercise price equals the PV of subsequent R&D outlays multiplied by the probabilities that the drug will pass each test and that the outlays after each test will be made.

Each candidate can be valued by solving back in a decision tree. At each stage of the tree,

APV = PV(net cash inflows from commercial sale, probability weighted) − PV(remaining cost of future R&D, probability weighted).

The second term of the APV creates R&D leverage because the obligation to pay for R&D is a fixed, or at least low-risk, obligation. The APV equation explains why investors should demand higher expected rates of return from early-stage than from later-stage R&D: not because the probability of failure is higher in early stages—the scientific lotteries create diversifiable risks— but because R&D leverage is greater.

Myers & Shyam-Sunder (1996) and Myers & Howe (1997) introduced the idea of R&D leverage and the APV valuation equation for pharmaceutical R&D. It is an interesting and practical real- options application. The application requires estimates of timing, costs, and success probabilities in each phase and a forecast of expected cash flows if and when the drug reaches the market. But all drugs go through the same R&D stages, and managers can refer to historical data on average outcomes.29 Big pharma companies could come up with reasonable estimates of the inputs for the Myers-Howe method. They were not flying blind.

Given the inputs, a single drug can be valued in a simple spreadsheet. Understanding the outcomes from a drug development program, where hundreds of drugs of different cohorts are working their way toward commercial launch, requires a Monte Carlo simulation. Myers & Howe (1997) described such a simulation model, which was later used in Healy, Howe & Myers (2002) as a test bed for analyzing alternative accounting methods for R&D.

The pharmaceutical industry is also a poster child for biased accounting. Year after year in the 1990s, Fortune identified pharmaceuticals as the most profitable US industry, with nary a word hinting that the industry’s returns on assets (ROAs) or capital (ROCs) might be upward-biased measures of the true economic rates of return. The biggest problem was that billions of R&D outlays were expensed and thus written off immediately. As a result, the denominators of ROAs and ROCs were grossly understated. Warnings by Solomon & Laya (1967) and Fisher & McGowan (1983) and in chapter 12 of Brealey-Myers were ignored.

Paul Healy and I proposed to set up and run an independent accounting authority that would restate big pharma 10-Ks by capitalizing and amortizing R&D and correcting other sources of bias. But the big pharma companies would not touch our proposal.30 Paul and I joked that the companies downplayed their high ROAs in Washington, D.C., but did not want to give up bragging about them on Wall Street.

29 For example, see the website of the Tufts Center for Drug Development, http://csdd.tufts.edu. 30 The one exception was Merck’s CFO Judy Lewent, who was willing at least to take preliminary steps for an independent pharmaceutical accounting authority.

www.annualreviews.org • Finance, Theoretical and Applied 19

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

4.5. Real Options, Taxes, and Leverage

Myers & Read (2014) explained how to calculate the after-tax APV of a real option. Finally, 40 years after Myers (1974)! I spend some extra space on this topic because it has implications for tests of the trade-off theory.

Real options are almost always valued in a risk-neutral setting in which certainty-equivalent cash flows are discounted at the risk-free interest rate. Everyone uses a pretax risk-free rate. Yet ordinary DCF discounts safe cash flows—contractual payments, for example—at the after-tax risk-free rate. (For an explanation and examples, see Brealey, Myers & Allen 2013, pp. 508–11).

Consider a real call option to expand. Exercise means investment in an underlying asset. The value of that asset is an APV, which includes the PV(interest tax shields) from debt that will be supported by the asset if the option is exercised. (The APV can be calculated in one step by discounting cash flows at an after-tax WACC if the assumptions behind the WACC calculation are satisfied.) But the real option also has debt capacity prior to exercise. The implications of that debt capacity have to be understood before the option can be valued. (Debt capacity is not the maximum possible debt, but the target amount of debt that the firm chooses to borrow against the option or the asset.)

Calculating the debt capacity of a real growth option is easy once you see what is happening. Suppose the firm has a target debt ratio of λ against the APV of assets in place. Thus, the debt capacity of the underlying asset is λAPV. The firm also has an option to expand worth APV(call). Split the call into a replicating portfolio:31

APV(call) = δAPV − DC , where δ is the option delta and DC is the debt that would be required for replication. δAPV tracks the underlying and should have the same proportional debt capacity over the next step in time. DC is equivalent to off-balance-sheet debt, which should displace explicit debt dollar for dollar. The debt capacity of the call is therefore λδAPV − DC .

The implicit interest on DC is not tax deductible, so interest tax shields are lost when option leverage displaces explicit debt. Thus, the implicit debt should be calculated as APV(DC ) to account for the lost tax shields. The adjustment can be done algebraically in the usual risk-neutral setup, where the risk-free discount rate is pretax. But the adjustment is automatic if valuation is done in an after-tax risk-neutral setting, where the expected rate of return on the underlying asset is assumed equal to the after-tax risk-free rate and that rate is used for discounting.

Myers & Read (2014) has both normative and empirical implications:

� Normative: Real options should be valued in an after-tax risk-neutral setting. We showed that the valuation errors from valuing real options in a pretax risk-free setup can be material.

� Empirical: The debt capacity of a real call option is usually negative and often greater than the value of the call. Suppose a firm’s target debt ratio for assets in place is λ. The firm has real call options with negative debt capacities. If it takes target debt ratios seriously, it will reduce explicit borrowing and operate at an observed debt ratio less than λ. It is easy to construct examples of firms with significant—but not extreme—growth opportunities in which the observed debt ratio is negative and the firm appears to be a net lender.

31 Real-options values can always be expressed as a replicating portfolio, even though actual replication by dynamic hedging is impossible. Real-options valuations require an asset pricing model for the underlying asset and its option. If the model works for the underlying, for example, in DCF format, it works for the option (see Brealey, Myers & Allen 2013, p. 577).

20 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

We also showed that the debt capacity of a real abandonment (put) option is always positive and always greater than the value of the put itself. A mature firm with significant abandonment options will, if it takes its target debt ratios seriously, operate at a debt ratio higher than the target λ for assets in place. An LBO motivated by abandonment put options should be financed by a higher-than-normal debt ratio. One does not have to invoke the disciplinary role of LBO debt to get this result. It follows directly from the trade-off theory once put options are recognized.

Myers & Read (2014) rationalized low or negative debt ratios for growth companies, but on balance, the paper delivers bad news for tests of the trade-off theory using samples of firms with valuable real call or put options. If a firm follows the trade-off theory, then it must account for the implicit debt in real options. That debt is large relative to the value of the options, and it fluctuates as the values of assets in place and corresponding options fluctuate. Therefore, observed debt ratios should not follow any simple target-adjustment specification.

Myers & Read (2014) reinforced an argument made above: Test the trade-off versus the pecking-order theories on a sample of mature firms for which the values of real options are rela- tively small. This is exactly what Shyam-Sunder & Myers (1999) did. Do not download thousands of firms from Compustat and try to fit one specification to them all.

4.6. Finance and Regulation

My interest in regulatory finance continues to this day. But here I emphasize testimony and academic papers from the late 1960s and the 1970s, when I helped to introduce modern finance to rate of return regulation.

