Taxes and Financing Decisions
Jonathan Lewellen & Katharina Lewellen
Overview
Taxes and corporate decisions
What are the tax effects of capital structure choices?
How do taxes affect the cost of capital?
How do taxes affect payout decisions?
How do taxes affect firms’ real investment decisions?
2
Trade-off theory
Firm value
VU + tax shields
VU + tax shields –
distress costs
Leverage
3
Trade-off theory
Firm value
VU + tax shields
Debt vs. equity: (1 – d) vs. (1 – c)(1 – e)
VU + tax shields –
distress costs
Leverage
4
Trade-off theory
Firm value
VU + tax shields
Debt vs. equity: (1 – d) vs. (1 – c)(1 – e)
VU + tax shields –
distress costs
Target capital
structure
Leverage
5
Main argument
Internal equity is cheaper than external equity
6
Main argument
Internal equity is cheaper than external equity
Cash distributions trigger personal taxes
Tax deferral benefit of retained earnings helps offset the tax
disadvantage of equity
7
Main argument
Internal equity is cheaper than external equity
Cash distributions trigger personal taxes
Tax deferral benefit of retained earnings helps offset the tax
disadvantage of equity
Our goal
Quantify this effect
Study the impact on capital structure, payout policy, and the cost
of capital
8
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
9
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Internal equity has tax benefits if c < i
10
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Internal equity has tax benefits if c < i
Trade-off: accelerate taxes vs. double taxation
11
Example
This paper
Clarify and generalize this idea (the example makes strong implicit
assumptions)
Miller (1977): (1 – c) (1 – e) > (1 – i)?
Understand the implications for a firm’s capital structure, payout
policy, and cost of capital
12
Overview
Clarify the literature
Capital structure
Miller (1977), Hennessy and Whited (2004)
Dividend taxes
King (1974), Auerbach (1979), Poterba and Summers (1985)
13
Outline
Simple model with two periods
Discuss the literature
Implications for corporate behavior
Empirical estimates of the tax costs of equity
14
Model
Study tax effects
No agency conflicts, information asymmetries, or distress costs
t=0
t=1
t=2
Investment opportunity, Y
Project pays Y + P1
Liquidation
Raise D0, S0
Repay debt
Cash: C0 = D0 + S0 – Y
Equity distribution, 1
Raise D1, S1
Invest cash in riskless asset
15
Model
Study tax effects
No agency conflicts, information asymmetries, or distress costs
t=0
t=1
t=2
Investment opportunity, Y
Project pays Y + P1
Liquidation
Raise D0, S0
Repay debt
Cash: C0 = D0 + S0 – Y
Equity distribution, 1
Raise D1, S1
Invest cash in riskless asset
16
Assumptions
Debt is short term
Risk neutral investors, interest rate = r
Project return P1 > Y r
[no bankruptcy]
Taxes
Corporate tax rate is c, personal tax rates are i, dv, cg
Capital gains taxed on realization (trading at t = 1 exogenous)
Liquidating dividends not taxed on portion that represents capital
repayment
17
Assumptions
Debt is short term
Risk neutral investors, interest rate = r
Project return P1 > Y r
[no bankruptcy]
Taxes
Corporate tax rate is c, personal tax rates are i, dv, cg
Capital gains taxed on realization (trading at t = 1 exogenous)
Liquidating dividends not taxed on portion that represents capital
repayment
18
Assumptions
Classic tax system
Corporate profits, after interest, taxed at c
Personal income taxed at i, dv, cg
Imputation system
Personal tax credit for corporate taxes already paid on
dividends
