a growth type explanation for capital structure persistence
DESCRIPTION
A Growth Type Explanation for Capital Structure Persistence. Xueping Wu and Chau Kin Au Yeung City University of Hong Kong. Presented at NTU on December 11, 2008, Taiwa n. Motivation. A: Unobserved firm heterogeneities. E xisting theories Our main idea A Growth Type S tory - PowerPoint PPT PresentationTRANSCRIPT
A Growth Type Explanationfor Capital Structure Persistence
Xueping Wu and Chau Kin Au YeungCity University of Hong Kong
Presented at NTU on December 11, 2008, Taiwan
Motivation
• A: Unobserved firm heterogeneities
Road Map1. Existing theories
2. Our main idea A Growth Type Story
3. Empirical: Persistence across growth types leverage ratios important firm fundamentals financing activities
1. Our view about market timing
2. Horse race: growth type vs. market timing
4. Conclusion
Literature• Competing Capital Structure Theories
1. Tradeoff theory
2. Pecking order
3. The market timing argument of Baker and Wurgler (2002)
Can they explain leverage persistence?
1. Tradeoff Theory• Focus on costs & benefits of debt finance Target leverage
Graham and Harvey (2001): Yes in survey
Tradeoff force = adjustment toward target (optimum)
An increase in debt has
Benefits: Tax shield, Disciplining role (for FCF)
Costs:Assets substitution (opportunism), Financial distress, Debt overhang
Tradeoff Theory (Problems)• Persistence = Target leverage? Conceptually, stay where you are ≠ at the target
• Tradeoff force? Titman & Wessels(1988); Rajan & Zingales(1995):
Profitability↑ leverage↓ (the big scar!) Graham (2000): Debt conservatism Fama and French (2002): “Snail” speed of adj. Chang & Dasgupta (2007): Mechanical Mean
Reversion
2. The Myers (1984) Pecking Order• Focus on costs of equity finance Myers and Majluf (1984): adverse selection effect Financial slack is valuable, weakening tradeoff force
(explaining the big scar in tradeoff theory.)
Myers (1984): Pecking order in financing Retained earnings < Debt < Equity (as the last
resort) Asymmetric Information costs
Slow adj., no target (persistently kept away)
The Pecking Order (A Problem)• Equity as last resort? Small growth firms issue a lot of new equity
Rajan & Zingales(1995), Fama and French (2002) and Frank and Goyal (2003)
Why no fear for the AI/adverse selection effect?
The last resort puzzle (the deep wound!)
3. Market Timing • Baker and Wurgler (2002) Capital structure is a result of past attempts to time
the market (especially via equity issues) Market timing is driven by overvaluations (cheap
equity, Stein, 1996).
So equity is not necessarily the last resort Like in Myers (1984): slow adj., no target
3. Market Timing (Debate) • Leary and Roberts (2005): Clustered quick rebalance
• Hovakimian (2005); Kayhan and Titman (2006): Wrong interpretation by Baker and Wurgler
• Lemmon, Roberts & Zender (2008): Leverage persistence: initially determined
• But, Huang & Ritter (2007): Address cheap equity!
Behavior finance: Overvaluation = Cheap new equity
This Paper’s View• MM Theorem:
If capital markets are imperfect, capital structure matters.
• Persistent agency conflicts and AI types across firms (persistent relative agency/AI costs)
Persistently distinct market imperfections across firms
Persistently distinct financing behavior (and capital structure) in response
• A type-determination story for leverage persistence
An Analogy of Type Determination• Baby » Adult:
We observe persistent characteristics, e.g., heavy vs. thin
• Gene determination: diet behavior is gene-driven A heavy kid/adult always eats a lot.
A thin kid/adult cannot eat much.