In 1968, my colleague Dan Holland recommended me to Haskell Wald and David Schwartz, the chief economists at the Federal Power Commission. They sought a fresh approach to setting the allowed rate of return for natural-gas pipelines. So in 1969 I submitted expert testimony on the cost of equity for Texas Eastern Transmission Co. In 1971, at age 31, I was the Federal Communication Commission’s (FCC’s) chief economic expert in the AT&T interstate rate case. (Recall that AT&T then had a near monopoly in long-distance telephone service. There was a massive amount of money at stake.) My recommended 10.5% equity return was accepted exactly by the FCC, although they waited until the Friday after Thanksgiving to announce it. I came to understand from hints and secondary sources that the FCC staff thought my 10.5% was too high and AT&T thought it was too low but both thought it was acceptable.

The US Supreme Court requires that “[t]he return to the equity holder should be commensu- rate with returns on investments in other enterprises having corresponding risks” (Federal Power Commission et al. v. Hope Natural Gas Co. 1944, p. 603). My Texas Eastern and AT&T reports interpreted this requirement as follows. I defined the return to the equity holder as the equilibrium expected rate of return in financial markets for investments with the same risk as the regulated firm’s common stock. This matches the definition of the cost of equity for an unregulated company. I estimated the cost of equity by the DCF method, in which the cost of equity is the discount rate that equates the PV of future cash dividends to the current stock price. I did not, however, simply use the “dividend yield plus growth rate” formula.32 Instead, I built DCF models with varying growth rates and rates of return on investment that were assumed to drift back to the cost of capital.

32 The formula is rE = DIV1 /P + g, where DIV1 is next period’s dividend, P the current price, and g a constant and perpetual future growth rate of dividends. The formula first appeared in Williams (1938) and was rediscovered by Gordon & Shapiro (1956).

www.annualreviews.org • Finance, Theoretical and Applied 21

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Myers (1972) explained why this cost of equity is the correct allowed rate of return on a traditional regulated rate base. This approach to rate of return regulation is universally accepted today, but was not in 1969 or 1971.33 I believe I was the first to apply it in major rate cases.

My next regulatory foray was for the Communications Satellite Corporation (Comsat). Comsat had been launched in 1963 with great fanfare. The company actually delivered what investors hoped for. Satellites were launched, with no significant failures, and they worked with boring regularity. Thus, in the early 1970s, Comsat found itself in a rate case before the FCC, looking like another dull, regulated telecom. I ran across Gene Brigham, Comsat’s cost of capital expert, and pointed out that Comsat’s beta had been substantially above 1.0 ever since the company’s stock started trading. Thus, the CAPM called for an allowed rate of return much higher than the FCC was used to. I was recruited to make that argument. My testimony explained the CAPM and argued that Comsat deserved an extra-generous allowed return.

My work for Comsat was, as far as I know, the first full-scale application of the CAPM in rate of return regulation. I starred in a Harvard Business School case about the Comsat proceeding (Mullins 1976). Unfortunately, I did not convince the FCC, which allowed only an ordinary rate of return. Hindsight provides a clearer view of what happened. Comsat’s beta was high in the company’s first decade because of investment leverage, which operated in the same way as R&D leverage. Put another way, Comsat’s early assets were growth options, which had higher betas than the assets finally in place. So in one sense, the FCC was right: On the one hand, the risk and cost of capital for Comsat’s assets in place were probably not unusually high, once the technology was proved and the satellites were operating routinely. On the other hand, Comsat deserved compensation for the risks it ran and avoided. If several satellite launches had failed, the FCC would not have bailed the company out.

Traditional rate-base regulation is not designed to cope with significant risks. Myers (1973a,b) proved that allowing a rate of return on investment equal to the regulated firm’s cost of capi- tal eliminates monopoly profits but cannot replicate competitive pricing and output decisions if demand is uncertain. Compensation for nonroutine risks is elusive, as Comsat found. The risks that it ran disappeared with hindsight and were not compensated. Kolbe, Tye & Myers (1993) described example after example of risks that were not provided for ex ante and not compensated by regulators ex post.

Next up was the Williams case before the Federal Energy Regulatory Commission (FERC). The FERC had inherited regulation of oil pipelines from the Interstate Commerce Commission (ICC) and had to decide whether to continue the ICC’s unique and complex procedures, to introduce utility-style rate bases and costs of equity, or to do something else entirely.

The biggest issue was inflation, which was raging in 1978, when I prepared an expert report. High inflation makes a mess of traditional utility regulation, which allows a nominal rate of return on an historical-cost rate base. Suppose, for example, that the real cost of capital is 6%, but inflation is running at 10% per year. In that case, the nominal cost of equity is about 16%. An asset costing $100 million is added to the rate base and depreciated straight line over 25 years, or by 4% per year. Then the annual capital charge is (0.16 + 0.04) × 100 = $20 million per year, constant over the 25-year life. But the real capital charge starts out double its level in a world of zero inflation

33 The prior comparable earnings method used the accounting returns on equity (ROE) for samples of other companies. But if the other companies are also regulated, then regulators using the method end up mimicking the returns other regulators allow. If the other companies are unregulated, then (a) it is difficult to match risk and (b) the unregulated companies’ ROEs are likely biased estimates of true economic rates of return. Note that the biases should not arise in regulated ROEs. If a regulated company earns its cost of capital on its rate base, then the market value of its equity equals the rate base and there is no bias (see Myers 1972, note 38, p. 97).

22 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

(20% versus 10%). The charge falls rapidly in real terms, to $8.5 million by year 10 and $3 million by year 20.

Inflation creates a front-end load of capital charges in traditional rate-base regulation. The front-end load creates a rate shock when new assets are added to the rate base. It makes capital charges depend on asset vintages, violating the old theorem that the price of tomatoes does not depend on the age of the farmer’s tractor.

The obvious solution is to allow a real rate of return on an inflation-indexed rate base. The mechanics of this trended original cost (TOC) method are identical to the mechanics of Treasury inflation-protected securities and other inflation-indexed bonds. I recommended that the FERC adopt TOC and provided detailed instructions for doing so.34

TOC sounds simple, but showing that it could be simple and practical took a long and method- ical report. I had to resolve prior debates among regulatory scholars about whether compensating for inflation was fair and efficient (for detailed coverage of prior literature, see Myers, Kolbe & Tye 1985, pp. 96–100). Convincing the FERC to accept TOC was an uphill battle. In the end, after a successful court challenge to the FERC’s first Williams decision, FERC Order 154-B adopted TOC but only for the equity-financed portion of the rate base.

US regulators have a queer compulsion to track costs of debt and equity separately. Life would be simpler if regulators allowed an overall WACC and let companies worry about financing, subject to a maximum debt ratio or to a credit-quality constraint. Regulated companies should have no first-order reason to operate at high debt ratios because interest tax shields are passed on to customers. But regulators insist on allowing only actual, “embedded” interest payments, thus passing interest-rate risk on to rate payers. Then they consider the cost of equity separately. They push for high debt ratios, usually forgetting MM’s proposition 2 that the cost of equity increases with financial risk. Companies push for low debt ratios because they expect that regulators will forget proposition 2.