Effectively: dv = 1 – (1 – i) / (1 – c)
19
Assumptions
Debt is short term
Risk neutral investors, interest rate = r
Project return P1 > Y r
[no bankruptcy]
Taxes
Corporate tax rate is c, personal tax rates are i, dv, cg
Capital gains taxed on realization (trading at t = 1 exogenous)
Liquidating dividends not taxed on portion that represents capital
repayment
20
Model
Exogenous trading
t=0
t=1
t=2
Investors:
Trade  of their shares
Realize gains of  (V1 – S0)
Tax basis = (1 – ) S0 +  V1
21
Model
Taxes
t=0
No taxes
t=1
t=2
Corporate tax
Corporate tax
Capital gains tax on a
fraction of shares
Personal tax on
dividends / liquidating
repurchase
Personal tax on
dividends / repurchases
22
Tax effects
Debt financing
New external equity
Internal equity / retained earnings
23
Cashflows
Firm’s cashflows
Arrival to date 1: C1 = Y + P1 (1 – c) + (C0 – D0) [1 + r (1 – c)]
Exit from date 1: C1 = C1 + D1 + S1 – 1
Arrival to date 2: C2 = (C1 – D1) [1 + r (1 – c)]
24
Cashflows
Firm’s cashflows
Arrival to date 1: C1 = Y + P1 (1 – c) + (C0 – D0) [1 + r (1 – c)]
Exit from date 1: C1 = C1 + D1 + S1 – 1
Arrival to date 2: C2 = (C1 – D1) [1 + r (1 – c)]
25
Transactions at date 1
Proposition
Issuing debt to hold as cash (i.e., to invest in the riskfree asset)
has no effect on value, regardless of tax rates
26
Transactions at date 1
Proposition
Issuing debt to hold as cash (i.e., to invest in the riskfree asset)
has no effect on value, regardless of tax rates
Implications
Debt does not create value, via interest tax shields; only important
if it changes equity
Using cash to pay down debt doesn’t affect value either
In transactions with equityholders, it doesn’t matter where cash
comes from or goes to
27
Transactions at date 1
Proposition
Issuing debt to hold as cash (i.e., to invest in the riskfree asset)
has no effect on value, regardless of tax rates
Implications
Debt does not create value, via interest tax shields; only important
if it changes equity
Using cash to pay down debt doesn’t affect value either
In transactions with equityholders, it doesn’t matter where cash
comes from or goes to
28
Equity financing at date 1
External equity
Raise S1 at date 1
Shareholders’ CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
29
Equity financing at date 1
External equity
Raise S1 at date 1
Shareholders’ CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
NPV
Invest S1 = $1 in the firm:
1 + r (1 – c)(1 – e)
Invest $1 outside the firm: 1 + r (1 – i)
30
Equity financing at date 1
Proposition 2 (Miller)*
If the firm uses repurchases, the tax benefit of external equity is:
PV(S1) = S1
r
[(1 – c)(1 – cg) – (1 – i)]
1  r (1  i )
If the firm uses dividends, the tax benefit of external equity is:
PV(S1) = S1
r
[(1 – c)(1 – dv) – (1 – i)]
1  r (1  i )
31
Equity financing at date 1
Internal equity: retained earnings vs. repurchases
32
Equity financing at date 1
Internal equity: retained earnings vs. repurchases
Distribute all cash at date 1
t = 1:
CF1 = C1 – cg (C1 – S0)
33
Equity financing at date 1
Internal equity: retained earnings vs. repurchases
Distribute all cash at date 1
t = 1:
CF1 = C1 – cg (C1 – S0)
Fully retain, distribute at date 2
t = 1:
CF1 = –  cg (V1 – S0)
t = 2:
CF2 = C2 – cg (C2 – TB1)
[V1 = PV(CF2)]
34
Equity financing at date 1
Proposition 3
The tax benefit of internal equity at date 1, or the PV of retained
cash vis-à-vis a share repurchase, is
PV(RE1) = RE1
r (1  cg )
1  r(1  i )  cg
[(1 – c)(1 – cg) – (1 – i)]
where  = TB1 / V1, the tax basis relative to current price when the