• The analogy: Gene = Persistent type behind leverage Diet behavior = Financing (incl. market timing)
Asset Type and Asymmetric Info Type(1) Asset Type (Myers, 1977): Tangibles vs. Intangibles(2) AI Type (correlated with Asset Type in reality):
Myers and Majluf (1984)
Type A: more AI about AIP than about growth Type B: more AI about growth than about AIP
Asymmetric Information (AI) arises from
Assets-in-place (AIP) Growth (NPVGO)
AI Types and Financing Patterns
The generalized Myers-Majluf ModelCooney & Kalay (1993) and Wu & Wang (2005)
Dominant Asymmetric Information about AIP
Dominant Asymmetric Information about Growth
The Classic MyMj Framework:
1. Separation of Overvalued Firms from Undervalued Firms at New Equity Issues
2. AI Can Inhibit New Equity Issues Due to Adverse Selection
3. Myers’(1984) Pecking Order in Financing
Generalized MyMj Framework:
1. Not All Growth-oriented Issuers Are Lemons Because Undervalued Firms May Issue. (Pooling)2. An Increase in Asymmetric Information Can Facilitate New Equity Issues.3. Equity Issuers Not Necessarily under Duress
Growth Type and Cost of Capital
A description of agency and asymmetric information costs based on Myers (1977) and the generalized MyMj model
Firm Characteristic Dominant Type of
Asymmetric Information
(AI)
Growth Type
Cost Structure of External
Finance
Market-to-book
Asset Tangibility Debt Equity
Low HighAI about Assets-in-
place Low Growth
(G1) Low High
High High Mixed Mixed Growth (G2)
↓ ↑Low Low Mixed Mixed Growth
(G2)
High Low AI about Growth
High Growth (G3) High Low
A Growth Type Story for Leverage Persistence 1. Lemmon, Roberts & Zender (2008) Initially determined (implying a type story)
2. This paper suggests Growth type gives rise to persistently distinct market
imperfections across firms Growth type preset persistently distinct financing
behavior in responseThus: Growth type explains leverage persistence
parsimoniously. Market Timing is a growth-type driven phenomenon
Test Design: Initial Growth Type
• CompuStat US non-utility & non-financials
• 1971-2005
• Initial value: average of event year 0, 1, 2.
• With breakpoints at medians, two way independent sort on initial market-to-book and initial tangibility
Low High
Low G2 G1(Low Growth)
High G3(High Growth)
G2
Initial Tangibility
Initial
Mar
ket-
to-b
ook
Fig. 1A: Group Mean Book Leverage with the IPO Sample (1971 – 2005)
0.00
0.10
0.20
0.30
0.40
0.50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
s
G1 G2 G3
Fig. 1B: Group Mean Market Leverage with the IPO Sample (1971 – 2005)
0.00
0.10
0.20
0.30
0.40
0.50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
s
G1 G2 G3
Fig. 1C: Group Mean Book Leverage with the Full Sample (1971 – 2005)
0.00
0.10
0.20
0.30
0.40
0.50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
s
G1 G2 G3
Fig. 1D: Group Mean Market Leverage with the Full Sample (1971 – 2005)
0.00
0.10
0.20
0.30
0.40
0.50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
s
G1 G2 G3
Table 3: Long Lived Effects of Initial Growth Type -- Event Time Cross-sectional OLS
Initial MtB overwhelms updated MtB in the long run. Updated Tang overwhelms initial Tang eventually.
Event
Year
t
Firm
Obs.
Initial
MtB
Initial
Tang MtB t-1 Tang t-1 Profit t-1 LnSize t-1
Ind_Median t-1 DivPayer t-1 Adj. R2
Panel A: Book Leverage
5 7,360 -0.0117 0.2200 -0.1195 0.0203 0.5766 -0.1030 0.191
(-8.49) (18.55) (-8.77) (14.65) (16.21) (-21.07)
-0.0084 0.0505 -0.0080 0.1854 -0.1179 0.0204 0.5214 -0.1058 0.207
T-value (-5.63) (2.31) (-5.49) (8.98) (-8.89) (14.78) (14.55) (-21.82)
10 4,170 -0.0116 0.1570 -0.1510 0.0186 0.4660 -0.0761 0.138
(-6.09) (9.46) (-6.00) (11.03) (9.11) (-11.76)
-0.0108 -0.0188 -0.0052 0.2306 -0.1577 0.0193 0.3880 -0.0821 0.171
(-5.63) (-0.82) (-3.25) (10.84) (-6.09) (11.61) (7.46) (-12.84)
15 2,354 -0.0136 0.0943 -0.1509 0.0191 0.4974 -0.0739 0.134
(-6.07) (4.16) (-4.23) (8.90) (8.37) (-8.57)
-0.0122 -0.0623 -0.0064 0.2438 -0.1672 0.0194 0.3991 -0.0780 0.181
(-5.58) (-2.24) (-2.57) (9.72) (-4.74) (9.38) (6.80) (-9.37)
20 1,448 -0.0117 0.1197 -0.1832 0.0223 0.3368 -0.0929 0.135
(-2.50) (4.05) (-2.65) (8.18) (4.91) (-7.71)
-0.0090 0.0166 -0.0170 0.1467 -0.1614 0.0233 0.2469 -0.0964 0.161
(-2.00) (0.46) (-3.71) (4.60) (-2.14) (8.76) (3.46) (-8.09)
Table 3: Long Lived Effects of Initial Growth Type (cont’d)
Initial growth type (especially via initial MtB) has a strong long lasting effect!
