Game theory and finance • Perfect competition: - Walras equilibrium in prices. • Game theory: - Each agent’s behaviour affects other agents’ payoffs. - Nash equilibrium in actions. Each agent’s action is optimum if other agents do their equilibrium actions. Tore Nilssen – Lecture 8: Corporate Finance II - 1 Applying game theory to corporate finance: • Uncertainty: Expected payoff - Each player knows the uncertainty he faces and the uncertainty each other player faces. • Time: Sequential rationality - Strategies versus actions: strategy = ”plan of actions” - Actions must be in equilibrium when they are made. - Think ahead. Reason back. • Uncertainty and time - One player’s action may affect other players’ beliefs. - Perfect Bayesian Equilibrium: both strategies and revised beliefs are in equilibrium. - Given the strategies in equilibrium, which revised beliefs are consistent with these strategies? - Given the way beliefs are revised in equilibrium, which strategies are in equilibrium? Tore Nilssen – Lecture 8: Corporate Finance II - 2 Signalling game Stage 1: The informed player chooses an action (signals) Stage 2: The uninformed player observes the action chosen at stage 1, revises his beliefs about the informed player, and chooses himself an action. The informed player’s private information – his type • θ ∈ T = {High, Low} The uninformed player’s beliefs about the other’s type: • Pr(High) = pH, Pr(Low) = pL = 1 – pH • stage 2: revised beliefs Equilibrium: actions and revised beliefs • Stage 1: • Stage 2: player 1’s action player 2’s revised beliefs player 2’s action Tore Nilssen – Lecture 8: Corporate Finance II - 3 Hidden information • Disregard conflicts of interest owners vs. management • Focus on conflict of interest owners vs. outside investors - insiders vs. outsiders • The firm has private information - the value of the firm (assets in place) - the value of a new project • The Pecking Order - Myers & Majluf J Fin Econ 1984 - Internal funds > Debt > Equity Tore Nilssen – Lecture 8: Corporate Finance II - 4 A simplified problem: Equity financing New equity is an expensive source of financing because, in the external capital market, high-value firms will have to - either ”mingle” with low-value firms at a high price - or stay out of the capital market - resorting to internal funds, or - not invest A model (Daniel & Titman, without ”money burning”) Value of the firm (assets in place): θ ∈ {H, L}, H > L. The value of the firm is its private information. Investors’ ex-ante beliefs: Pr(θ = H) = p Eθ = θ = pH + (1 – p)L Project revenue: V Project costs: I < V Project value (NPV): V – I > 0. Tore Nilssen – Lecture 8: Corporate Finance II - 5 Endogenous variables: • The firm’s decision whether to invest, for each type θ. • Outsiders’ revised beliefs, given observed decision: q = outsiders’ subjective probability that θ = H, given the firm’s decision (invest or not). If the firm invests and outsiders have beliefs q, then outsiders expect the firm’s value to become: q(H + V) + (1 – q)(L + V) = V + qH + (1 – q)L Capital market is competitive. Investors are risk neutral. ⇒ Investors’ expected revenue = expected costs ⇒ f[V + qH + (1 – q)L] = I, where f = share of the firm owned by new owners / equity holders ⇒ f= I V + qH + (1 − q) L • df/dq < 0 Tore Nilssen – Lecture 8: Corporate Finance II - 6 • Current owners are left with: (1 – f)(θ + V), θ ∈ {H, L} Investment is profitable if: (1 – f)(θ + V) ≥ θ ⇔ f≤ V θ +V Case (i): The firm has low value: θ = L Investment is profitable if V I ≤ , V + qH + (1 − q) L L + V which always holds. I < V, and V + qH + (1 – q)L ≥ V + L. Thus: Low-value firms always invest. Tore Nilssen – Lecture 8: Corporate Finance II - 7 Case (ii): The firm has high value: θ = H Investment is profitable if V I ≤ ⇔ V + qH + (1 − q) L H + V q≥ I ( H +V ) −V ( L +V ) V ( H − L) Consistency of beliefs: (a) Since type L always invests, type H investing is consistent with beliefs only if: q = p Thus, an equilibrium exists in which - the firm invests, whether it has a high value or a low value, and - q = p, if: p≥ I ( H +V ) −V ( L +V ) ⇔ V ( H − L) V H +V ≥ I θ +V Tore Nilssen – Lecture 8: Corporate Finance II - 8 (b) Since type L always