firm doesn’t repurchase
35
Equity financing at date 1
Proposition 3
The tax benefit of internal equity at date 1, or the PV of retained
cash vis-à-vis a share repurchase, is
PV(RE1) = RE1
r (1  cg )
1  r(1  i )  cg
[(1 – c)(1 – cg) – (1 – i)]
where  = TB1 / V1, the tax basis relative to current price when the
firm doesn’t repurchase
 determines how much tax is triggered by repurchase at t = 1
36
Equity financing at date 1
Case 1:  = 0 [ = 0, S0 = 0]
37
Equity financing at date 1
Case 1:  = 0 [ = 0, S0 = 0]
Internal equity has tax benefit (better than debt) if c < i
PV(RE1) = RE1
r (1  cg )
1  r(1  i )
[(1 – c) – (1 – i)]
38
Equity financing at date 1
Case 1:  = 0 [ = 0, S0 = 0]
Internal equity has tax benefit (better than debt) if c < i
PV(RE1) = RE1
r (1  cg )
1  r(1  i )
[(1 – c) – (1 – i)]
Trade-off
If firm distributes at date 1, shareholders get C1 (1 – cg), grows to
C1 (1 – cg) [1 + r (1 – i)]
If firm retains the cash, it grows to C1 [1 + r (1 – c)], shareholders
get C1 (1 – cg) [1 + r (1 – c)]
39
Equity financing at date 1
Case 2:  = 1 [accrual taxation,  = 1]
40
Equity financing at date 1
Case 2:  = 1 [accrual taxation,  = 1]
Internal and external equity are equivalent
PV(RE1) = RE1
 RE1
r (1  cg )
1  r(1  i )  cg
[(1 – c)(1 – cg) – (1 – i)]
r
[(1 – c)(1 – cg) – (1 – i)]
1  r (1  i )
41
Equity financing at date 1
Case 2:  = 1 [accrual taxation,  = 1]
Internal and external equity are equivalent
PV(RE1) = RE1
 RE1
r (1  cg )
1  r(1  i )  cg
[(1 – c)(1 – cg) – (1 – i)]
r
[(1 – c)(1 – cg) – (1 – i)]
1  r (1  i )
Intuition
Payout triggers no new taxes  no deferral benefit
Shareholders pay tax on first-period earnings regardless of payout
decision
42
Equity financing at date 1
Proposition 4: Retained earnings vs. dividends
43
Equity financing at date 1
Proposition 4: Retained earnings vs. dividends
With dividends, the tax benefit of internal equity at date 1, or the
PV of retained cash vis-à-vis dividends, is
PV(RE1) = RE1
r (1   dv )
[(1 – c)(1 – cg) – (1 – i)]
1  r(1  i )  cg
Observations
1 – dv in numerator not 1 – cg
 not  in brackets
Tax cost of dividends depends, in sign, on cg not dv
44
Equity financing at date 1
Proposition 4: Retained earnings vs. dividends
With dividends, the tax benefit of internal equity at date 1, or the
PV of retained cash vis-à-vis dividends, is
PV(RE1) = RE1
r (1   dv )
[(1 – c)(1 – cg) – (1 – i)]
1  r(1  i )  cg
Observations
1 – dv in numerator not 1 – cg
 not  in brackets
Tax cost of dividends depends, in sign, on cg not dv
45
Equity financing at date 1
Proposition 4: Retained earnings vs. dividends
With dividends, the tax benefit of internal equity at date 1, or the
PV of retained cash vis-à-vis dividends, is
PV(RE1) = RE1
r (1   dv )
[(1 – c)(1 – cg) – (1 – i)]
1  r(1  i )  cg
Observations
1 – dv in numerator not 1 – cg
 not  in brackets
Tax cost of dividends depends, in sign, on cg not dv
46
Equity financing at date 1
Proposition 4: Retained earnings vs. dividends
With dividends, the tax benefit of internal equity at date 1, or the
PV of retained cash vis-à-vis dividends, is
PV(RE1) = RE1
r (1   dv )
[(1 – c)(1 – cg) – (1 – i)]
1  r(1  i )  cg
Observations
1 – dv in numerator not 1 – cg
 not  in brackets
Tax cost of dividends depends, in sign, on cg not dv
47
Equity financing at date 1
Why is cg important?
Suppose  = 0
Example from introduction: dividend tax is a sunk cost
PV(RE1) = RE1
r (1   dv )
[(1 – c) – (1 – i)]
1  r (1  i )
48
Equity financing at date 1
Why is cg important?