4.7. Railroads and Stand-Alone Cost

I briefly describe my work on stand-alone cost (SAC) in the railroad industry. Railroad rates in the United States are mostly unregulated, but in some circumstances, shippers can appeal to the Surface Transportation Board for rate relief. Relief can come if a shipper can show that a railroad is charging more than the SAC of providing service. SAC equals what a new, efficiently equipped and configured railroad would have to charge to achieve breakeven NPV.

The SAC rule is cumbersome and unpopular because of the cost of inventing and documenting a hypothetical new railroad. But set that problem aside and consider the SAC rule itself. It is based on the idea of contestable markets, in which the threat of entry by a new, efficient entrant constrains prices charged by an incumbent (Baumol, Panzar & Willig 1982).

A potential entrant faces downside risk because much of the required investment for a new railroad, including rights of way, tracks, bridges, and tunnels, is irreversible. But if demand takes off, entrant 1 has to worry about a later entrant 2. The threat from entrant 2 thus limits the upside to entrant 1 as well as the incumbent. Entrant 2 would have to worry about a later entrant 3, etc. All entrants are deterred by the constrained upside and therefore would demand a cushion of extra current revenue, that is, current revenue higher than would be required in a risk-free world, before exercising their options to invest and enter. I estimated the size of the required cushion using

34 Here I am skipping past my analysis of the prior ICC method, which was ad hoc but worked reasonably well, and in my opinion would have been acceptable going forward. See Myers, Kolbe & Tye (1984) for details.

www.annualreviews.org • Finance, Theoretical and Applied 23

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

reasonable assumptions about uncertainty. Unfortunately, the Surface Transportation Board has yet to adopt this real-option version of SAC.

This result is an interesting twist to simple real-options analyses, in which upside and downside payoffs on the underlying asset are assumed symmetric. With contestable markets and SAC, the upside is truncated. Entrant x acquires an asset with returns constrained on the upside because of the threat of entry by Entrants x + 1, x + 2, etc. Valuation of the real option to enter requires numerical methods (for details, see Hausman & Myers 2002).

4.8. Litigation and Arbitration

Since the 1980s, I have been less engaged in regulatory finance and more in financial issues arising in litigation and arbitrations. Often the assignments have been fascinating. I have been fortunate to work on the aftermaths of the Penn Central, Enron, and Parmalat bankruptcies; on calculating damages in antitrust and patent infringement cases; on international transfer pricing for pharmaceuticals; on pricing and profitability in the credit-card business; and on corporate- owned life insurance programs. I have valued oil fields expropriated by Iran and Venezuela. I have calculated damages to rescuers of failing savings and loans when the US Congress passed legislation nullifying the “regulatory forbearance” granted in the rescue contracts.

Expert witnesses are portrayed by some as hired guns. I was hired. I worked with lawyers to define the questions that needed answers. My task was to design a gun and see if I could hit the target. I wrote my own reports and directed all numerical analyses. With very few exceptions, I never got pressure from clients’ lawyers to change my analysis or conclusions. That may happen with inferior lawyers and supple experts but was not a problem for me. The greater danger came when opposing lawyers and experts said I was wrong and implied that I was stupid. Then came the temptation to join my client’s adversarial team too vigorously.

I have been fortunate to work with excellent lawyers, and I have received first-rate professional support. I particularly note Larry Kolbe, Bill Tye, Lynda Borucki, Jamie Read, and Frank Graves, longstanding colleagues at The Brattle Group, Inc.

4.9. Insurance

Insurance is an important branch of finance, although one has to learn the lingo before tackling insurance problems. In the early 1980s, my late MIT colleague Richard Cohn and I were asked how to calculate a zero-NPV insurance premium. Automobile and workers’ compensation insurance was regulated in Massachusetts, and local insurance companies wanted us to take a look. At the time, the state of the art was Fairley (1979), who derived a formula based on the CAPM.

Rich and I drew on corporate finance for an APV equation, later published as Myers & Cohn (1987):

Premium = P = (1 − T ) × PV(losses and expenses) + PV(tax cost of holding P + S).

PV(losses and expenses) is an actuarial expectation discounted to PV. T is the corporate tax rate. P + S is the sum of premium and surplus, which is assumed invested in a portfolio of securities. Surplus is an allocation of equity to assure that the portfolio is adequate collateral for payment of losses.

The tax cost of holding collateral is the corporate tax on investment income—the tax that would not have to be paid if shareholders could invest the collateral on their own. If corporate borrowing generates valuable interest tax shields, then corporate lending must generate negative

24 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

tax shields, that is, tax costs. Thus, PV(tax cost of holding P + S) is valued in the same way as PV(interest tax shields), but with signs changed, of course.

The APV equation assumes an efficient market, so that the NPV of investing is zero, regardless of risk. Thus, the APV calculation can ignore how P + S is invested. Fairley (1979) also realized that, for rate setting, an insurance company can be assumed to invest premiums and surplus in risk-free assets. Shifting investment from safe to risky assets is zero NPV in an efficient market. But how does the tax cost of investing depend on risk?

It doesn’t. The tax cost for a one-period investment in any risky asset can be calculated as if the asset were risk free. If the risk-free rate is rf , the tax cost per dollar invested = Trf /(1 + rf ). This is the “Myers formula” in Derrig (1994). The reader should be able to derive the formula in three lines of simple algebra.

The Myers-Cohn formula is not too difficult to apply. One needs good estimates of losses and expenses, of course, and a reasonable discount rate for these outflows. In Massachusetts, the discount rate was calculated from the CAPM, assuming a small negative beta.

From 1993 to 1998, I was a director of CAT, Ltd., a reinsurance company that wrote protection against hurricanes, earthquakes, and other natural disasters. CAT (short for catastrophe) was based in Bermuda, largely because the corporate tax rate there is zero.

Relief from taxes is crucial for companies such as CAT, which insure against low-probability events. The PV(tax cost of holding P + S ) is much larger, compared with the actuarial PV of losses, for natural disasters than for a portfolio of routine auto or homeowner’s policies. Suppose a CAT policy covers up to $100 million of hurricane damage at a defined location. If the probability of a hurricane with that damage at that location is 1%, then the PV of losses and expenses is in the neighborhood of $1 million. But the collateral required to assure payment is P + S of $100 million, roughly 100 times the premium. Thus, the tax costs of collateral would take over the Myers-Cohn formula if this hurricane insurance were written in the US. But PV(tax cost of holding P + S ) is minimal in Bermuda and other low-tax jurisdictions.

The next big insurance issue was to figure out how to allocate surplus to a line of business, defined as a portfolio of similar policies. An insurance company does not finance lines separately. If it defaults on one line, it defaults on them all. Nevertheless it is important to know which lines soak up the most surplus and which lines consume relatively little.

Suppose a company’s surplus is costly or constrained. It wants to calculate breakeven (zero- NPV) premiums for line i. It defines Si as the allocation of the company’s overall surplus (equity) to the ith line35 and uses Si to charge back part of the cost or shadow price of surplus to line i. Finding the optimal values for Si is called the surplus allocation problem.