Event
Year
t
Firm
Obs.
Initial
MtB
Initial
Tang MtB t-1 Tang t-1 Profit t-1 LnSize t-1
Ind_Median t-1 DivPayer t-1 Adj. R2
Panel B: Market Leverage
5 7,360 -0.0344 0.2584 -0.1790 0.0253 0.5016 -0.1279 0.287
(-19.95) (18.57) (-12.03) (15.64) (22.03) (-19.82)
-0.0241 0.0728 -0.0262 0.1955 -0.1859 0.0248 0.4243 -0.1271 0.316
(-13.21) (2.98) (-8.14) (8.38) (-13.02) (15.43) (17.96) (-20.04)
10 4,170 -0.0307 0.1994 -0.2545 0.0231 0.4939 -0.0792 0.247
(-13.21) (10.38) (-7.52) (11.49) (16.96) (-9.73)
-0.0258 -0.0191 -0.0257 0.2706 -0.2739 0.0233 0.3962 -0.0837 0.298
(-6.52) (-0.73) (-1.82) (10.94) (-7.67) (11.56) (10.15) (-10.64)
15 2,354 -0.0287 0.1344 -0.2718 0.0187 0.4460 -0.0927 0.201
(-11.39) (4.99) (-6.98) (7.51) (9.66) (-9.38)
-0.0236 -0.0397 -0.0230 0.2545 -0.2970 0.0190 0.3337 -0.0961 0.259
(-5.97) (-1.22) (-1.75) (8.81) (-7.86) (7.92) (6.86) (-10.22)
20 1,448 -0.0256 0.1367 -0.4572 0.0195 0.4359 -0.1133 0.219
(-7.84) (4.04) (-5.78) (6.05) (8.86) (-8.10)
-0.0179 0.0143 -0.0474 0.1514 -0.3787 0.0207 0.3140 -0.1171 0.271
(-4.77) (0.35) (-5.31) (4.18) (-5.27) (6.67) (6.19) (-8.70)
Table 4: How Stable is GT with Updated Tangibility?
The dominant diagonal effect indicates stable growth type.
Transition Matrix from Initial Growth Type, IGT, to GTt
Growth Type with Initial Tangibility replaced by Updated Tangibility
in Event Year t (GTt)
G1 G2 G3
Panel A: Aggregate Average
IGTG1 97% 3% 0%G2 5% 94% 1%G3 0% 5% 95%
Panel B: Simple average
IGTG1 94% 6% 0%G2 3% 94% 3%G3 0% 2% 98%
Fig. 2A: Group Mean Book Leverage with Full Sample in Calendar Time
0.00
0.10
0.20
0.30
0.40
0.50
0.60
1971 1975 1979 1983 1987 1991 1995 1999 2003
Lev
erag
e R
atio
s
G1 G2 G3
Fig. 2B: Group Mean Market Leverage with Full Sample in Calendar Time
0.00
0.10
0.20
0.30
0.40
0.50
0.60
1971 1975 1979 1983 1987 1991 1995 1999 2003
Lev
erag
e R
atio
s
G1 G2 G3
Fig. 3A: Industry-adjusted Book Leverage with Full Sample in Event Time
-0.12
-0.08
-0.04
0.00
0.04
0.08
0.12
0.16
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
min
us I
ndus
try
Med
ian
n
G1 G2 G3
Fig. 3B: Industry-adjusted Market Leverage with Full Sample in Event Time
-0.12
-0.08
-0.04
0.00
0.04
0.08
0.12
0.16
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Lev
erag
e R
atio
min
us I
ndus
try
Med
ian
n
G1 G2 G3
Fig. 3C: Industry-adjusted Book Leverage with Full Sample in Calendar Time
-0.21
-0.18
-0.15
-0.12
-0.09
-0.06
-0.03
0.00
0.03
0.06
0.09
0.12
0.15
0.18
0.21
1971 1975 1979 1983 1987 1991 1995 1999 2003
Lev
erag
e R
atio
min
us I
ndus
try
Med
ian
n
G1 G2 G3
