Efficient Capital Markets, Inefficient Firms: A Model of Myopic Corporate Behavior

Stein, Jeremy, “Efficient Capital Markets, Inefficient Firms: A Model of Myopic Corporate Behavior,” The Quarterly Journal of Economics (1989), 655-669.

Claim: Managers will invest myopically, even when they care about stock price and even when markets behave rationally.

Model:

  • Stock markets determine company valuation based on current earnings.
  • Managers have an incentive to forgo profitable investments now in order to boost earnings (cash flow) now.
  • Market are rational, so they will expect managers to do this.
  • Managers who care about stock prices will be trapped into behaving myopically, since they will be penalized if they do not boost earnings now.
  • In a steady-state “signal-jamming” model, managers will inflate earnings by “borrowing” from the future (i.e. not investing optimally), and markets will correctly estimate this borrowing.
  • Capital market pressure determines the strength of this phenomenon, and may take one of the following forms:
    • Threats of takeover (when a firm’s stock price is low)
    • Lack of financial slack (the ability to undertake investments without issuing new stock)
    • Distance between the firm and its creditors (creditors who are intimate with a firm will rely less on the stock price to appraise firm value)
    • The degree to which current earnings are a good signal of future earnings.

Conclusions:

  • Myopic behavior, where managers underinvest to boost short-term earnings and market rationally expect this and adjust valuations accordingly, is a Nash equilibrium.
  • This positive market reaction to announced investments is not evidence that managers do not underinvest, as argued by Jensen (1986).
    • Underinvesting managers only undertake the best projects, so markets are pleased when myopic managers invest.
  • Startups with high stock prices and high investments are not evidence that managers are not myopic.
    • For startups, current earnings have little correlation with long-run success (most startups have negative earnings). The link between stock prices and underinvestment does not yet exist.
  • Corporate divestitures and breakups are evidence supporting the signal-jamming model.
    • Markets can better interpret the investments and future earnings of stand-alone companies than of conglomerates, so breakups reduce the signal-jamming inefficiencies

Managerial Incentive Problems: A Dynamic Perspective

Holmström, B., “Managerial Incentive Problems: A Dynamic Perspective,” Review of Economic Studies 66 (1999), 169-182.

Purpose: To show how a manager’s desire to build his resume and firm owners’ desire to earn financial returns can conflict or harmonize.

Findings:

  • Fama (1980) posited that market forces eliminate the agency problem over time.  A manager will choose to maximize shareholder returns because, in the long run, he is concerned about his long-term career prospects.
    • Under some narrow assumptions, Fama is right, but in general, he is not.
  • When the market is uncertain about a manager’s ability, he will work harder to achieve good outcomes and look like a high-ability manager.
    • Ability changes over time and is never fully known, so managers can always substitute effort for ability.
  • Managers’ supply of effort/output in every period is less than the socially optimal level.
  • A risk-averse manager may not want to invest, since investing poorly could reveal low managerial ability, and since it cannot be proved that investments not made would have been successful.

The Basic Model:

  • Effort is unobservable and cannot be contracted, so managers in each period are paid in advance for their efforts.
  • A manager’s ability is revealed over time through the outcomes of his effort.
  • A risk-neutral market pays a manager wages w_t(y^{t-1} = E[\eta | y^{t-1}] + a_t(y^{t-1}).
    • y^{t-1} is the series of historical output through time t-1.
    • w_t is wages paid at the beginning of period t.
    • \eta represents belief about manager ability.
    • a_t \in [0,1] is the manager’s decision rule, or how much effort he gives in period t.
  • The manager maximizes the expected discounted payoff of his current and future wages, minus the current and expected disutility of effort g(a_t), where g(\cdot) is an increasing and convex function.
  • Solving these two equation gives the equilibrium wage and managerial decision rule.
  • When there is uncertainty, there are gains to effort.
    • High effort biases the market’s ability estimating process upward, and lower effort biases it downward.
  • Managerial ability changes over time, and so never becomes fully known.

