Econ 234C Corporate Finance Lecture 11: Capital Structure

Transcription

1 Econ 234C Corporae Finance Lecure 11: Capial Srucure Ulrike Malmendier UC Berkeley April 16, 2008

2 Homework 3 Tesing a new capial srucure heory he Depression Baby Theory of Capial Srucure Decisions How does a Depression Baby CEO s decision differ from he decision of Non-Depression Babies? Noe: Depression Babies are someimes defined as people born in he 1930s ; someimes as people who lived hroug he 1930s in heir eenageyearsandearlyadulhood(e.g.,borninhe1920s).youmay wan o ry boh definiions. In paricular, are Depression Babies averse o accessing he marke for exernal financing? Are hey paricularly averse o equiy issues? Or o deb issues? How do you inerpre he resuls? (Alernaive explanaions.)

3 Daa: Compusa and Execucomp (wrds.wharon.upenn.edu). You do no need SDC daa. Sofware: STATA. Programming: Rerieving he daa. Infiling he daa ino STATA. Calculaing deb/equiy issues. (Go o recen capial srucure lieraure when deciding abou he correc daa definiions, variable choices, ec.) Regression (Go o recen capial srucure lieraure and empirical regression specificaions, conrols ec. > papers handed ou las ime.)

4 Purpose: Geing o know wo imporan daa ses for Empirical Corporae (even beer). GeingoknowSTATA(evenbeer). Promising research direcion in capial srucure: managerial fixed-effecs.

5 Exra office hours on Research Proposals Wed, 4/23 afernoon. OHs saring a 1pm. Shee on my door: pick 20min slo beween 1and 3pm; if filled up sar adding afer 3pm...

6 1 Capial Srucure Theory Modigliani-Miller Theorem Proposiion (1958): Capial srucure irrelevance. Inuiion: Value addiiviy. If operaing cashflows are fixed, value of he pie unaffeced by spli-up of he pie. Pracical message: If here is an opimal capial srucure, i should reflec axes and/or specific marke imperfecions. [Myers 1993] leads o

7 Trade-off Theory Opimal capial srucure rades off ax savings from deb financing (ax-deducibiliy of ineres paymens on deb) agains coss of financial disress from deb financing (agency coss of issuing risky deb; deadweigh coss of liquidaion or reorganizaion; coss of deb overhang [Myers 1977]). versus Pecking-Order Theory Firms prefer inernal funds  safe deb  risky deb  quasi-equiy (e.g. converibles)  equiy.

8 2 Capial Srucure Empirics Empirical Tess Tradiional empirical approach: Analyze wha ype of financing is used o fill he financing defici. Financing defici = asse growh minus liabiliies growh minus growh in reained earnings. Financing defici mus be filled wih (ne) sales of new securiies. Specificaion D i = α + βdef i + ε i Predicion PO heory: β 1.

9 Incorporae TO heory deerminans of capial D i = α + β DEF DEF i + β T T i +β Q Q i + β size S i + β Π Π i + ε i wih T = asse angibiliy, Q = book-o-marke Size = log sales (al.: log asses) Π = profi = DEF has lile explanaory power.

10 1. Some sylized facs on financial policies Low leverage puzzle Firms seem o use exernal deb financing oo conservaively relaive o wha convenional rade-off models would predic Too many firms have almos no deb financing Also: Graham (JF 2000): finding on financial conservaism Quasi-Marke Leverage Mean Median < 1% < 5% N daa: COMPUSTAT/CRSP merged file; Book Deb: D9+D34; Marke Equiy: D25*D199; condiions:no financials, book asses> 10. Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

11 1.1 Sylized facs: Con d Inernaional comparisons Similar resuls across boh developed and developing counries Suble differences are sill imporan o explain: E.g.: size effec in Germany Quasi-Marke Leverage and Ineres Coverage Raio Developed Counries Leverage Coverage Raio Developing Leverage Coverage Raio USA 28/ Souh Korea 64 Japan 29/ India 35 Germany 23/ Malaysia 20 France 41/ Pakisan 19 Ialy 46/ Turkey 11 UK 19/ Brazil 10 Canada 35/ Mexico 14 Source: Rajan and Zingales (JF 1995) for developed counries (period: 1991; daa: Global Vanage; leverage: deb o capial (a/b: a: non-adjused; b: adjused); ineres coverage raio: EBITDA/Ineres; medians repored), Booh, Aivazian, Demirguc-Kun, and Maksimovic (JF 2001) for developing counries (period: ; daa: IFC; leverage: liabiliy-based esimaion; for Brazil and Mexico: book equiy Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

