Introduction Some Stylized Facts Model Estimation Counterfactuals Conclusion Equity Market Misvaluation, Financing, and Investment

Transcription

1 Equity Market, Financing, and Investment Missaka Warusawitharana Toni M. Whited North America meetings of the Econometric Society, June 2014

2 Question Do managers react to perceived equity mispricing? How do managers react?

3 Heavily researched but hard question One cannot observe mispricing. Reactions to mispricing are endogenous. Two exclusion restrictions. Literature lacks a quantitative assessment.

4 Alternative approach Estimate a dynamic model of investment and financing. Deviations of intrinsic from market value. Issuances and repurchases when equity is over or under valued. Model parameters dictate The magnitude of these reactions The uses and sources of these funds

5 Results Model fits a broad set of data moments: large heterogeneous samples across industries shocks: statistically significant. Equity issuers time the market, but proceeds primarily go in and out of net saving. Timing adds a small amount of value for long term shareholders.

6 Let s start with some stylized facts about equity Equity Issuance, Repurchases, Dividends Return Repurchases Dividends Equity Issuance Return

7 Returns and cash saving are highly correlated Saving, Debt Issuance Return Return Saving Debt Issuance

8 Investment is much smoother than returns Investment Return Return Investment

9 The model of the firm Infinite horizon, discrete time, partial equilibrium Manager chooses stocks of physical and financial assets Manager maximizes the value of a controlling block of shareholders

10 The real side of the model is standard zk constant returns profit function ln (z ) = µ + ρ z ln (z) + ε z law of motion for z I K (1 δ)k investment A (I, K) λi2 2K physical adjustment costs

11 Financing from internal cash, external debt, and external equity C 0 Net cash earns/pays an interest rate r f one-period discount bond E Issuances (+) or repurchases ( ) a 0 Fixed equity issuance cost T = (zk δk + Cr f )τ c Tax bill

12 Dividends are separate from repurchases Dividends equal sources minus uses of funds Sources D = zk I λi2 2K T + C(1 + r) C + E a 0 I(E > 0) Either Uses

13 Market values may diverge from intrinsic V (K, C, ψ, z) Intrinsic value V = ψ (V (K, C, ψ, z) D) of ex-dividend value σ2 ψ ln (ψ ) = + ρ 2(1+ρ ψ ) ψ ln (ψ) + ε ψ Law of motion for ψ Issuances and repurchases are valued at V

14 Dilution Mispricing affects dilution (or concentration) of the controlling shareholders block stake. V V + E = ψ(v D(1 τ)) ψ(v D(1 τ)) + E Market participants draw inference on the size of misvaluation from the transaction size, with a concave price impact function. ( ) ( ) E 2 E 2 ν(e, K) ν i I(E > 0) + ν r I(E 0) 2K 2K The price impact adversely affects the dilution (or concentration) of the controlling stake. V V + E + ν(e, K) = ψ(v D(1 τ)) ψ(v D(1 τ)) + E + ν(e, K)

15 Bellman equation and intuition V (K, C, ψ, z) = max K,C,E ( + β D ψ(v D(1 τ)) ψ(v D(1 τ)) + E + ν(e, K) ) V ( K, C, ψ, z ) dg ψ ( ε ) ψ, ε ψ dg z ( ε ) z, ε z Current Dividend Dilution/Concentration Ratio Continuation Value

16 Policy and value functions Policy Function States cash productivity shock misvaluation shock c z ψ Decisions tomorrow s net cash c investment i issuances/repurchases e dividends d Value Function States net cash productivity shock misvaluation shock c z ψ Value intrinsic value market value v v

17 Optimal policies: {c, i, e, d} = h(c, ψ, z) for estimated parameters 0.10 Net cash Dividends Investment Equity transactions Shock Net cash Dividends Investment Equity transactions Profitability Shock

18 Model estimation Estimate some parameters directly by calculating averages: r f, δ, τ c For the rest of the parameters use SMM (Ingram and Lee, 1991). Draw from (z, ψ) Compute v (c, ψ, z) and (c, e, i) = h (c, ψ, z) Repeat Simulated data Interesting moments Real data Interesting moments

19 Identification Use moments that vary when the underlying structural parameters vary. Do not cherry-pick. Use many moments. Identify the variance and autocorrelation of misvaluation shocks. The moments involving profits are only affected by the profitability shock parameters.