Fig. 3D: Industry-adjusted Market Leverage with Full Sample in Calendar Time
-0.21
-0.18
-0.15
-0.12
-0.09
-0.06
-0.03
0.00
0.03
0.06
0.09
0.12
0.15
0.18
0.21
1971 1975 1979 1983 1987 1991 1995 1999 2003
Lev
erag
e R
atio
min
us I
ndus
try
Med
ian
n
G1 G2 G3
Table 5: Significance of Differences in Leverage Between the Growth Types
Annual Average Number of Firms
Book Leverage Market Leverage
G1 – G2 G2 – G3 G1 – G2 G2 – G3
EventYear G1 G2 G3
t-valuefor
MeanDifference
p-valuefor
MedianDifference
t-valuefor
MeanDifference
p-valuefor
MedianDifference
t-valuefor
MeanDifference
p-valuefor
MedianDifference
t-valuefor
MeanDifference
p-valuefor
MedianDifference
3 to 5 2,373 2,863 2,983 12.74 0.00 24.77 0.00 30.39 0.00 37.08 0.00
6 to 10 1,716 1,884 1,744 9.27 0.00 22.48 0.00 23.08 0.00 33.20 0.00
11 to 15 1,067 1,111 875 11.18 0.00 8.75 0.00 16.71 0.00 16.69 0.00
16 to 20 658 670 441 7.43 0.00 9.51 0.00 10.08 0.00 14.78 0.00
Table 6: Persistent Firm Fundamentalsacross the Growth Types
Growth types are persistently distinct.
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year Market-to-Book Tangibility
3 to 5 0.71 1.25 2.45 -31.14 -36.62 0.67 0.52 0.31 40.07 55.33
6 to 10 0.83 1.22 2.19 -24.17 -32.66 0.64 0.54 0.33 30.60 56.61
11 to 15 0.97 1.29 2.15 -15.15 -21.13 0.60 0.52 0.35 18.16 35.58
16 to 20 1.01 1.32 2.04 -10.06 -12.37 0.58 0.50 0.35 13.42 24.71
Table 6: Persistent Firm Fundamentals (cont’d 1)
• Firm size: G1 > G2 > G3 (Gaps persist.)
• Profitability:G1 > G2 > G3 (stable bumpy)
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year LnSize Profitability
3 to 5 18.86 18.27 17.84 15.02 13.29 13.68 10.13 -1.82 12.05 27.64
6 to 10 19.15 18.59 18.16 14.82 12.61 14.07 12.21 2.19 7.51 26.71
11 to 15 19.44 19.01 18.51 8.62 10.54 12.21 11.76 5.30 1.55 14.48
16 to 20 19.70 19.42 18.83 4.29 8.44 12.37 12.09 7.16 0.82 8.59
Table 6: Persistent Firm Fundamentals (cont’d 2)
• Among profitable firms: G1<G2> G3• Among loss-making firms: G1 > G3 G3: smaller-sized with the largest dispersion in
profits skewed towards losses
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year Profitability > 0 Profitability < 0
3 to 5 14.98 15.72 15.04 -4.09 3.38 -10.49 -17.54 -27.82 5.22 11.22
6 to 10 15.37 15.94 14.98 -3.62 5.34 -10.04 -16.66 -24.74 5.50 8.00
11 to 15 13.88 14.86 14.67 -5.05 0.86 -8.09 -14.05 -23.07 4.67 6.83
16 to 20 13.72 14.67 14.77 -4.07 -0.33 -9.12 -13.23 -21.50 2.55 4.52
Table 6: Persistent Firm Fundamentals (cont’d 3)
• Annual sales growth (SGR): G1 < G2 < G3
G3 firms do deliver the highest sales growth.