12 1.2 Sylized facs: Con d Persisence of leverage Leverage is heavily pah-dependen and persisen Explanaions: Baker and Wurgler (JF 2002), Welch (JPE 2004), Srebulaev (2004) Persisence of leverage Panel B: / lowes highes Panel A: / lowes highes Source: My esimaion; Daa: COMPUSTAT/CRSP annual merged; period: ; quanile 1: lowes leverage; rows: iniial leverage; columns: leverage in 1/10 years Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

13 1.3 Sylized facs: Con d Cross-secional deerminans of leverage Hisorically he mos acceped empirical ool Srebulaev (2004): criique Cross-secional deerminans of leverage Book Leverage Q-Marke Leverage Consan (22.92) (21.20) Marke-o-Book (-1.85) (-14.47) Tangibiliy (27.80) (21.88) Profiabiliy (-12.94) (-13.48) Log Size ( 1.28) ( 4.05) R N 52/ / Source: My esimaion; Daa: COMPUSTAT/CRSP annual merged; period: ; no financials, asses(d6)>10; Mehod: Fama-McBeh (1973); no adjusmen for -sas Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

14 1.4 Sylized facs: Con d Wha is ax advanage o deb In he absence of deb: π U = δ(1 τ) In he presence of deb: π L = (δ c)(1 τ) + c = π U + cτ Definiion of he marginal ax rae How large is ax advanage? The aggregae ax benefis of deb Gross Benefi Ne Benefi Los Gross Benefi Los Ne Benefi 4.7 N % of firm value; daa: from Graham (2000), COMPUSTAT and CRSP Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

15 1.5 Sylized facs: Con d Deb policy facors: CFO s perspecive: (Graham and Harvey (JFE 2001)) 1. Financial flexibiliy (59%) 2. Credi raing (57%) 3. Earnings and cash flow volailiy (48%) 4. Insufficien inernal funds (45%) 5. Tax advanage (45%) Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

16 1.6 Sylized facs: Con d Missing ables... Mean reversion Frequency and ypes of (a) defaul; (b) financial disress Privae vs public deb usage Sources of invesmen: inernally generaed cash, equiy, deb Complexiy of deb srucure: (a) disribuion of insrumens/rusees Covenans used in deb conracs Credi raings Response o business cycles Ilya Srebulaev Dynamic Corporae Financial Policies: Nov

17 Capial Srucure and Marke Timing A varian of Myers and Majluf (1984): 1. Like Myers-Majluf: Managers have he incenive o ry o ime he marke because hey care more abou exising shareholders. Invesors reac o financing decisions, and his adverse selecion dominaes oher consideraions, so here is no opimal capial srucure. 2. Unlike Myers-Majluf: Managers hink ha hey can successfully ime he marke, believing Shares are occasionally under- or overvalued. Invesors underreac o new issues.

18 Implicaion: Temporary flucuaions in marke value have a lasing impac on capial srucure oucomes Managers respond o he (over-/under-)valuaion of heir firm wih securiy issuances. Such marke iming have a persisen effec. Appeal of marke iming Inuiive Persisence no easy o explain wih exising heory

19 Empirical Approach (Baker and Wurgler (2002): Marke Timing and Capial Srucure) Trace he evoluion of capial srucure as firms maure Sar from IPO dae. Trace he deerminans of capial srucure as firms maure. Capial srucure; focus on leverage D /A Compusa coverage from 1969 o 1998 Sep 1: analyze year-by-year capial srucure and documen he link beween marke value and financing decisions Sep 2: analyze (persisen?) effec of marke valuaions on capial srucure over ime