invests, type H not investing is consistent with beliefs only if: q = 0 Investment is not profitable if q≤ I ( H +V ) −V ( L +V ) V ( H − L) An equilibrium exists in which - the firm invests if and only if it has a low value, and - q = 0, if: I ( H +V ) −V ( L +V ) ≥0 ⇔ V ( H − L) V H +V ≤ I L +V Tore Nilssen – Lecture 8: Corporate Finance II - 9 Conclusion • Two equilibria • Adverse selection when only type L invests. V +H V +θ 1 only type L invests V +H V +L both equilibria are possible V I both types invest Open questions • Isn’t there anything a high-value firm can do in order to convince investors about its true quality? - ”money burning” • What if we introduce debt financing as an alternative to equity financing? Tore Nilssen – Lecture 8: Corporate Finance II - 10 ”Money burning” What can a high-value firm do in order to convince outside investors? – Waste money. • Inflate investment costs. • Deflate project value. (a) Inflate investment costs? – No, doesn’t work. (CHANGE? – consider taking this point (a) out) • Can the firm ask for more funds from investors than I in order to convince them that it is high-value? - equity financed money burning • Wasted money = C Tore Nilssen – Lecture 8: Corporate Finance II - 11 • If this does work, then the high-value firm asks for I + C, the low-value firm asks for I, and both types of firm invest. Outside investors will only believe this if - it is better for the low-value type not to pretend being high value: ⎛ I ⎜ 1 − ⎜⎜ L +V ⎝ C≥ ⎞ ⎟ ⎟⎟ L +V ⎠ ( )≥ ⎛ I +C ⎜ 1 − ⎜⎜ H +V ⎝ ⎞ ⎟ ⎟⎟ L +V ⎠ ( ) ⇔ I (H − L) L +V - it is better for the high-type not to pretend being low-value: ⎛ I +C ⎜ ⎜1− ⎜ H +V ⎝ C≤ ⎞ ⎟⎛ ⎞ ⎟⎜⎝ H +V ⎟⎠ ⎟ ⎠ ⎛ ≥ ⎜⎜1− ⎜ ⎝ I ⎞⎟⎛ ⎞ ⎟⎜⎝ H +V ⎟⎠ ⇔ L +V ⎟⎠ I (H − L) L +V • Both types indifferent at C = I(H – L)/(L + V). • If the H-type asks for more money, so can also the Ltype ask for more. Thus, C has to be very high to convince outsiders – actually, too high. Tore Nilssen – Lecture 8: Corporate Finance II - 12 (b) Deflate project value? – Yes, might work. • The firm may propose a project that is so ill-organized that it will bring about a revenue of V – C instead of V. • If this works, then the high-value firm presents a project with value V – C, the low-value firm presents a project with value V, and both types invest. Outside investors believe this if: - It is better for the low-value type not to pretend being high value: ⎛ ⎞ ⎛ ⎞⎟ I ⎜⎜1− I ⎟⎟( L + V ) ≥ ⎜⎜1− ⎟( L + V − C ) ⇔ ⎜⎝ L + V ⎠⎟ ⎝⎜ H + V − C ⎠⎟ L +V − C C + ≥1 ⇔ H +V − C I C2 – (H + V)C + (H – L)I ≤ 0 The lowest C satisfying this condition is: C*= ⎤ 1 ⎡⎢ H +V − (H +V )2 − 4I (H − L )⎥ ⎥⎦ 2 ⎢⎣ Tore Nilssen – Lecture 8: Corporate Finance II - 13 - It is better for the high-type to invest with a deflated project than not to invest: ⎛ ⎞ I ⎜ ⎟⎛ ⎞ ⎜1− ⎟⎜⎝ H +V −C ⎟⎠ ≥ H ⇔ ⎜ H +V −C ⎟⎠ ⎝ V–C≥I (CHANGE – consider a third constraint: better for H type to invest with deflated project than to invest with standard project, being believed to be low-risk. This constraint makes the second condition above redundant (is itself redundant) if V/I > (<) (H + V)/(L + V), which is the the critical value for underinvestment in the case aobe of no money burning.) • Deflated project value works as a signal of high value if: C* ≤ V – I ⇔ (ψ) (V – I)(H + I) – I(H – L) ≥ 0 ⇔ H −L ≤ (V − I )(H + I ) I i.e., if the types are not too different. We now have a separating equilibrium always: • If (ψ) holds, then both types invest and the Htype deflates its project’s value by C*. Tore Nilssen – Lecture 8: Corporate Finance II - 14 • If (ψ) does not hold, then only the L-type invests. In both kinds of equilibria, there are inefficiencies: • wasted resources, or • underinvestment. Question: Is the firm able to commit to deflate the project’s value after the investment capital has arrived? Tore Nilssen – Lecture 8: Corporate Finance II - 15 What about debt vs. equity? No uncertainty about project value ⇒ Debt without risk of loss, D = I. ⇒ Any project with V > I can be financed by debt. A more realistic model: • uncertainty in the firm’s value → equity financing costly • uncertainty in the project’s value → debt financing costly Tore Nilssen – Lecture 8: Corporate Finance II - 16