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year Asset Growth Rate (%) Sales Growth Rate (%)
3 to 5 9.65 13.46 22.75 -6.07 -9.35 12.36 17.22 30.93 -5.96 -10.77
6 to 10 12.48 13.91 18.41 -2.38 -4.90 12.20 13.91 22.19 -2.93 -7.94
11 to 15 8.78 10.89 16.62 -2.80 -4.97 6.11 9.53 17.55 -4.40 -6.26
16 to 20 7.54 7.43 12.48 0.14 -4.36 7.89 8.17 13.20 -0.31 -3.90
Table 6: Persistent Firm Fundamentals (cont’d 4)
Distinct investment styles:
• Tangible Investment (Capex): G1 > G2 > G3
• Intangible investment (R&D): G1 < G2 < G3 G1 = tangible-growth type; G3 = intangible-growth type
Persistently large R&D for G3 supports its high market to book and high sales growth rate.
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year Tangible Investment or Capext/Assett-1 (%) Intangible Investment or R&Dt/Assett-1 (%)
3 to 5 7.11 7.32 5.54 -1.47 13.99 2.07 4.65 14.48 -21.16 -42.01
6 to 10 7.60 7.00 5.35 4.97 14.82 2.32 4.39 12.58 -19.27 -39.53
11 to 15 7.02 6.41 5.13 4.56 10.11 2.81 4.67 11.32 -13.20 -24.57
16 to 20 6.31 5.87 4.63 3.06 8.63 2.91 4.34 10.17 -9.52 -18.64
Table 6: Persistent Firm Fundamentals (cont’d 5)
• Cash Holdings: G1 < G2 < G3
• %Payers: G1>G2>G3
G1: low cash holdings and typical div payers
G3: high cash holdings and typical div non-payers
EventYear Pooled Mean
t-value forMean Difference Pooled Mean
t-value forMean Difference
G1 G2 G3 G1 - G2 G2 - G3 G1 G2 G3 G1 - G2 G2 - G3
Year Casht/Assett-1 (%) Percentage of Dividend Payers (%)
3 to 5 7.37 14.54 38.38 -20.68 -35.81 56.20 32.60 10.25 21.87 28.17
6 to 10 8.27 13.17 32.35 -14.52 -32.69 64.21 42.12 15.63 21.85 31.69
11 to 15 9.99 13.48 29.16 -7.85 -20.93 65.93 48.06 23.28 13.88 20.92
16 to 20 8.49 11.54 26.13 -7.29 -18.19 66.38 50.97 29.79 9.50 12.85
How about Funding Activities? Fig. 4A: Net Debt Issue by Initial Growth Type
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Net
Deb
t Iss
ue
s
G1
G2
G3
Little differences here
Fig. 4B: Net Equity Issue by Initial Growth Type
• Difference in Net Equity Issues: G1 < G2 << G3 Heavy equity finance is a growth type phenomenon.
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Net
Equ
ity I
ssue
s
G1
G2
G3
Fig. 4C: Change in Retained Earningsby Initial Growth Type
• Difference in RE Changes: G1 > G2 >> G3 (negative!) G3 firms grow largely through equity financing.
-0.12
-0.10
-0.08
-0.06
-0.04
-0.02
0.00
0.02
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Cha
nge
in R
etai
ned
Ear
ning
s
s
G1
G2
G3
For G3, heavy losses, due to the expensing of large R&D investments that pay off slowly, offset new equity issues.
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Net
Equ
ity I
ssue
s
G1
G2
G3
-0.12
-0.10
-0.08
-0.06
-0.04
-0.02
0.00
0.02
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Event Year
Chan
ge
in R
etai
ned
Ear
nin
gs
s
G1
G2
G3
Market Timing (MT)1. Traditional, rational Views (No ex ante under-/over-valuation)
2. Irrational or behavioral view MT is driven purely by market overvaluation
3. Generalized View(Rational, Type-based View)
(1) Traditional Views• Market conditions drive equity issue spikes (MT)
(e.g., Market-to-book, stock price run-ups)
(1a) A decrease in AI of AIP
(Korajczyk, Lucas, McDonald, 1990, 1993; Choe, Masulis, and Nanda, 1993)
(1b) A decrease in the adverse selection effect due to project delay ability (Lucas & McDonald, 1992)
Both are rational views but are silent about σ(growth opportunities)↑.