Debt and Seniority: An Analysis of the Role of Hard Claims in Constraining Management

Hart, Oliver, and John Moore, “Debt and Seniority: An Analysis of the Role of Hard Claims in Constraining Management,” The American Economic Review, Vol 85, No 3 (1995), 567-585.

  • Assume that managers will always invest in every project (due to empire-building, prestige, resumé-building, etc.)
    • If managers have too much cash flow, they will invest in bad (negative NPV) projects along with good ones.
    • Debt can fix the agency problem by forcing managers to pay out cash flows before investing.
  • The debt overhang problem
    • If managers have too little cash, they will have to forgo investing in good projects.
    • If managers have too much debt, they will not risk losing their job and, again, will forgo investing in good but risky projects.
  • Long-term debt can outperform complicated contracts in solving agency problems.
    • Contracts that attempt to align manager compensation with shareholders’ interests can’t address every agency problem.
    • Long-term debt restricts managers from borrowing against future cash flows to invest in bad projects.

Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure

.Jensen, Michael C. and William H. Meckling, “Theory of the Firm:  Managerial Behavior, Agency Costs and Ownership Structure,” Journal of Financial Economics, Vol 3, No 4 (1976), 305-360.

Purpose:  To show how agency costs influence the firm’s ownership structure.

Agency costs are the sum of:

  • Monitoring costs incurred by the principal
  • Bonding costs incurred by the agent (costs incurred by the manager to guarantee or to prove his fidelity)
  • The residual loss (monitoring and bonding cannot usually prevent all agency-related problems)

This paper takes a positive approach, seeking to explain contracting methods rather than recommend methods.

A “firm” is a nexus of contracts between individuals, rather than a cohesive body with clear boundaries.  Therefore, firm behavior, like market behavior, is an equilibrium outcome.

Agency costs of outside equity:

  • If an owner-manager sells a fraction of the firm, his incentive to maximize firm value declines and he will increase his perquisites, thus reducing firm value.
  • In a rational market, equity purchasers of X% of the firm will only pay X times the value they expect the firm to have once the manager increases his perquisites.
  • Firm value is lower after a sale of equity, and since the outside buyer paid a fair price given the expected reduction in firm value, the loss of wealth is entirely imposed on the owner-manager who sold the equity.
  • The manager may be forced to sell equity to raise money for investment, but his total welfare may still increase if the investment is profitable enough.
  • monitoring and bonding can increase firm value and manager wealth regardless of who (principal or agent) incurs the costs.

Agency costs of debt:

  • Consider two investments requiring the same outlay, but with different payoffs and probabilities of success.
  • The high-variance investment, with a higher expected payoff but a lower probability of success, imposes risk on the bondholders.
  • A manager can increase his wealth by promising to invest in the low-variance project, selling bonds, and subsequently investing in the high-variance project.
  • In a rational market, bondholders will expect the manager to do this, so they will not pay the full (low-variance project) price for the bonds.
  • If the manager could fully self-finance, he would choose the high-variance project.
  • If the manager cannot finance the project and must sell bonds, then the bond buyers will not pay enough to finance the high-variance project.  The agency costs are the loss of value associated with forgoing the higher-expected-value project.
  • Monitoring costs (external audits, bond covenants, etc) can sometimes put a burden on the firm.
  • If the manager can produce information for monitors more cheaply than the monitors themselves (having internally collected data simply certified by external auditors), it will be worth it to do so.  This is referred to as bonding.
  • Bankruptcy costs and reorganization costs may also be incurred if the firm cannot meet its debt obligations.

Conclusions:

  • Managers want to spend a lot of money on perks.
  • higher leverage means managers’ equity forms a larger part of the whole, so managers’ incentives are more aligned with shareholders.
  • If all outside financing is debt, or if all is equity, agency costs one way or the other will likely be too high; there is a balancing point.
  • Inside vs. Outside Financing:
    • Inside financing avoids agency costs, but limits the manager’s ability to diversify his personal investments and limits the size of projects he can undertake.
    • Outside financing is often necessary to raise the capital needed for an investment.
    • Outside financing is often preferred by a risk-averse manager who wishes to diversify.