Myers & Read (2001) solved this problem. Start with the value of the insurance company’s default put over the next period, say, the next year. The default put measures the default risk absorbed by policyholders. Calculate the derivative of the value of the put with respect to a change in the scale of each line. Allocate surplus proportional to these marginal default values. In other words, set marginal surplus allocations to offset differences in marginal default values, so that all lines have the same net marginal impact on the value of the default put. The resulting allocations always add up exactly to total surplus, regardless of the joint probability distribution of line-by- line losses. We presented extensive numerical examples showing how and why these allocations work. The allocations can be derived in closed form for joint normal distributions of line-by-line losses.

35 A company such as CAT, which writes a relatively small number of large policies, should allocate surplus to individual policies.

www.annualreviews.org • Finance, Theoretical and Applied 25

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

4.10. Risk Capital

Myers & Read (2001) won both the annual best-paper awards from the Journal of Risk and In- surance. But our allocation method was hardly noticed outside the insurance industry. This was disappointing because the basic result, that risk capital (surplus or equity) should be allocated on the basis of marginal default values, clearly applied to banks and other financial institutions. So Isil Erel, Jaime Read, and I undertook a general paper on the allocation of risk capital.

We made several false starts that were largely my fault. For example, readers of early versions saw our paper as “all about risk-based capital requirements for banks” and expected discussions of bank regulation, the Basel rules, etc. We mentioned deposit insurance, which opened a door to other detours from our analysis. We also had to contend with a common belief that allocations of risk capital are dangerous and should be avoided.36

Erel, Myers & Read (2015) finally arrived at a general theory of risk capital, a significant extension of Myers & Read (2001), although risk capital is still allocated based on marginal default values. Many of the theory’s implications for practice are at odds with actual practice. Here are three examples.

4.10.1. Value at risk. Capital requirements in practice are often based on value at risk (VaR), which measures the downside risk of an asset or portfolio over a short time period.37 Higher variance of return means higher VaR. Thus, it is natural to assume that a risk-free asset with zero variance and zero VaR requires zero supporting capital. Wrong: The capital allocation for a risk-free asset is always negative when held in a portfolio with risky assets. Expansion of the risk-free asset decreases the value of the default put and frees up capital for other uses. Of course, we agree that the stand-alone capital allocation for a risk-free asset is zero.

4.10.2. APV versus risk-adjusted return on capital. Suppose a financial firm’s risk capital (eq- uity) is costly because of taxes (no interest tax shields on equity) and because capital is constrained in the short run. The firm is considering additional investment in assets Ai for line of business i. The investment adds value if its APV is positive:

APVi = ∂ V ∂ Ai

= NPV( Ai ) − (τ + κ)c i > 0. (4)

Here V is firm value, τ is the one-period tax cost (not the tax rate) per dollar of equity, κ is the shadow price of the capital constraint, and ci is the capital allocation based on marginal default value.

APV depends on profitability and market risks, which determine the NPV of additional in- vestment δAi, and on the portfolio of market- and firm-specific risks, which determine the capital allocation ci and the risk-capital charges c i (τ + κ). The APV formula clarifies the difficulties built in to risk-adjusted return on capital (RAROC) as a measure of risk-adjusted profitability. RAROC is an after-tax rate of return on the “economic” capital allocated to (and implicitly invested in) a business or bundle of assets. RAROC can be thought of as a hurdle rate for NPV. But there is

36 We agreed that allocations based on marginal default values do not apply when an entire new business is added or subtracted. This is the case considered by Merton & Perold (1993). But our allocations nevertheless apply when a firm chooses its portfolio of lines of business. Assume a firm maximizes market value, subject to a credit-quality constraint. Credit quality is defined as the ratio of default-put value to the face value of outstanding debt. Capital allocations based on marginal default values are part of the conditions for an optimum portfolio. 37 The period for capital allocation has to be short because of the liquidity of many assets held by financial institutions; see Myers & Rajan (1998).

26 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

no practical way to stuff two different risk premiums in one rate. The risks are (a) the market risk of business i, for example, its CAPM beta, and (b) the marginal default value of business i, which depends on risks specific to the firm’s portfolio of businesses. You can always find a discount rate that calculates NPV = APV in one step, but only if you know APV first.

Again and again I have found that APV is required to solve problems in corporate finance. Once the comforting assumptions of WACC are left behind, there is no practical alternative to APV.

APV is simple: Just add two or more PVs. Theory, once you get it right, usually yields simple measures. The shortcuts and ad hoc methods that many practitioners adopt often make practice more complicated.

4.10.3. Digression. Notice that I said “simple,” not “easy.” A golf swing is simple, not easy. Even with APV, valuing risky assets with confidence is irreducibly difficult, absent a trusted market price, because it requires not only forecasting, but also understanding the future. As Keynes (1936, p. 149) noted, “Our knowledge of the factors which will govern the yield of an investment some years hence is usually very slight and often negligible.” Understanding the past is hard enough.

This was Keynes’s main point in his famous chapter in The General Theory of Employment, Interest and Money on investing, the stock market, and beauty contests.

Investment based on genuine long-term expectation is so difficult today as to be scarcely practical. . .In abnormal times. . .when. . .continuance of the present state of affairs is less plausible than usual, even if there are no express grounds to anticipate a definite change, the market will be subject to waves of optimistic and pessimistic sentiment, which are unreasoning, but in a sense legitimate, where no basis exists for a reasonable calculation. (Keynes 1936, pp. 154, 157; emphasis supplied)

The difficulty of valuing risky assets creates space for Keynes’s oft-quoted beauty contest in which investors’ attention is diverted from fundamental values to guessing what other investors think. It is tempting to interpret the beauty contest as a parable showing why investors are silly to ignore fundamentals, but Keynes was really saying that even the smartest investors often have no choice.

Economists may assume that the actors in their models know what fundamental value is, but they should be careful not to believe that they can calculate fundamental values personally. Keynes inspired my paper on fuzzy efficiency, which I wrote as a commentary on the crash of 1986. “Fundamental value is not a definite number, but a fuzzy band of possible values” (Myers 1988, p. 8; see also Black 1986 on “noise”).

4.10.4. NPV and APV: pretax or after tax? In corporate finance, the tax-adjustment term in APV is usually expressed as a tax advantage of debt rather than a tax cost of equity. NPV is calculated after tax at an opportunity cost of capital, as if the investment were all-equity financed and the PV of interest tax shields is then added to get APV. Interest tax shields depend on the amount of debt supported by investment. But if a financial corporation allocates tax or other costs of risk capital, as in Erel, Myers & Read (2015), NPV should be calculated as if investment were 100% financed by tax-free financing, that is, debt. The tax cost of the allocated capital required to support investment is then subtracted.

The two APV formats are of course equivalent. The corporate-finance format makes sense for nonfinancial firms, where the default financing is equity. APV in Equation 4 makes sense for banks and other financial firms, where the default financing is debt. Myers & Read (2014) defined APV in the corporate-finance framework, but we could have used Equation 4.

www.annualreviews.org • Finance, Theoretical and Applied 27

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

5. ACT 3: BACK TO THE THEORY OF CORPORATE FINANCE

A dark night in a city that knows how to keep its secrets, but one man is still trying to find the answer to life’s persistent questions

—Guy Noir, Private Eye, in the introduction to Guy Noir skits on Prairie Home Companion

Corporate finance also knows how to keep its secrets. After 50 years of work in finance, I am still searching for the secret of optimal capital structure as well as the secrets to the persistent questions about payout policy, capital investment, corporate governance, mergers and acquisitions (tendentiously called “the market for corporate control”), and risk management.