Suppose  = 0
Example from introduction: dividend tax is a sunk cost
PV(RE1) = RE1
r (1   dv )
[(1 – c) – (1 – i)]
1  r (1  i )
If  > 0
Same except shareholders also pay capital gains taxes at t = 1
determined by  cg
49
Interpretation
Impact on capital structure?
Impact on payout policy?
Impact on the cost of capital and investment?
50
Capital structure
Trade-off theory
No distinction between internal and external equity
Target leverage ratio
Tax cost of equity depends on (1 – c)(1 – e) – (1 – i)
51
Capital structure
Trade-off theory
No distinction between internal and external equity
Target leverage ratio
Tax cost of equity depends on (1 – c)(1 – e) – (1 – i)
[e ambiguous; avg. of dv and cg]
52
Capital structure
Our results
Internal equity generally less costly than external equity
Equivalent only if  =  = 1 and either
1: Firms use repurchases
2: Firms use dividends and cg = dv
53
Capital structure
Our results
Internal equity generally less costly than external equity
Equivalent only if  =  = 1 and either
1: Firms use repurchases
2: Firms use dividends and cg = dv
[with dividends, internal equity is cheaper if  cg < dv]
[King, 1974; Auerbach, 1979]
54
Capital structure
Our results
Incentives to lever up smaller than often assumed
For firms with internal cash, trade-off between debt and retained
earnings, not debt and new equity
Dividends: (1 – c)(1 – cg) – (1 – i)
Repurch:
(1 – c)(1 – cg) – (1 – i)
Neither depends
on dv
55
Capital structure
Our results
Incentives to lever up smaller than often assumed
For firms with internal cash, trade-off between debt and retained
earnings, not debt and new equity
Dividends: (1 – c)(1 – cg) – (1 – i)
Repurch:
(1 – c)(1 – cg) – (1 – i)
Neither depends
on dv
Profitable firms (w/ internal cash) should have lower leverage
56
Capital structure
Our results
Incentives to lever up smaller than often assumed
For firms with internal cash, trade-off between debt and retained
earnings, not debt and new equity
Dividends: (1 – c)(1 – cg) – (1 – i)
Repurch:
(1 – c)(1 – cg) – (1 – i)
Neither depends
on dv
Profitable firms (w/ internal cash) should have lower leverage
Internal equity may have tax benefits even if firms never want
to issue new equity
57
Capital structure
Our results
No target debt ratio
Debt ratio should be a function of internal cashflows
Leverage up when the firm has a cash deficit, down when it has a
cash surplus
Pecking order?
58
Payout policy
Dividend puzzle
Form: why do firms use dividends not repurchases?
Timing: why do firms pay dividends vs. retain the cash?
dv vs. cg
59
Payout policy
Dividend puzzle
Form: why do firms use dividends not repurchases?
Timing: why do firms pay dividends vs. retain the cash?
dv vs. cg
Observation 1
Retaining good if (1 – c)(1 – cg) – (1 – i), which doesn’t depend
on dividend taxes
60
Payout policy
Dividend puzzle
Form: why do firms use dividends not repurchases?
Timing: why do firms pay dividends vs. retain the cash?