How about Growth Uncertainty?• Two different views on this observation:
σ(growth)↑ equity issuing costs ↓
(2) Behavioral view Stein (1996), Backer and Wurgler (2002) Exploitable market timing based on Overpricing
& dumb equity investors (Stein, 1996)
(3) The generalized MyMj view Cooney and Kalay (1993); Wu and Wang (2005) Type-based fair market timing
Growth-type-based Fair Market Timing
• Implications of market conditions (MtBt)
Type A: AI mainly about assets-in-place
MtBt ↑ AIt ↓ Equity issuet ↑
Type B: AI mainly about growth opportunities
MtBt ↑ AIt ↑ Equity issuet ↑
“Fair”=Rational expectation is imposed.
Table 7: D and E Issues by IGTFull sample 1971-2005, Pooled OLS with Firm FE
Fair market timing: during market upturn, MtBt ↑
AI about AIP ↓ (good for G1), and AI about growth ↑ (good for G3)
A growth-type-based pecking order in financing
Baker and Wurglar (2002): Low (high) leverage firms tend to raise fund when their valuations are high (low). (Untrue!)
ΔDebt/Assett ΔNet Equity/Assett
Book Value Market ValueMtBt-1 G1 0.0110 (8.72) 0.0118 (10.00) 0.0287 (13.68)
G2 0.0068 (9.75) 0.0152 (23.09) 0.0311 (26.68)
G3 0.0023 (6.83) 0.0161 (50.96) 0.0349 (62.36)Tangt-1 G1 0.0165 (2.90) 0.0158 (2.97) -0.0001 (-0.01)
G2 0.0062 (1.18) 0.0490 (9.93) 0.0312 (3.57)
G3 0.0106 (1.77) 0.0953 (17.05) 0.0714 (7.20)
Table 7 (cont’d): D and E Issues by Growth Type
• Profits: A tradeoff force for debt issue at G1 but weak at G3.
• Size: A maturity effect (external finance↓, but on diff. paths)
Taken together: Growth-type determined dynamic external finance
Profitt-1 G1 0.0745 (10.21) -0.0455 (-6.65) -0.0596 (-4.91)
G2 0.0554 (9.76) -0.0975 (-18.31) -0.1430 (-15.15)
G3 0.0167 (4.19) -0.1857 (-49.56) -0.2292 (-34.49)LnSizet-1 G1 -0.0112 (-11.12) -0.0108 (-11.39) -0.0158 (-9.43)
G2 -0.0093 (-10.19) -0.0144 (-16.74) -0.0193 (-12.66)
G3 -0.0080 (-8.49) -0.0308 (-34.67) -0.0524 (-33.31)Ind_mediant-1 G1 -0.0394 (-2.72) 0.0015 (0.11) -0.0187 (-0.78)
G2 -0.0876 (-6.29) -0.0089 (-0.68) -0.0858 (-3.71)
G3 -0.0865 (-5.24) -0.0055 (-0.36) -0.0878 (-3.20)DivPayert-1 G1 0.0253 (13.81) 0.0019 (1.09) -0.0009 (-0.30)
G2 0.0234 (11.91) 0.0014 (0.74) 0.0023 (0.70)
G3 0.0154 (5.45) 0.0041 (1.54) 0.0042 (0.88)
Firm FE Yes Yes Yes
Obs. 76,454 76,454 76,454
R2 0.180 0.512 0.533
Horse Race: Growth Type vs. MT• Baker and Wurgler’s (2002)
For each firm at t:
• es=net equity issue in event year s
• ds=net debt issue in event year s
• Two-Horse Race: (1) B&W MtBefwa, t-1 (Baker-Wurgler MT factor)
versus (2) Initial MtB efwa, t-1: (our growth-type-based variant)
The same weighting scheme but with time-varying annual M/Bs being replaced by initial MtB
It is better than the trailing average MtB, as suggested in the literature: confounded by long-term (or average) market timing
1
10, 1
0
ts s
tsefwa t s
r rr
e dM M
B Be d
Table 8: Horse Race(OLS a la Baker and Wurgler, 2002, Y=book average)
EventYear t N
InitialMtBefwa, t-1
B&WMtBefwa, t-1 MtB t-1 PPE t-1 Profit t-1 LnSale t-1
Adj. R2
Year 3 5,904 -1.81 -3.08 0.12 -0.18 3.56 0.179
(-6.68) (-10.37) (10.00) (-9.13) (20.32)
-2.16 -2.92 0.12 -0.19 3.54 0.186
(-9.23) (-10.79) (9.58) (-9.77) (20.22)
-2.36 0.29 -3.01 0.12 -0.19 3.54 0.186
(-6.40) (0.68) (-10.06) (9.57) (-9.76) (20.22)
Year 5 4,799 -2.02 -3.06 0.12 -0.24 3.05 0.177
(-7.93) (-10.32) (8.70) (-9.25) (15.71)
-2.11 -3.43 0.11 -0.24 3.05 0.188
(-11.14) (-13.38) (8.12) (-9.18) (15.80)
-2.32 0.34 -3.57 0.11 -0.24 3.06 0.188
(-7.76) (0.85) (-11.80) (8.12) (-9.16) (15.84)
It is growth type and not market timing that best explains capital structure.