This search may be quixotic. Many financing decisions are just tactics. The financial behavior of corporations may be as diverse as life itself and just as resistant to specific prediction. But I keep hoping for some compact theory that will provide fresh insight.

My research strategy for these persistent questions has evolved as follows:

1. Do not try for a “theory of everything” that covers all types of firms and circumstances. Look first for a theory that works for mature, blue-chip public corporations in countries with good financial markets and good corporate governance.

2. Stick to models of agency. It is much better to assume that managers act entirely in their own interests, subject to governance constraint, than to assume that they are acting, however imperfectly, to maximize shareholder value. We have made great strides with theory based on value maximization, but diminishing returns have set in.

3. Avoid private benefits if they are purely psychological, for example, the pride and prestige of running a larger corporation. Do not assume that such private benefits will drive man- agers to overinvest whenever they can. Overinvestment should be a modeling result, not an assumption.

4. Explore dynamics. Write down a model with repeated decisions over an indefinite horizon. The model may be in discrete or continuous time, but the risks borne at any point in time should be ordinary probability distributions, not two or three matchstick outcomes.

5. Take care when using models designed for individual principals and agents to describe public corporations, which are complex organizations. Do not assume that a CEO is the only agent. CEOs are not puppet masters who can control all that goes on in their firms. CEOs’ actions are constrained because they must motivate following generations of managers to work hard and specialize their human capital to the firm’s assets (see Acharya, Myers & Rajan 2011).

6. Do not invoke frictions and capital-market imperfections unless absolutely necessary. No doubt they exist, but by invoking them, one risks ending up with a coatrack model of the sort that I criticize in Act 1.

7. Sources and uses of cash must match. Consider the three big decisions about borrowing, CAPEX, and payout (dividends plus repurchases), taking the current period’s operating cash flow after interest and taxes as given:

Operating cash flow + change in debt = CAPEX + payout. There are three decisions but only two degrees of freedom. Thus, there cannot be three separate theories of debt policy, CAPEX, and payout. If one assumes that CAPEX responds solely to investment opportunities, then a theory of payout must also be a theory of debt policy. Theories of debt policy that take CAPEX as given must assume that payout is the residual shock absorber. (Equity issued to pay down debt amounts to negative payout.) The assumption of payout as a shock absorber is embedded in most capital structure models, including mine, but rarely mentioned explicitly.

28 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Myers (2000) and follow-on papers all start with the expanded market-value balance sheet anticipated in Myers (1999):

A Assets D Debt

R PV(managers’ rents)

E Outside shareholders’ equity

V Firm value V Firm value

Managers can extract rents because governance is imperfect. Outside equity investors have full property rights. They can intervene and throw out the managers, forcing managers’ rents to zero. But there is a cost of intervention. Outside investors’ net payoff from taking over is α(V − D ), where α < 1, say, 0.8. Thus, α < 1 creates the space for managers to extract rents, but they pay out enough in each period so that investors do not intervene. They pay out no more than enough; the governance constraint is binding. Outside investors end up sharing cash flow and value, say, 80% to investors and 20% to managers.

Myers (2000) focused on the relationship between managers and outside equity investors. The paper is overly technical and did not explore financing. It did introduce several ideas:

1. Outside equity survives because of an intertemporal constraint: Payout today must always ensure investors’ participation for at least one more period.

2. CAPEX by (risk-neutral) managers can be value maximizing. The market-value balance sheet above has two classes of equity: common stock (E) and the PV of managerial rents (R). In equilibrium, the binding governance constraint requires E and R to share in gains and losses proportionally. Therefore, managers’ and shareholders’ interests can be perfectly aligned. Managers must coinvest, however. If a firm buys a new asset for $1, then governance constraint forces managers to put up 1 − α cents by cutting back rents. (In Lambrecht & Myers 2012, 2015, CAPEX is mostly debt financed, and managers coinvest by assuming the fraction 1 − α of future debt service.)

3. Firms whose future value depends on human effort and risk-taking may go public to reduce the bargaining power of outside equity. Public ownership avoids the threat of holdups by private equity.

4. It can be much more difficult to confirm new investment than the existence of assets in place. As a result, outside equity investors must monitor the disposition of operating cash flows among operating costs, investment, and rents. Monitoring can then lead to evasive actions by managers, for example, transfer of pecuniary rents to less-efficient perquisites.

The first three follow-up papers examined specific cases. Jin & Myers (2006) explained the finding by Morck, Yueng & Yu (2000) that firm-specific risk is lower, relative to market risk, for common stocks of companies in countries with low per-capita GDP and less well-developed financial systems. The key is opaqueness, which usually comes along with poor governance and helps sustain it. Opaqueness allows managers to capture more rents out of sight of investors. The bad news for managers is that they have to absorb more firm-specific risks.

Jin & Myers (2006) extended the ideas in Myers (2000) to a model of transparency versus opaqueness and the allocation of risk bearing. We recognized, however, that managers’ capacity to absorb hidden bad news is limited. At some point, they will give up and proclaim a credible crisis. Thus, we predicted that extreme negative firm-specific returns should be more frequent in countries with poor governance and less transparency. We confirmed this prediction empirically.

Lambrecht & Myers (2007) considered takeovers of declining firms that sooner or later will disinvest. We assumed that investors’ property rights are automatically enforced in bankruptcy,

www.annualreviews.org • Finance, Theoretical and Applied 29

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

so that they receive full value V = A in liquidation. If V declines sufficiently, then managers voluntarily give up, abandoning their rents instead of paying out cash to generate an adequate rate of return for investors on V. Managers give up inefficiently late, however. They always exercise the abandonment put at an asset value below the value-maximizing exercise value.

Jensen (1986) proposed that managers will overinvest free cash flow if given the chance. Jensen thus explained hostile takeovers and greenmail agreements in the 1980s. I believe his explanation of those events was correct. Jensen’s insight was then expanded to a “free cash-flow theory” of corporate finance, in which unconstrained managers always overinvest free cash flow. But the theory, if one can call it that, gives no satisfactory explanation of why managers would always want to overinvest and of what limits investment absent takeovers. There is no overinvestment in Lambrecht & Myers (2007). Investment is inefficient, not because managers overinvest free cash flow in good times, but because they wait too long to disinvest in bad times.

Hostile takeovers are one solution for inefficient disinvestment. Lambrecht & Myers (2007) showed that a raider will take over and exercise the target’s abandonment put at exactly the value- maximizing point. A hostile takeover by another corporation is likewise efficient, although there may be a temptation not to disinvest after the takeover is complete. Takeovers are efficient because protagonists’ costs of collective action are nearly zero. We attributed the takeover premiums paid to target-firm shareholders, not to synergies, but to recapture by shareholders of the value of rents that would otherwise go to managers. Thus, we provided a formal model for the views expressed in Shleifer & Summers (1988).