dv vs. cg
Observation 1
Retaining good if (1 – c)(1 – cg) – (1 – i), which doesn’t depend
on dividend taxes
Observation 2
Inconsistent with view that profitable firms have too little leverage
61
Cost of capital
Trade-off theory
WACC =
D
E
(1 – c) rD +
rE
V
V
62
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
63
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
64
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
65
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
Cost of internal equity doesn’t depend on dv
66
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
Cost of internal equity doesn’t depend on dv
Investment-to-cashflow sensitivity
67
Cost of capital
Our results
Cost of capital also depends on the firm’s mix of internal and
external equity
External equity
Internal equity
Dividends
Repurchases
r (1 – i) / (1 – dv)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
r (1 – i) / (1 – cg)
Cost of internal equity doesn’t depend on dv
Investment-to-cashflow sensitivity
Cost of capital  expected stock return
68
Literature
Hennessy and Whited (2004)
Dynamic model
Taxes, uncertainty, issuance costs, bankruptcy costs
69
Literature
Hennessy and Whited (2004)
Dynamic model
Taxes, uncertainty, issuance costs, bankruptcy costs
Flat tax on distributions, no other personal taxes
In essence: firms use dividends and  cg = 0
70
Literature
Hennessy and Whited (2004)
Dynamic model
Taxes, uncertainty, issuance costs, bankruptcy costs
Flat tax on distributions, no other personal taxes
In essence: firms use dividends and  cg = 0
Maximizes the tax advantage of retained earnings
Drives many of their ‘debt dynamics’
Same as example in introduction: retaining better if c < i
71
Literature
Trapped equity
Auerbach (1979), Poterba and Summers (1985)
Dividend policy is irrelevant even with taxes
72
Literature
Trapped equity
Auerbach (1979), Poterba and Summers (1985)
Dividend policy is irrelevant even with taxes
1   dv
( A t  Dt )
Equity value =
1  cg
If pay $1 today, shareholders get 1 – dv
If retain, value goes up by
1   dv
; after capital gains taxes = 1 – dv
1  cg
73
Literature
Trapped equity
Auerbach (1979), Poterba and Summers (1985)
Dividend policy is irrelevant even with taxes
1   dv
( A t  Dt )
Equity value =
1  cg
If pay $1 today, shareholders get 1 – dv
If retain, value goes up by
1   dv
; after capital gains taxes = 1 – dv
1  cg
Implication: dv doesn’t affect cost of capital or investment
74
Literature
Our results
Trapped equity only if tax rates are just right: only if internal
equity has zero tax costs
Miller-like equilibrum: (1 – c)(1 – cg) = (1 – i)
75
Literature
Our results
Trapped equity only if tax rates are just right: only if internal
equity has zero tax costs
Miller-like equilibrum: (1 – c)(1 – cg) = (1 – i)
Even if trapped equity doesn’t hold, dv does NOT affect the
cost of internal equity
If retained earnings or debt is the marginal source of funds, dv
doesn’t affect investment decisions
76
Empirical results
Estimate tax costs of equity for a large sample of U.S. firms
Tax costs depend on
Tax rates
Interest rates
Fraction of capital gains taxed each period / tax basis of shares
Perspective of representative investor
Typical tax rates
Average tax basis of all shareholders
77
Tax rates, 1966 – 2003
60%
Corporate
Dividend
Interest
LT cap gain
50%
40%
30%
20%
10%
0%
1966
1970
1974
1978
1982
1986
1990
1994
1998
2002
78
Tax basis of shares
Tax basis = average purchase price
Estimate using prices and trading volume
Proportional trading