Dynamic Agency and the q Theory of Investment

DeMarzo, Peter M., Michael J. Fishman, Zhiguo He, and Neng Wang, 2012, “Dynamic Agency and the q Theory of Investment,” The Journal of Finance, Vol. 67, No. 6 (2012), 2295-2340.

Purpose:  To introduce an agency problem into the standard q theory of investment; to show that cash flow is not the best predictor of investment.

Motivation:  A large body of literature uses cash flow to predict firm investment levels.  This paper argues that a better proxy is “financial slack,” which is directly related to the agency problem.

Model:

  • Productivity is a Brownian motion, and the agent controls the drift but not the volatility.
  • w = W/K is the agent’s total expected payoff per unit of capital, and must be high enough to incentivize the agent to maximize productivity.
  • The level of w depends on λ, σ, and historical firm profitability.
    • λ is a measure of the extent of the agency problem.
    • σ is the volatility of firm productivity.
    • Past productivity raises or lowers w, and the agent loses his job when w = 0.
    • A portion of w is deferred, giving the agent a stake in continued firm success.
  • Investor’s expected payoff per unit of capital, p(w), is a function of how much they pay the agent.
  • Average q is total firm value per unit of capital stock, or qa = p(w) + w.
  • Marginal q, or qm = p(w)wp’(w).
    • Firms invest when marginal q is less than 1, so investment is a function of w. It follows that investment depends upon λ, σ, and past firm performance.
  • “Financial slack” equals w/λ, and is the largest productivity shock the firm can suffer without changing agents.
  • The agent accumulates cash and available credit equal to the firm’s “financial slack,” then distributes excess income to shareholders.

Conclusions:

  • Financial slack is a better predictor of investment than cash flow.
  • Average q is higher than marginal q because an increase in capital stock K reduces w, and hence reduces the agent’s [historically determined] incentives to maximize productivity.
  • Financial slack and profitability are substitutes in determining average q.
  • When the firm is profitable, w rises, the agent’s incentives grow, and return on investment increases.
    • Investment is serially correlated.
  • The cost of incentivizing the agent leads to underinvestment in every state of the world.

Why do U.S. Firms Hold So Much More Cash than They Used To?

Bates, Thomas W., Kathleen M. Kahle, and René M. Stulz, 2009, “Why Do U.S. Firms Hold So Much More Cash than They Used To?” Journal of Finance, Vol 64, No. 5 (2009), 1985-2021.

Purpose:  To document the increase in U.S. firms’ cash holdings from 1980-2006 and investigate its causes.

Motivation:  U.S. firms in 2006 hold twice as much cash (as a percentage of assets) than they did in 1980.  The average firm in 2006 has enough cash on hand to pay off all its debts.  This cash buildup has excited considerable popular attention.

Findings:

  • Firms of all sizes have a much higher cash-to-asset ratio in 2006 than in 1980.
  • This also holds for firms with no foreign income, so repatriation taxes do not explain it.
  • Cash buildup is observed in firms that do not pay dividends, but not in dividend-payers.
  • Firms whose idiosyncratic risk increased saw more growth in cash holdings.
    • Firms in industries that grew riskier
    • Younger firms (with more recent IPOs)
  • As the increase in idiosyncratic risk has reversed in recent years, cash holdings have declined slightly.

Data/Methods:

  • Data is a sample of 13,599 U.S.-based, nonfinancial, non-utility firms with positive sales and positive assets, over the period 1980-2006 (from WRDS merged CRSP/Compustat database).
  • Look at descriptive statistics of the average cash ratios, with the data segmented by size, industry risk, year of IPO, dividends paid, and net income.
  • Ask whether firms’ cash demand curve shifted or whether firms moved along the curve.
    • Regress cash-to-assets ratio on several firm characteristics (R&D expenditures, size, etc).
  • Connect the rising cash levels to specific firm characteristics.
    • Use Fama-Macbeth regressions with 1980s data to predict cash holdings, based on several firm characteristics, in the 1990s and 2000s.
    • Compare prediction to actual holdings.
  • Examine whether increased cash holdings is due to agency problems.
    • Compare yearly average cash for quintiles based on the GIM index of management entrenchment.