Lambrecht & Myers (2008) introduced debt financing and showed that the burden of debt service can accelerate abandonment. The level of debt can in principle be set to assure optimal abandonment. If the abandonment value of assets is a fixed amount K, optimal debt is D = K and safe. Optimal debt is D > K, however, if K is correlated with uncertain asset value. In that case, managers encounter debt overhang problems and temptations for risk-shifting, which can distort CAPEX for expansion, just as in Myers (1977).

Lambrecht & Myers (2007, 2008) assumed risk-neutral managers. They assumed that any leftover free cash flow is split between managers and outside investors and consumed immediately. Thus, managers do not manage their flow of rents over time. These papers consider the effects of different debt levels but do not say anything about debt and payout dynamics.

Lambrecht & Myers (2012) analyzed payout policy for mature firms that generate free cash flow and can make regular payouts to shareholders. Payout equals the sum of dividends and repurchases. We introduced managerial risk aversion. We held the capital stock fixed and explained the dynamics of payout, rents, and borrowing. We ignored default risk, however. If the governance constraint is binding, with fixed α, then we proved that rents and payout must move in lockstep. If managers want to smooth rents over time, they must also smooth payout.

There is no doubt that mature corporations smooth dividends and, more recently, overall payout (see Skinner 2008, table 6). The usual explanations of smoothing are nearly circular, however. One might hear, “Dividend changes convey information. Investors expect managers to smooth dividends and raise dividends only when confident that the increase can be maintained. Managers smooth dividends because they do not want to send false signals to investors.” We say that payout is smoothed because rents are smoothed, and payout and rents move together. Investors do not have to care.

If managers (acting as a coalition) have negative exponential utility and habit formation for rents, we can derive Lintner’s (1956) target-adjustment payout equation, 64 years after it was published. The finance profession still uses this equation empirically, but no one had derived it.

Lambrecht & Myers (2012) held the capital stock fixed. Payout and rents move in lockstep and are smoothed. There is only one possible shock absorber, corporate borrowing or lending.

30 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Managers use corporate borrowing or lending to implement their optimal lifetime path of rents. They save for the future by paying down corporate borrowing. With time and adequate profits, their precautionary savings push the firm to negative debt; the firm becomes a net lender.

The role of debt as the main corporate-finance shock absorber is especially intriguing. It suggests that mature firms may follow the pecking-order theory, not because of information asymmetry, but because managers as agents act in their long-term self-interest. But Lambrecht & Myers (2012) is primarily about payout policy. The paper does not include taxes or default, so the level of debt is completely irrelevant to shareholders. The capital stock is fixed, so debt is not needed to finance CAPEX. The debt level matters to managers only as a long-term limit on the PV of their rents.

Lambrecht & Myers (2015) introduced taxes and valuable interest tax shields. CAPEX is now endogenous: Managers reset the optimal level of the capital stock every period. We solved simul- taneously for the optimal paths of CAPEX, debt, and payout plus rents. I believe this is a first in dynamic agency modeling.

This paper is not yet published, and I do not go through it in detail here. Instead, I highlight our results for debt policy, the persistent question that Alex Robichek set me to work on in 1964. First, debt remains the shock absorber, as in Lambrecht & Myers (2012). Changes in debt are much more volatile than payout and rents. Debt follows a kind of pecking order.

Second, differences in debt levels can last for long periods, consistent with (dangerous words again) Lemmon, Roberts & Zender (2008). If firm A has borrowed more than B has, then A’s debt level will stay above B’s forever if both experience the same sequence of economic shocks. Third, managers’ precautionary saving will still push debt levels down, profitability permitting, even when interest is tax deductible.

Fourth, managers may have to increase payout to compensate investors for foregone interest tax shields. But managers of firms with low debt levels will not rebalance to higher debt ratios. Once managers are on their optimal path of rents, they do not gain by issuing an additional chunk of debt and using the proceeds for extra payout and rents. Such rebalancing cannot improve managers’ expected lifetime utility; their local optimum is also their global optimum. Managers’ time value of money at the optimum, defined by the marginal utilities of future versus current rents, conforms to their corporation’s after-tax borrowing rate. The NPV of additional borrowing is zero if after-tax debt service is also discounted at the after-tax rate. We arrived at a Miller (1977) equilibrium, but with managers rather than shareholders calling the shots.38

Thus, we have a new explanation of why successful corporations often operate at low, even negative, debt ratios and leave interest tax shields unexploited. The pecking-order theory oper- ates again, although in a more complicated fashion, because payout, rents, and CAPEX are all endogenous.

Our model leaves out many things. We should probably try adding default risk and costs of financial distress—at the sacrifice of closed-form solutions—although I do not see how including them could generate an optimal target debt ratio for managers.

In Lambrecht & Myers (2012, 2015), all corporate governance is reduced to one parameter, α. We should think harder about how corporate governance works for firms in different stages of their life cycle, particularly for younger firms with valuable growth options. We should think harder about our assumed coalition of managers and ask whether the coalition is bound together by career concerns, as in Acharya, Myers & Rajan (2011), or by joint investments in firm-specific

38 In the Miller (1977) equilibrium, shareholders’ time value of money, which is the rate of return on safe equity, equals the after-tax corporate rate.

www.annualreviews.org • Finance, Theoretical and Applied 31

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

human capital. We should think harder about the implications of our models for the design of corporate governance. Stay tuned.

DISCLOSURE STATEMENT

The author is not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

LITERATURE CITED

Acharya V, Myers SC, Rajan R. 2011. The internal governance of firms. J. Finance 66(3):689–720 Arrow KJ. 1964. The role of securities in the optimal allocation of risk-bearing. Rev. Econ. Stud. 31(2):91–96 Asquith P, Mullins DW. 1986. Equity issues and offering dilution. J. Financ. Econ. 15(1–2):61–89 Baumol W, Panzar JG, Willig RD. 1982. Contestable Markets and the Theory of Industry Structure. New York.

Harcourt Brace Jovanovich Black F. 1986. Noise. J. Finance 41(3):529–43 Brealey RA, Myers SC. 1981. Principles of Corporate Finance. New York: McGraw-Hill. 1st ed. Brealey RA, Myers SC, Allen F. 2014. Principles of Corporate Finance. New York: McGraw-Hill. 11th ed. Bhattacharya S. 1978. Project valuation with mean-reacting cash flows. J. Finance 33:1317–31 Derrig R. 1994. Theoretical considerations of the effect of federal income taxes on investment income in

property-liability ratemaking. J. Risk Insur. 61:691–709 Diamond DA, He Z. 2014. A theory of debt maturity: the long and short of debt overhang. J. Finance

69(2):719–62 Donaldson G. 1961. Debt Capacity: A Study of Corporate Debt Policy and the Determination of Debt Capacity.