All investors holding a stock are equally likely to sell
Different propensities to trade
Recent purchasers are more likely to sell than long-time investors
79
Trading probabilities
Ivkovic, Poterba, and Weisbenner (2004)
70%
60%
50%
40%
Cumulative prob. of sale
30%
Hazard rate (prob of a sale conditional
on holding to month t)
20%
10%
0%
1
4
7
10
13
16
19
22
25
28
31
34
Month after buying
80
Tax basis of shares
Proportional trading
Tax basis evolves as TBt = vt Pt + (1 – vt) TBt-1
Recursively substituting:
TBt =
t i
w
i t Pt i
i1
w tt i  v t i j0 (1  v t  j )
Examples
TB1 = v1 P1 + (1 – v1) TB0
TB2 = v2 P2 + (1 – v2) v1 P1 + (1 – v2)(1 – v1) TB0
:
81
Tax basis of shares
IPW hazard rates
Hazard rates, hi, imply that trading volume evolves as
v1 = h1
v2 = h1 v1 + h2 (1 – v1)
v3 = h1 v2 + h2 (1 – h1) v1 + h3 (1 – h2) (1 – v1),
Infer the fraction of shares held today that were bought last month,
the month before, and so on
Treat abnormal trading volume in three ways
Ignore completely
Scale hazard rates up and down
Scale hazard rates down, truncate volume at predicted level
82
Data
1966 – 2003
7,066 NYSE and Amex stocks on CRSP
Daily (proportional) and monthly (IPW hazard rates) data
 is annual trading volume
Truncate estimates of  and  at one
83
Trading volume and tax basis, 1966 – 2003
Pooled time series and cross section of monthly estimates
Variable
Mean
Std
Q1
Median
Q3
Trading volume
0.47
0.32
0.20
0.38
0.72
Estimates of , assuming the initial basis = .5P1
Proportional
0.85
0.18
0.72
IPW
0.77
0.22
0.59
IPW scaled
0.80
0.20
0.64
IPW truncated
0.73
0.23
0.54
0.92
0.80
0.84
0.73
1.00
1.00
1.00
1.00
Estimates of , assuming the initial basis = P1
Proportional
0.88
0.16
0.79
IPW
0.81
0.21
0.64
IPW scaled
0.85
0.18
0.72
IPW truncated
0.80
0.22
0.62
0.98
0.88
0.94
0.87
1.00
1.00
1.00
1.00
84
Tax basis, 1966 – 2003
1.05
PROP
IPW truncated
0.95
0.85
0.75
0.65
0.55
0.45
1966
1970
1974
1978
1982
1986
1990
1994
1998
2002
85
Trading volume, 1966 – 2003
0.80
0.60
0.40
0.20
0.00
1966
1970
1974
1978
1982
1986
1990
1994
1998
2002
86
Tax costs of equity
If firms use dividends
External:
r
[(1 – c)(1 –dv) – (1 – i)]
1  r (1  i )
Internal:
r (1   dv )
[(1 – c)(1 – cg) – (1 – i)]
1  r(1  i )  cg
If firms use repurchases
External:
Internal:
r
[(1 – c)(1 –cg) – (1 – i)]
1  r (1  i )
r (1  cg )
1  r(1  i )  cg
[(1 – c)(1 – cg) – (1 – i)]
87
Tax costs of equity, 1966 – 2003
Pooled time series and cross section
c = top
c = .66 top
c = .33 top
c = 0
Mean Std
Mean Std
Mean Std
Mean Std
Tax costs if firms use dividends (%)
External --2.31 1.07
Internal --1.02 0.45
-1.70
-0.42
0.82
0.30
-1.11
0.17
0.59
0.32
-0.52
0.76
0.39
0.49
Tax costs if firms use repurchases (%)
External --1.80 0.72
Internal Prop
-1.62 0.68
IPW
-1.54 0.68
IPW-scale
-1.57 0.67
IPW-trunc
-1.51 0.67
-1.05
-0.87
-0.80
-0.82
-0.76
0.39
0.38
0.39
0.38
0.38
-0.31
-0.15
-0.07
-0.10
-0.03
0.23
0.29
0.31
0.30
0.32
0.42
0.58
0.66
0.63
0.69
0.45
0.52
0.53
0.53
0.54
 est.
88
Tax costs of equity, 1966 – 2003
Pooled time series and cross section
c = top
c = .66 top
c = .33 top
c = 0
Mean Std
Mean Std
Mean Std
Mean Std
Tax costs if firms use dividends (%)
External --2.31 1.07
Internal --1.02 0.45
-1.70
-0.42
0.82
0.30
-1.11
0.17
0.59
0.32
-0.52
0.76
0.39
0.49
Tax costs if firms use repurchases (%)
External --1.80 0.72
Internal Prop
-1.62 0.68
IPW
-1.54 0.68
IPW-scale
-1.57 0.67
IPW-trunc
-1.51 0.67
-1.05
-0.87
-0.80
-0.82
-0.76
0.39
0.38
0.39
0.38
0.38
-0.31
-0.15
-0.07
-0.10
-0.03
0.23
0.29
0.31
0.30
0.32
0.42
0.58
0.66
0.63
0.69
0.45
0.52
0.53
0.53
0.54
 est.