Conclusions:

  • Firms whose cash holdings increased the most are newer, do not pay dividends, and are in industries that have become riskier since 1980.
  • Cash holdings rose because inventories fell, R&D expenses increased, capital expenditures fell, and cash flow risk increased.
  • Cash holdings did not increase more quickly for firms with entrenched management.

Measuring investment distortions arising from stockholder-bondholder conflicts

Parrino, Robert, and Michael S. Weisbach, 1999, “Measuring investment distortions arising from stockholder-bondholder conflicts,” The Journal of Financial Economics 53, 3-42.

Purpose:  This paper calculates the expected wealth transfer between stock- and bondholders occurring when a firm begins a new project.  It also estimates how stockholder-bondholder conflict impacts investment decisions, and whether it can explain cross-sectional variation in capital structure.

Findings:

  • There will be underinvestment when the firm is faced with safe projects.
    • Stockholders demand a higher return than the CAPM rate.
      • This effect is stronger for high-leverage firms.
    • Safe projects with low returns benefit bondholders at the expense of equity holders.
  • There will be overinvestment when the firm is faced with risky projects.
    • Stockholders are willing to gamble, and will even invest in negative NPV projects if the potential payoff is high.
      • This effect is also stronger for high-leverage firms.
    • Risky projects with negative expected (but high potential) payoffs benefit equity holders at the expense of bondholders.
  • Longer debt duration is vulnerable to larger agency problems
  • Lower marginal tax rates lead to slightly larger distortions
  • The stockholder-bondholder conflict does exist and there do seem to be empirical investment distortions, but these distortions are too small to be useful in explaining most firms’ capital structure decisions.
    • The distortion is only 0.14% for a firm with 20% debt-to-capital ratio, compared with a 3% noise factor in measuring cost of capital

parrinoweisbach

Motivation:  Managers seek to maximize shareholder value, not necessarily firm value.  The “underinvestment problem” is when managers avoid a positive NPV project that would increase firm value but would lower stockholder value.  The “overinvestment problem” is when managers undertake a negative NPV project that lowers firm value but raises stockholder value.  These agency problems have been widely discussed in the literature for decades, but there is no consensus on their magnitude or on how important they really are.

Data/Methods:  Use numerical simulations to estimate the impact of debt on the investment decisions of a levered firm whose managers seek only to maximize stockholder value.  Compute the stockholder-bondholder wealth transfer accompanying projects with known characteristics.

  • Compustat data for firms from 1981-1995
  • Monte Carlo:  Assume a firm with known cash flows following a random walk without drift for 30 years, after which cash flows are static.  Assume a project financed entirely with equity, whose cash flows similarly follow a random walk without drift for 30 years.  Assume a correlation of 0.5 between firm and project cash flows, run the simulation 5,000 times, and compute the ex ante value of debt and equity each time.
    • Value of debt is the sum of discounted future cash flows to bondholders.
    • Value of equity is the discounted cash flows to stockholders plus a terminal value.

Conclusions:  Distortions in stockholder-bondholder required investment returns vary along several dimensions.  However, for the typical firm they are much smaller than the noise in cost-of-capital measurement.  The effect exists but is too small to explain cross-sectional variation in capital structure.

Enjoying the Quiet Life? –commentary

This paper uses a clever check for reverse causality.  The authors find that for firms whose incorporating state adopts anti-takeover laws in the preceding twelve months, wages rise and productivity falls.  The reverse causality story is that rising wages (perhaps due to union pressure) and falling productivity threaten firms in a certain state, and therefore make that state more likely to enact anti-takeover laws.  The authors’ story depends on causality in the right direction.