Boston, MA: Div. Res. Harvard Grad. Stud. Bus. Admin. Erel I, Myers SC, Read JA Jr. 2015. A theory of risk capital. J. Financ. Econ. In press. doi:10.1016/

j.jfineco.2014.10.006 Fairley WB. 1979. Investment income and profit margins in property-liability insurance. Bell J. Econ. 10:191–

210 Fama EF. 1977. Risk adjusted discount rates and capital budgeting under uncertainty. J. Financ. Econ. 5(1):3–24 Fama EF. 1996. Discounting under uncertainty. J. Bus. 69(4):415–29 Fama EF, French K. 2005. Financing decisions: Who issues stock? J. Financ. Econ. 76(3):549–82 Federal Power Commission et al. v. Hope Natural Gas Co., 320 U.S. 591 (1944) Fisher FM, McGowan JJ. 1983. On the misuse of accounting rates of return to infer monopoly profits. Am.

Econ. Rev. 73:82–97 Frank MZ, Goyal VK. 2003. Testing the pecking order theory of capital structure. J. Financ. Econ. 67:217–48 Fulghieri P, Garcia D, Hackbarth D. 2014. Asymmetric information and the pecking (dis-)order. Work. Pap.,

Boston Univ., Univ. N.C. Gordon MJ, Shapiro E. 1956. Capital equipment analysis: the required rate of profit. Manag. Sci. 3:102–10 Graham JR, Leary MT. 2011. A review of empirical capital structure research and directions for the future.

Annu. Rev. Financ. Econ. 3:309–45 Hausman J, Myers SC. 2002. Regulating U.S. railroads: the effects of sunk costs and asymmetric risk. J. Regul.

Econ. 22(3):287–310 Healy P, Howe C, Myers SC. 2002. R&D accounting and the tradeoff between relevance and objectivity: a

pharmaceutical industry simulation. J. Account. Res. 40(3):677–710 Hirshleifer J. 1965. Investment decision under uncertainty: choice-theoretic approaches. Q. J. Econ. 79(4):509–

36 Hirshleifer J. 1966. Investment decision under certainty: applications of the state-preference approach. Q. J.

Econ. 80(2):252–77 Jensen MC. 1986. Agency costs of free cash flow, corporate finance and takeovers. Am. Econ. Rev. 76(2):323–29 Jin L, Myers SC. 2006. R2 around the world: new theory and new tests. J. Financ. Econ. 79(2):257–92 Kaplan S, Ruback R. 1995. The valuation of cash flow forecasts: an empirical analysis. J. Finance 50(4):1059–93

32 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Keynes JM. 1936. The General Theory of Employment, Interest and Money. New York: Macmillan Kolbe AL, Tye WB, Myers SC. 1993. Regulatory Risk: Economic Principles and Applications to Natural Gas Pipelines

and Other Industries. New York: Springer Lambrecht B, Myers SC. 2007. A theory of takeovers and disinvestment. J. Finance 62(2):809–45 Lambrecht B, Myers SC. 2008. Debt and managerial rents in a real-options model of the firm. J. Financ. Econ.

89(2):209–31 Lambrecht B, Myers SC. 2012. A Lintner model of dividends and managerial rents. J. Finance 67(5):1761–810 Lambrecht B, Myers SC. 2015. The dynamics of investment, payout and debt. Work. Pap., Mass. Inst. Technol. Lemmon ML, Roberts MR, Zender J. 2008. Back to the beginning: persistence and the cross-section of

corporate capital structure. J. Finance 63(4):1575–608 Lemmon ML, Zender J. 2010. Debt capacity and tests of capital structure theories. J. Financ. Quant. Anal.

45:1161–87 Lintner J. 1956. Distribution of incomes of corporations between dividends, retained earnings and taxes. Am.

Econ. Rev. 46:97–113 Lintner J. 1965. Optimal dividends and corporate growth under uncertainty. Q. J. Econ. 77:59–95 Majd S. Myers SC. 1987. Tax asymmetries and corporate income tax reform. In The Effects of Taxation on

Capital Accumulation, ed. M Feldstein, pp. 343–73. Chicago, IL: Univ. Chic. Press Majluf NS. 1978. Study on mergers: a rationale for conglomerate mergers. PhD Thesis, Mass. Inst. Technol.,

Cambridge Merton RC, Perold AF. 1993. Theory of risk capital in financial firms. J. Appl. Corp. Finance 6:16–32 Miles J, Ezzell R. 1980. The weighted average cost of capital, perfect capital markets and project life: a

clarification. J. Financ. Quant. Anal. 15(3):719–30 Miller MH. 1977. Debt and taxes. J. Finance 32:261–76 Miller MH, Modigliani F. 1961. Dividend policy, growth and the valuation of shares. J. Bus. 34:411–33 Miller MH, Modigliani F. 1966. Some estimates of the cost of capital to the electric utility industry, 1954–57.

Am. Econ. Rev. 56:333–91 Modigliani F, Miller MH. 1958. The cost of capital, corporation finance and the theory of investment. Am.

Econ. Rev. 48(3):261–96 Modigliani F, Miller MH. 1963. Corporate income taxes and the cost of capital: a correction. Am. Econ. Rev.

53:433–43 Morck R, Yeung BY, Yu W. 2000. The information content on stock markets: Why do emerging markets

have synchronous stock price movements? J. Financ. Econ. 58:215–60 Mullins DW Jr. 1976. Communications Satellite Corp. Case Study 276195, Harvard Bus. Sch., Cambridge, MA Myers SC. 1967. Effects of uncertainty on the valuation of securities and the financial decisions of the firm. Ph.D.

Thesis, Stanford Univ., Stanford, CA Myers SC. 1968a. A time-state-preference model of security valuation. J. Financ. Quant. Anal. 3:1–33 Myers SC. 1968b. Procedures for capital budgeting under uncertainty. Ind. Manag. Rev. 9(3):1–20 Myers SC. 1972. Application of finance theory to public utility rates cases. Bell J. Econ. 3:58–97 Myers SC. 1973a. A simple model of firm behavior under regulation and uncertainty. Bell J. Econ. 4:304–15 Myers SC. 1973b. On public utility regulation under uncertainty. In Risk and Regulated Firms, ed. RH Howard,

pp. 32–46. East Lansing: Div. Res. Grad. Sch. Bus., Mich. State Univ. Myers SC. 1974. Interactions of corporate financing and investment decisions: implications for capital bud-

geting. J. Finance 29:1–25 Myers SC. 1977. Determinants of corporate borrowing. J. Financ. Econ. 5(2):147–75 Myers SC. 1984a. Finance theory and financial strategy. Interfaces 14:126–37 Myers SC. 1984b. The capital structure puzzle. J. Finance 39(3):575–92 Myers SC. 1988. Fuzzy efficiency. Inst. Invest. 1988(Dec.):8–9 Myers SC. 1989. Still searching for optimal capital structure. In Are the Distinctions Between Debt and Equity

Disappearing?, ed. RW Kopke, ES Rosengren, pp. 80–95. Boston: Fed. Reserve Bank Myers SC. 1999. Financial architecture. Eur. Financ. Manag. 5(2):133–41 Myers SC. 2000. Outside equity. J. Finance 55(3):1005–37

www.annualreviews.org • Finance, Theoretical and Applied 33

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07CH01-Myers ARI 6 November 2015 16:51

Myers SC, Cohn R. 1987. A discounted cash flow approach to property-liability insurance rate regulation. In Fair Rate of Return in Property Liability Insurance, ed. JD Cummings, S Harrington, pp. 55–78. Dordrecht, Neth.: Kluwer-Nijhoff

Myers SC, Dill DA, Bautista AJ. 1976. Valuation of financial lease contracts. J. Finance 31(3):799–819 Myers SC, Howe C. 1997. A life cycle model of pharmaceutical R&D. Work. Pap., Progr. Pharm. Ind., Mass.