89
Tax costs of equity, 1966 – 2003
Pooled time series and cross section
c = top
c = .66 top
c = .33 top
c = 0
Mean Std
Mean Std
Mean Std
Mean Std
Tax costs if firms use dividends (%)
External --2.31 1.07
Internal --1.02 0.45
-1.70
-0.42
0.82
0.30
-1.11
0.17
0.59
0.32
-0.52
0.76
0.39
0.49
Tax costs if firms use repurchases (%)
External --1.80 0.72
Internal Prop
-1.62 0.68
IPW
-1.54 0.68
IPW-scale
-1.57 0.67
IPW-trunc
-1.51 0.67
-1.05
-0.87
-0.80
-0.82
-0.76
0.39
0.38
0.39
0.38
0.38
-0.31
-0.15
-0.07
-0.10
-0.03
0.23
0.29
0.31
0.30
0.32
0.42
0.58
0.66
0.63
0.69
0.45
0.52
0.53
0.53
0.54
 est.
90
Tax costs of equity, 1966 – 2003
0.0%
-0.5%
-1.0%
-1.5%
-2.0%
External equity
Internal equity
-2.5%
-3.0%
1966
1970
1974
1978
1982
1986
1990
1994
1998
2002
91
Summary
Debt vs. internal equity vs. external equity
Tax advantage of internal equity depends on capital gains
taxation
Implications for capital structure, payout policy, and cost of
capital
92
Extras …
93
Trade-off theory notes
Miller (1977) – at least partially offset at personal level
DeAngelo and Masulis (1981) – tax shields drop as leverage go up
Target capital structure / debt ratio: main focus of most empirical
tests
94
Clarify the literature
Capital structure
Miller (1977), Hennessy and Whited (2004)
Internal vs. external equity
(1 – c)(1 – e) vs. (1 – i)
Dividend taxes
King (1974), Auerbach (1979), Poterba and Summers (1985)
Retained earnings cheaper than new equity (w/ dividends)
Trapped equity view
95
Implications
Internal equity less costly than external equity
Discuss the literature
Implications for capital structure, payout policy, and cost of capital
96
Empirical results
Tax costs of equity depend on
Tax rates, interest rates
Fraction of capital gains taxed each period / tax basis of shares
Estimate for a large sample of U.S. corporations
Representative investor with typical marginal tax rates
Tax basis equal to average tax basis of all investors
97
Equity financing at date 1
Internal equity (retained earnings vs. repurchases)
Distribute all cash at date 1
t = 1:
CF1 = C1 (1 – cg) + cg S0
Fully retain, distribute at date 2
t = 1:
CF1 = –  cg (V1 – S0),
V1 = PV(CF2)
t = 2:
CF2 = C2 (1 – cg) + cg TB1,
TB1 = (1 – ) S0 +  V1
98
Capital structure literature
Miller (1977)
No distinction between internal and external equity
Tax cost of equity depends on (1 – c)(1 – e) – (1 – i)
99
Capital structure literature
Miller (1977)
No distinction between internal and external equity
Tax cost of equity depends on (1 – c)(1 – e) – (1 – i)
First statement generally false (unless  = 1 and either firms use
share repurchases or cg = dv)
Second statement incomplete and unclear about e
100
Implications
Main result: internal equity less costly than external equity
Impact on capital structure?
Impact on payout policy?
Impact on the cost of capital and investment?