EventYear t N
InitialMtBefwa, t-1
B&WMtBefwa, t-1 MtB t-1 PPE t-1 Profit t-1 LnSale t-1
Adj. R2
Year 10 2,707 -2.67 -2.72 0.07 -0.39 2.14 0.170 (-7.66) (-6.17) (4.23) (-10.10) (8.80)
-2.11 -3.68 0.07 -0.37 2.24 0.173 (-9.35) (-9.29) (3.87) (-9.53) (9.42)-1.40 -1.27 -3.13 0.07 -0.38 2.16 0.175 (-4.08) (-2.39) (-6.76) (3.86) (-9.66) (8.93)
Year 15 1,490 -1.27 -4.54 0.04 -0.40 2.33 0.145 (-2.71) (-7.35) (1.52) (-7.61) (7.15)
-1.45 -4.82 0.03 -0.39 2.36 0.153 (-4.98) (-8.24) (1.28) (-7.30) (7.46)-1.46 0.03 -4.84 0.03 -0.39 2.37 0.153 (-3.80) (0.04) (-7.47) (1.27) (-7.28) (7.35)
FM 1,790 -1.79 -1.83 0.06 -0.51 2.79 71-05 (-19.27) (-3.78) (7.27) (-9.52) (10.73)
-1.63 -2.02 0.05 -0.51 2.87 (-22.66) (-3.69) (6.45) (-9.74) (10.99) -1.65 -0.07 -1.84 0.05 -0.51 2.87 (-9.84) (-0.30) (-3.56) (6.57) (-9.67) (11.41)
ConclusionA Growth Type Story for leverage persistence:Growth type gives rise to persistently distinct market imperfections across
firms and presets persistently distinct financing behavior in response
Findings: low (G1), mixed (G2) and high growth type (G3)1. Future leverage ratios persistently differ across the three Initial growth
types—explaining leverage persistence parsimoniously!
2. The growth types are rooted in firm fundamentals:(1) G1=tangible-growth type, low sales growth and cash, high and stable
profits, div payerG3=intangible-growth type, high sales growth and cash, high uncertain profits, div non-payer
(2) Equity issues and changes in retained earnings largely offset for G3 where R&D investments are relentless
Conclusion (Cont’d)3. Market Timing is preset by growth type.
During market price upturns, new equity issues↑ for all growth types, but the reasons differ across growth types:(1) AI of Type A ↓ good for low growth type (G1)(2) AI of Type B ↑ good for high growth type (G3)
consistent with the generalized Myers-Majluf viewA growth-type-based pecking order in financing
4. A two-horse race shows It is growth type and not market timing that best explains long-run capital structure. Fair but not exploitable view on market timing is compatible with our growth type determination story.