Inst. Technol. Myers SC, Kolbe AL, Tye WB. 1984. Regulation and capital formation in the oil pipeline industry. Transp.

J. 23(4):25–49 Myers SC, Kolbe AL, Tye WB. 1985. Inflation and rate of return regulation. Res. Transp. Econ. 2:83–119 Myers SC. Majd S. 1990. Abandonment value and project life. In Advances in Futures and Options Research,

Vol. 4, ed. F Fabozzi, pp. 1–21. Greenwich, CT: JAI Myers SC, Majluf NS. 1984. Corporate financing and investment decisions when firms have information that

investors do not have. J. Financ. Econ. 13(2):187–221 Myers SC, Pogue GA. 1974. A programming approach to corporate financial management. J. Finance

29(2):579–99 Myers SC, Rajan R. 1998. The paradox of liquidity. Q. J. Econ. 113(3):733–71 Myers SC, Read JA Jr. 2001. Capital allocation for insurance companies. J. Risk Insur. 68(4):545–80 Myers SC, Read JA Jr. 2014. Real options, taxes and leverage. Work. Pap., Mass. Inst. Technol. Myers SC, Shyam-Sunder L. 1996. Measuring pharmaceutical industry risk and the cost of capital. In Compet-

itive Strategies in the Pharmaceutical Industry, ed. RB Helms, pp. 208–37. Washington, DC: Am. Enterp. Inst.

Myers SC, Turnbull SM. 1977. Capital budgeting and the capital asset pricing model: good news and bad news. J. Finance 32(2):321–33

Robichek AA, Myers SC. 1965. Optimal Financing Decisions. Upper Saddle River, NJ: Prentice-Hall Robichek AA, Myers SC. 1966a. Conceptual problems in the use of risk-adjusted discount rates. J. Finance

21(4):727–30 Robichek AA, Myers SC. 1966b. Problems in the theory of optimal capital structure. J. Finance Quant. Anal.

1(2):1–35 Robichek AA, Myers SC. 1966c. Valuation of the firm: effects of uncertainty in a market context. J. Finance

21(2):215–27 Rosenbaum J, Pearl J. 2013. Investment Banking. New York: Wiley. 2nd ed. Schwartz E. 1959. The theory of the capital structure of the firm. J. Finance 14:18–39 Sharpe WF. 1965. Capital asset prices: a theory of market equilibrium under conditions of risk. J. Finance

19(3):425–42 Shleifer A, Summers LH. 1988. Breach of trust in hostile takeovers. In Corporate Takeovers: Causes and Conse-

quences, ed. AJ Auerbach, pp. 33–56. Chicago, IL: Univ. Chic. Press Shyam-Sunder L, Myers SC. 1999. Testing static tradeoff against pecking order models of capital structure.

J. Financ. Econ. 51(2):219–44 Skinner DJ. 2008. The evolving relation between earnings, dividends and stock repurchases. J. Financ. Econ.

87:582–609 Solomon E, Laya J. 1967. Measurement of company profitability: some systematic errors in the accounting

rate of return. In Financial Research and Management Decisions, ed. AA Robichek, pp. 152–83. New York: Wiley

Van Horne J. 1966. Capital budgeting decisions involving combinations of risky assets. Manag. Sci. 19:B84–92 Williams JB. 1938. The Theory of Investment Value. Cambridge, MA: Harvard. Univ. Press

34 Myers

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07-FrontMatter ARI 6 November 2015 15:6

Annual Review of Financial Economics

Volume 7, 2015Contents

Finance, Theoretical and Applied Stewart C. Myers � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Consumption-Based Asset Pricing, Part 1: Classic Theory and Tests, Measurement Issues, and Limited Participation Douglas T. Breeden, Robert H. Litzenberger, and Tingyan Jia � � � � � � � � � � � � � � � � � � � � � � � � � � � �35

Consumption-Based Asset Pricing, Part 2: Habit Formation, Conditional Risks, Long-Run Risks, and Rare Disasters Douglas T. Breeden, Robert H. Litzenberger, and Tingyan Jia � � � � � � � � � � � � � � � � � � � � � � � � � � � �85

Behavioral Finance David Hirshleifer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 133

Contributions to Defined Contribution Pension Plans James J. Choi � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 161

The Value and Risk of Human Capital Luca Benzoni and Olena Chyruk � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 179

Asset Price Bubbles Robert A. Jarrow � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 201

Disaster Risk and Its Implications for Asset Pricing Jerry Tsai and Jessica A. Wachter � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 219

Analytics of Insurance Markets Edward W. Frees � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 253

The Role of Risk Management in Corporate Governance Andrew Ellul � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 279

The Axiomatic Approach to Risk Measures for Capital Determination Hans Föllmer and Stefan Weber � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 301

Supervisory Stress Tests Beverly Hirtle and Andreas Lehnert � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 339

Financial Stability Monitoring Tobias Adrian, Daniel Covitz, and Nellie Liang � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 357

v

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .

FE07-FrontMatter ARI 6 November 2015 15:6

An Overview of Macroprudential Policy Tools Stijn Claessens � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 397

A Review of Empirical Research on the Design and Impact of Regulation in the Banking Sector Sanja Jakovljević, Hans Degryse, and Steven Ongena � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 423

Financing Innovation William R. Kerr and Ramana Nanda � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 445

Peer-to-Peer Crowdfunding: Information and the Potential for Disruption in Consumer Lending Adair Morse � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 463

Hedge Funds: A Dynamic Industry in Transition Mila Getmansky, Peter A. Lee, and Andrew W. Lo � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 483

Recent Advances in Research on Hedge Fund Activism: Value Creation and Identification Alon Brav, Wei Jiang, and Hyunseob Kim � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 579

Private Equity Performance: A Survey Steven N. Kaplan and Berk A. Sensoy � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 597

Real Estate Price Indices and Price Dynamics: An Overview from an Investments Perspective David Geltner � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 615

Governance of Family Firms Belén Villalonga, Raphael Amit, Marı́a-Andrea Trujillo, and Alexander Guzmán � � � � 635

Indexes

Cumulative Index of Contributing Authors, Volumes 1–7 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 655

Cumulative Index of Articles Titles, Volumes 1–7 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 658

Errata

An online log of corrections to Annual Review of Financial Economics articles may be found at http://www.annualreviews.org/errata/financial

vi Contents

A nn

u. R

ev . F

in an

c. E

co n.

2 01

5. 7:

1- 34

. D ow

nl oa

de d

fr om

w w

w .a

nn ua

lr ev

ie w

s. or

g A

cc es

s pr

ov id

ed b

y O

kl ah

om a

S ta

te U

ni ve

rs it

y -

S ti

ll w

at er

o n

11 /0

9/ 16

. F or

p er

so na

l us

e on

ly .