101
Capital structure
Our results
Internal equity generally less costly than external equity
If firms use repurchases, holds if ,  < 1
If firms use dividends, holds if  cg < dv
Incentives to lever up smaller than typically assumed, if firms
at least partially face a trade-off between debt and retained
earnings
Tax cost of internal equity
102
Equity financing at date 1
External equity
Firm’s cash at date 2: C2 = (C1 – D1) [1 + r (1 – c)]
C2 = (C1 + S1 – 1) [1 + r (1 – c)]
Shareholders’ cashflow: CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
Benchmark: S1 [1 + r (1 – i)]
Difference = tax costs of equity
103
Equity financing at date 1
Firm’s cashflow
C2 = (C1 – D1) [1 + r (1 – c)]
C2 = (C1 + S1 – 1) [1 + r (1 – c)]
Shareholders’ cashflow from external equity
After-tax CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
Benchmark: S1 [1 + r (1 – i)]
104
Payout policy
Example from introduction
Firm has $1
Distribute now: investors have (1 – e) [1 + r (1 – i)]
Distribute next year: investors have (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Ignores capital gains taxes (valid only if firms use dividends
and either  or cg is zero)
Implicit in Hennessy and Whited (2004); their assumptions
maximize the tax advantage of retaining cash, driving many
of their key ‘debt dynamics’
105
Cashflows
Firm’s cashflows
On arrival to date 1: C1 = I + P1 (1 – c) + (C0 – D0) [1 + r (1 – c)]
C1 = 1 + S0 [1 + r (1 – c)]
On exit from date 1: C1 = C1 + D1 + S1 – 1  0
On arrival to date 2: C2 = (C1 – D1) [1 + r (1 – c)]
106
Equity financing at date 1
External equity
Raise S1 at date 1
Shareholders’ CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
Benchmark: S1 [1 + r (1 – i)]
Difference = tax costs of equity
107
Equity financing at date 1
Internal equity: retained earnings vs. repurchases
Distribute all cash at date 1
t = 1:
CF1 = S0 + (C1 – S0)(1 – cg)
Fully retain, distribute at date 2
t = 1:
t = 2:
CF1 = –  cg (V1 – S0)
CF2 = C2 + (C2 – TB1)(1 – cg)
V1 = PV(CF2)
108
Equity financing at date 1
Internal equity: retained earnings vs. dividends (< total NI)
Pay dividend at date 1
t = 1:
CF1 = NI1 (1 – dv) –  cg (V1,div – S0)
t = 2:
CF2 = S0 [1 + r (1 – c)(1 – dv)] +  cg (V1,div – S0)
Retain until date 2
t = 1:
CF1 = –  cg (V1,retain – S0)
t = 2:
CF2 = NI1 (1 – dv) [1 + r (1 – c)] + S0 [1 + r (1 – c)(1 – dv)]
+  cg (V1,retain – S0)
109
Financing decisions at date 1
Issue debt
Issue equity
Retained earnings vs. distribute cash
110
Equity financing at date 1
External equity
Raise S1 at date 1
Shareholders’ CF2 = 2 + S1 [1 + r (1 – c)(1 – e)]
e is either cg or dv
NPV of external equity
1  r(1  c )(1  e )
PV(S1) = –S1 + S1
1  r(1  i )
111
Equity financing at date 1
Proposition 2 (Miller)
If the firm uses repurchases, the tax benefit of external equity at
date 1, or the NPV of an equity issue, is
PV(S1) = S1
r
[(1 – c)(1 – cg) – (1 – i)]
1  r (1  i )
Invest S1 = $1 in the firm: 1 + r (1 – c)(1 – cg)
Invest $1 outside the firm: 1 + r (1 – i)
112
Equity financing at date 1
Why is cg important?
Suppose  = 0
Distribute: investors get C1 (1 – dv), grows to C1 (1 – dv) [1 + r(1 – i)]
Retain: cash grows in firm, investors get C1 (1 – dv) [1 + r(1 – c)]
Retain vs. pay: (1 – c) vs. (1 – i)
113
Literature
Trapped equity
Auerbach (1979), Poterba and Summers (1985)
Dividend policy is irrelevant even with taxes
1   dv
Equity value =
1  cg
D t i
i(1  )i
If pay $1 today, shareholders get 1 – dv
If retain, value goes up by
1   dv
; after capital gains taxes = 1 – dv
1  cg
114
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Internal equity has tax benefits if c < i
Miller: equity better if (1  c)(1  e) > (1  i)
Trade-off here: accelerate taxes vs. double taxation
115
Equity financing at date 1
Why is cg important?
Suppose  = 0, c = i
Example from introduction: dividend tax is a sunk cost
PV(RE1) = 0
If  > 0
Same except shareholders also pay capital gains taxes at t = 1
determined by  cg
116
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Internal equity has tax benefits if c < i
Trade-off: accelerate taxes vs. double taxation
117
Example
Firm has $1
Distribute now
Investors get (1 – e), grows to (1 – e) [1 + r (1 – i)]
Distribute next year
Grows to 1 + r (1 – c), investors get (1 – e) [1 + r (1 – c)]
Retaining better if c < i
Internal equity has tax benefits if c < i
Trade-off: accelerate taxes vs. double taxation
118