5. In short, if managers cannot change growth type, they cannot alter long-run capital structure through market timing.
Total Debt Short-term Debt (34) + Long-term Debt (9)Market Equity Stock Price (199) * Common Shares Outstanding (54)
Asset Total assets (6)
Leverage (L)(i)Book Leverage(ii)Market Leverage
Total Debt / AssetTotal Debt / (Total Debt + Market Equity)
Market-to-Book (MtB) [Market Equity + Total Debt + Preferred Stock (10) – Deferred Tax (35)] / Asset
Tangibility (Tang) [Inventory (3) + Property, Plant and Equipment (8)] / Asset
Profitability (Profit) Operating Income before Depreciation (13) / Asset
Firm Size (LnSize) Natural log of (Asset * 1,000,000), where Asset is deflated by GDP deflator (in 2000 dollar)
Ind_median Median industrial leverage according to the Fama and French classification of 38 industries
DivPayer Dummy variable: 1 for dividend payer and 0 for non-payer
Asset Growth Rate [Assett – Assett-1] / Assett-1
Sales Growth Rate [Salest (12) – Salest-1] / Salest-1
Investment Expenditure
Capext (128) / Assett-1; R&Dt (46) / Assett-1 (Note that only R&D missing values are replaced by zero.)
Cash Holdings Casht (1) / Assett-1
Propensity to Pay (PTP), or %Payers
The percentage of dividend payers (of a firm group)
Debt The change in Total Debt, or net debt issue
Net Equity (i)Book Value(ii)Market Value
The change in net equity, or net equity issue(i)[ of Common and Preference Stock (108)– Purchase of Common and Preference Stock (115)](i)[Sharest (25) * Adjustt (27) – Sharest-1 * Adjustt-1] * [Pricet-1 (199)/ Adjustt-1 + Pricet / Adjustt ]/2
RE The change in retained earnings (36)
Thank You!
Does competition from new equity mitigate bank rent-seeking at high growth firms? Insights from Japanese Data
Xueping Wu, Piet Sercu and Jun Yao Information asymmetry
mainly about growth Debt overhang likely Monitoring and information
production needed
Limited investment opportunities
Screening needed
Better prospects and reputation
Information asymmetry mainly about assets-in-place
Equity as last resort
Bank loans Total debt
Market-to-book ratio (Growth)
Banks’ rent-seeking attempts
Competition from bonds Competition from equity
E =1.00 E =2.00 E =3.00 E =1.00 E =2.00 E =3.00
sA=
0.50 1.41 0.73 0.23 0.96 0.64 0.330.75 1.26 0.54 0.07 0.92 0.58 0.261.00 1.06 0.29 -0.21 0.83 0.48 0.171.25 0.83 -0.08 -0.66 0.74 0.35 0.001.50 0.49 -0.59 -1.33 0.62 0.16 -0.221.75 0.15 -1.15 -2.23 0.49 -0.07 -0.452.00 -0.28 -1.93 -3.38 0.31 -0.35 -0.782.25 -0.73 -2.80 -4.57 0.15 -0.63 -1.19
-2.00 -7.24 -27.82 -45.25 -1.05 -5.65 -10.10-1.00 -1.37 -10.08 -19.63 0.60 -1.84 -4.850.00 1.22 -1.59 -4.60 1.42 0.27 -0.861.00 1.10 0.24 -0.19 0.93 0.41 0.122.00 0.16 0.05 0.00 0.17 0.06 0.023.00 0.00 0.00 0.00 0.01 0.01 0.004.00 0.00 0.00 0.00 0.00 0.00 0.005.00 0.00 0.00 0.00 0.00 0.00 0.00
0.50 -0.28 -1.17 -1.33 0.12 -0.20 -0.341.25 1.41 0.73 0.21 1.02 0.79 0.372.00 1.63 1.59 0.99 1.28 1.34 1.022.75 1.51 1.97 1.44 1.24 1.58 1.453.50 1.21 1.96 1.62 1.18 1.66 1.684.25 1.67 1.82 1.73 1.15 1.72 1.745.00 0.97 1.73 1.77 1.06 1.65 1.805.75 0.92 1.63 1.69 0.92 1.63 1.76
Panel A: Announcement Return and Relative Uncertainty about Growth
Panel C: Announcement Return and Expected Growth and Uncertainty in Tandem
Panel B: Announcement Return and Expected Growth
5A
B B s
10A
0.51 wE/10,c 1.00,σ A
0.51 wE/10,c 1.00,σ 1.00,B B
0.51 wE/10,c 1.00,σ 1.00, BA s B
Wu and Wang (JCF 2005)