20 Capial srucure changes: Δ = Δ A A E A RE A e A D e = ne equiy issue = new issues minues repurchases = change in book equiy change in balance shee reained earnings = acive change in capial srucure Newly reained earnings = passive change in capial srucure Residual change in asses ==> Break he equiy-o-asses raio ino wo main componens: Ne new equiy issues and reained earnings

21 Sep 1: Deerminans of capial srucure changes M/B (marke iming heory) Oher variables from Rajan and Zingales (deerminans of capial srucure in several counries): Fixed asses inensiy (defined as PPE/A - more collaeral, more deb capaciy under a radeoff heory of capial srucure) Profiabiliy (defined as EBITDA/A - more inernal funds, less deb needed under a pecking-order heory or, more free cash flow and so more deb required under a radeoff heory) Size (log sales large firms may be more sable, less likely o ener financial disress, so more deb under a radeoff heory)

22 Resuls (a and f no repored):

23 Sep 2: Persisence of M/B effec on capial srucure Empirical approach: Summarize he hisorical pah of valuaions wih a single saisic Main measure: weighed average marke-o-book raio Definiion weighs = amoun of exernal finance (deb plus equiy) raised in each year from he IPO hrough -1 Idea: Financing evens represen a pracical opporuniy o change capial srucure Noe: resriced o non-negaive exernal finance, i.e., negaive values excluded from weighing scheme. (Eliminaes he possibiliy of a negaive overall weighed average and a negaive denominaor.) M 1 s = s = 0 1 B efwa, 1 r = 0 e + e r d s + d r M B s

24 An example of he weighing scheme (sample firm): Year E/A F/A F max(f,0) M/B M/B efwa 1986 (IPO) Explanaion: Firm wen public in (Capial srucure in nex 2 columns.) Use dollar amouns of exernal finance (3rd column) o weigh he pas marke-o-book, wih he excepion of negaive values (runcae a 0). The fifh column shows he paern of marke-o-book, he las column shows he weighed average.

25 Cross-secion regressions: efwa u B M c B M b a A D = x 1 k 1 efwa IPO pre u B M c B M b a A D A D = x 1 k 1 Conrols x include Fixed asses inensiy Profiabiliy Firm size Also, FF (2000) conrols

26

27 Economic and saisical significance of efwa-m/b effec Since he sandard deviaion of he weighed average is ypically abou 1, you can inerpre hese coefficiens as 1 STD effecs on capial srucure. In oher words, a one sandard deviaion in he weighed average, holding everyhing else consan, changes capial srucure by abou 10 percen for firms 10 years ou. Including he weighed average doubles he R-squared from 10 o 20 percen. Conclusion: Firms issue equiy when heir marke value is relaively high. They do no rebalance subsequenly. Pu differenly, emporary flucuaions in marke value have a lasing impac on capial srucure.

28 Possible Explanaions 1. Trade-off heories Taxes, coss of financial disress, and agency lead o an opimal leverage raio Marke-o-book could be conneced o Cosly financial disress Deb overhang Agency Perhaps ax benefis = Ancillary predicion: Temporary flucuaions in marke-o-book (or anyhing else) should have a emporary impac. = Refued.

29 2. Pecking order Adverse selecion dominaes oher consideraions, leading o a pecking order High marke-o-book means invesmen opporuniies exceed inernally generaed funds and deb capaciy. = Ancillary predicion: Temporary increases in marke-o-book should lead o lower cash balances or higher fuure invesmen. = Ancillary finding: Increases in M/B have a permanen negaive impac on cash balances and no lasing impac on invesmen. = Refued. 3. Marke iming Managers believe hey can ime he marke.

30 Flipside of Marke Timing: Biased managers face raional invesors = Remember I/CF discussion: relucance o issue equiy if perceived o be undervalued in he marke.

31 Readings for nex week: * Loughran, Tim and Jay Rier (1995), The New Issues Puzzle, Journal of Finance 50, pp Fama, Eugene and Kenneh French (2005), Financing Decisions: Who Issues Sock? Journal of Financial Economics 76: Asquih, Paul and David Mullins (1986), Equiy Issues and Offering Diluion, Journal of Financial Economics 15, pp * Jay Rier (1991), The long-run performance of iniial public offerings, Journal of Finance 42, pp