The paper’s main tests use a binary variable BC to indicate whether a firm’s incorporating state enacted anti-takeover laws in the prior 12 months.  The coefficients on BC in the several regressions are the paper’s main findings.

In the reverse-causality check, the authors replace BC with four dummy variables:

  • Before-1 for firms incorporated in states that had not yet passed legislation but would do so in the next 12 months
  • Before0 for firms incorporated in states that had passed legislation in the prior 12 months
  • After 1 for firms incorporated in states that had passed legislation between 12 and 24 months prior
  • After2+ for firms incorporated in states that had passed legislation two or more years previously

Significant coefficients on Before-1 would have suggested that legislation was passed in response to an already-changing business environment.

Before0 was the variable of primary interest.

Small and Insignificant coefficients on After 1 and After2+ would have suggested that the effects of legislation were short-lived and therefore economically much less interesting.

In this case, the coefficients on Before-1 were small and statistically insignificant, ruling out the reverse causality.  The coefficients on After 1 and After2+ were significant and even larger than the coefficients on Before0, strengthening support for the authors’ story since their reported effects of anti-takeover laws continued to grow in the longer term.

Enjoying the Quiet Life? Corporate Governance and Managerial Preferences

Bertrand, Marianne, and Sendhil Mullainathan, 2003, “Enjoying the Quiet Life?  Corporate Governance and Managerial Preferences,” The Journal of Political Economy 111 (5), 1043-1075.

Purpose:  This paper examines the effect of anti-takeover laws on a variety of firm behaviors.

Findings:  Following the passage of anti-takeover laws, blue-collar wages rise 1%, white-collar wages rise 4%, and plant creation and destruction both fall so that firm size does not significantly change.  Capital expenditures are unaffected.  Total factor productivity falls.   Return on capital falls by 1%.  Findings contradict stakeholder theory that proposes increased efficiency when workers are paid more.  Findings contradict “empire-building” theories of corporate governance that suggest unfettered managers opt to increase firm size.

Motivation:  The reduced-form agency problem is our assumption that managers desire to pursue their own goals, which may not align with shareholders’ best interests.  There are many theories—but no consensus—regarding what managers’ personal goals actually are.

Data/Methods:

  • Longitudinal Research Database details plant-level employment and wages and plant creation and destruction.
    • The LRD does not include data on workers’ age, education, or tenure.
  • Compustat includes firm-level financial data, and each firm’s state of incorporation.
  • The Census of Auxiliary Establishments has better data on white-collar workers than the LRD.
    • These data are limited to the firm level (matching to specific plants is not possible)
    • These data are not available in every year, so we cannot analyze the trend
  • Corporate governance and firm behavior are endogenous, but this is overcome by studying anti-takeover laws passed by several states at different times.  The laws weakened governance by limiting the threat of hostile takeover, but were not driven by specific characteristics of any firm.  Laws were also passed at different times, so many firms belong both to the treatment group and to the control group in different years.
  • Analysis of firm-level outcomes
    • Difference-in-differences using firms incorporated in states that recently passed anti-takeover laws as the treatment group
  • Analysis of plant-level outcomes
    • The laws passed affected all firms incorporated in the passing state, regardless of the actual plant location
    • Control for regional economic and political variation by considering plants located in the same regions, one incorporated in a passing state, and one incorporated in some other state
  • Check for reverse causality, whereby states with rising wage pressures are more likely to pass anti-takeover legislation.  We do not find significant evidence of this.

Conclusions:  Anti-takeover legislation does change firm behavior.  Managers pay blue-collar workers more and pay white-collar workers much more, thus transferring more benefits to stakeholders.  Contrary to stakeholder theory, this benefit to stakeholders does not create overall improvement, as firm efficiency declines.  Managers avoid either opening or closing plants, undermining the “empire-building” view of manager preferences.  Managers appear to prefer “the quiet life,” with less employee conflict and fewer hard decisions.

commentary