20 Matched moments Early/Small Early/Large Late/Small Late/Large Data Model t-stat Data Model t-stat Data Model t-stat Data Model t-stat Average net cash Std. dev. of net cash Autocorrelation net cash Average investment Std. dev. of investment Autocorrelation investment Average profits Std. dev. of profits Autocorrelation profits Std. dev. of Tobin s q Autocorrelation Tobin s q Return std. dev Return serial correlation Average equity issuance Std. dev. of issuance Average repurchases Std. dev. of repurchases Issuance-return sensitivity Issuance Incidence

21 Parameter estimates Early Small Early Large Late Small Late Large λ δ ρ ψ σ ψ µ ρz σz ν i νr ρ zψ a 1 φ (1.36) (0.01) (0.12) (0.03) (0.10) (0.04) (0.04) (4.36) (2.20) (0.15) (0.00) (0.06) (0.60) (0.01) (0.17) (0.04) (0.16) (0.09) (0.05) (3.94) (1.12) (0.22) (0.00) (0.61) (2.08) (0.01) (0.13) (0.02) (0.08) (0.03) (0.04) (3.31) (2.01) (0.06) (0.01) (0.36) (2.40) (0.01) (0.19) (0.03) (0.09) (0.05) (0.03) (1.41) (1.15) (0.26) (0.21) (0.02)

22 Maybe we are just picking up a risk factor When we add a pricing kernel, not much happens.

23 Industry Estimation Actual Net Cash Actual Investment Simulated Net Cash Simulated Investment

24 Industry Estimation (Contd.) Actual Equity Issuance Simulated Equity Issuance Actual Repurchases Simulated Repurchases

25 The parameter estimates make sense SIC13 SIC 20 SIC 28 SIC 35 SIC 36 SIC 38 SIC 50 SIC 73 λ (0.713) (15.310) (2.144) (8.748) (1.975) (9.515) (17.169) (6.447) δ (0.004) (0.017) (0.052) (0.013) (0.017) (0.011) (0.046) (0.024) ρ ψ (1.025) (0.432) (0.110) (0.806) (0.767) (1.245) (0.724) (0.801) σ ψ (0.114) (0.114) (0.111) (0.192) (0.135) (0.158) (0.179) (0.107) µ (0.042) (0.212) (0.552) (0.191) (0.215) (0.288) (0.316) (0.201) ρz (0.032) (0.105) (0.231) (0.059) (0.112) (0.151) (0.126) (0.110) σz (0.064) (0.138) (0.168) (0.174) (0.126) (0.107) (0.166) (0.081) ν i (3.260) (3.228) (8.946) (4.070) (33.227) (5.818) (7.072) (2.054) νr (58.041) (7.478) (22.729) (4.283) (0.882) (2.556) (8.110) (1.501) ρ zψ (1.591) (0.517) (0.201) (0.666) (0.699) (1.071) (0.531) (0.721) a (0.031) (0.154) (0.224) (0.011) (0.033) (0.020) (0.012) (0.006) φ (0.453) (0.068) (0.035) (0.123) (0.084) (0.083) (0.441) (0.195) Average R&D

26 The model fits out-of-sample correlations Early/Small Early/Large Late/Small Late/Large Actual Simulated Actual Simulated Actual Simulated Actual Simulated Investment Net Saving Equity Issuance Repurchases Dividends

27 Contribution to firm value Timing value 1.10 lal shock standard deviation

28 Impulse response functions 0.02 Profit shock shock 0.01 Net Cash Time 0.02 Profit shock shock 0.01 Investment Time

29 Conclusion Use structural estimation to infer the effects of equity misvaluation on corporate policies. The standard deviation of misvaluation shocks is significant. Firms issue equity in response to overvaluation and repurchase equity in response to undervaluation. Proceeds mostly funneled into cash. does help ease the effect of costly external finance on investment.