International and Development Economics: Working Papers
International and Development Economics: Working Papers
International and Development Economics: Working Papers
Satish Chand
02–1
http://apseg.anu.edu.au
sdf
© Satish Chand 2002
Abstract
If capital for corporate finance was available from a common global pool and at zero transaction
cost, then does after-tax arbitrage require harmonisation of income tax rates across jurisdictions?
This paper shows that the answer is in the negative. When a corporation has the choice in
deciding the fraction of income that it distributes as dividends with the remainder held for future
capitalisation, then such choice brings about arbitrage in after-tax rates of return to investors
facing a common pre-tax return but different rates of income taxes. Policy implications are drawn
from this result.
http://apseg.anu.edu.au
Satish Chand
The Australian National University
Canberra, ACT 0200
AUSTRALIA
Email: Satish.Chand@anu.edu.au
Phone: 612 6125 0773
Fax: 612 657 2886
02/20/02 11:09 AM
Abstract
If capital for corporate finance was available from a common global pool and at zero
transaction cost, then does after-tax arbitrage require harmonisation of income tax
rates across jurisdictions? This paper shows that the answer is in the negative. When
a corporation has the choice in deciding the fraction of income that it distributes as
dividends with the remainder held for future capitalisation, then such choice brings
about arbitrage in after-tax rates of return to investors facing a common pre-tax return
but different rates of income taxes. Policy implications are drawn from this result.
1
We gratefully acknowledge helpful comments from two anonymous referees of this journal, though
the usual disclaimer still applies.
1
1. Introduction
have been calls for harmonisation of tax rates across nations. It is argued that the
failure to harmonise tax rates between countries will lead to tax competition. This in
turn, it is pointed out, could have a deleterious effect on trade and growth similar to
the 1930s (see Vann 1998). One strategy suggested to bring about harmonisation of
tax rates is through the creation of a supranational taxing authority. This is unlikely in
the short–term since such action puts in question the legitimacy of the modern state.
The establishment of a World Tax Organisation may take time as suggested by the
latter, the International Monetary Fund (IMF), the Organisation for Economic
Corporation and Development (OECD), and the World Trade Organisation (WTO)
have attended to issues of taxation on an international basis. The IMF in fulfilling its
role as the lender of last resort to economies facing liquidity crises has, as part of its
rescue package, included reforms to domestic taxation. The preference by the IMF
has been for the implementation of a broad-based consumption tax. The OECD is the
defacto supranational body on broad economic policies for its members, all except
Mexico and the Czech Republic being industrial countries. The OECD takes an
active role in taxation amongst its members by providing model tax treaties and
publication of Revenue Statistics for all of its members. The aim of policy
coordination by the OECD has been the pursuance of the goal of attaining the highest
2
sustainable economic growth and employment in member countries, but this
indirect taxation only, particularly those pertaining to tariffs and excise duties. None
of the above agencies fulfil the role of a supranational taxing authority and hence the
Calls for fiscal reciprocity and harmonisation of tax rates across countries are not new
and will continue, but this is not strictly necessary for attainment of arbitrage in the
after-tax rate of return on investment in the corporate sector. 2 The analysis in this
paper shows that if domestic firms can choose the portion of income to be paid as
dividends as against the rest which is re-invested for future capitalisation, then
arbitrage in after-tax returns to investors residing in these jurisdictions. Given that the
one, extent of variation possible in corporate and personal-income tax rates is also
bounded; these bounds are wider when dividends are franked than otherwise.
Here we consider the specific case where institutional investors seek funds from their
respective jurisdictions, each with a unique set of corporate and personal income tax
rates. These institutions operate in perfectly competitive markets such that the after-
in the financial institution which then invests the pool of funds in a global company.
Let the return to this investment by the global company be constant over time. The
3
puzzle is as to whether the institutional investor can return a common after-tax return
in the face of diverse taxing rates across jurisdictions. 3 A numerical example is used
The rest of the paper is organised as follows. Section 2 presents the analytical
Section 4 provides implications for tax policy from this analysis. Conclusions and
For tractability and ease of exposition we start off with several strong assumptions,
some of these will be relaxed later. Let an investor at time 0 invest P dollars for T
years in a stock which earns an annual return of ρ. We first consider the flows at each
t: post corporate tax yield (y) from this outlay at time t can be written as
(1)
where δ is the fraction of income distributed (such that 1-δ is re-invested and 0 ≤ δ ≤
1) and τ is the rate of corporate taxation; tax payments (x) arising from the investment
are given as
2
See Miller (1902) on early work on fiscal reciprocity and Miller and Modigliani (1961) on the role of
dividend policy in valuation of shares.
3
We relax the Miller-Modgliani (MM) (1961) zero-tax assumption.
4
where µ is the marginal tax rate faced by the investor; dividends, net of tax, received
d (t ) = δ (1 − µ) y( t )
(3);
f (t ) = δτy (t )
(4).
Since the investment is made for a period of T years, the net present value (NPV) of
stocks at time 0 with a discount factor ε is given as follows: the NPV of yield is
ρP 1 + σ T σ = ρ(1 − δ )(1 − τ)
Y= [( ) − 1],
σ −ε 1+ε
(5), 4
σ being the compound growth rate of after-tax retained earnings; that for total tax paid
is
X = [τ + µ − τµ − 2δµ(1 + τ )]Y
(6);
dividends collected is
D = δ (1 − µ)Y
(7);
4
Note that capital letters are used to denote stocks of the respective flow variables in small letters.
5
F = δτ Y
(8);
and, capital gains which is equivalent to the capitalisation of all retained earnings is
R = Y − D − X = (1 − δ)(1 − τ)Y ≡ K
(9).
I = D + (1 − µ )K + F = [1 − µ − τ + τ (2δ + µ − δµ )]Y
(10),
normalising P to unity allows equation (10) to reflect the rate of return on an initial
investment. Capital gains, as shown by equation (10), are taxed at the rate of the
Several of the above equations follow from the definition of the variables on the left-
hand-side of the equality sign, we elaborate on the meaning of the important equations
only. Equation (1) shows that the flow of after-tax income at each time period (t) is
the return to accumulated capital. The tax payments in equation (2) are the sum of
personal tax applying to dividends distributed and corporate taxes paid on retained
earnings. Equation (5) shows that the net present value of after-tax income flows
depends on the gap between the after-tax returns on capital accumulation, this being
represented by the parameter σ, and the discount rate (ε); the length of investment (T)
has a crucial role in determining the stock of yield (Y). The rest of the NPV
computations are intuitive and are derived from the NPV of yield (Y).
6
It is clear from equation (10) that a dividend imputation system with full franking
does not guarantee immunity from double taxation of after-tax corporate yield (Y);
1− µ
δ= ≡δ*
2−µ
(11),
the satisfaction of the above condition can only be coincidental and particularly so if
firms choose δ to arbitrage the rate of return accruing to investors across different
taxing jurisdictions. This is the first result of this paper. A numerical example is used
in the next section to illustrate arbitrage of after-tax returns from splitting corporate
3. Numerical Simulations
(AustCo) earns a ten percent rate of return on its investment from which it distributes
residents facing the top marginal rate of income tax of 48.5 percent. 5 Furthermore,
assume that the investment is made for 10 years, the shares cost unity, and the
discount rate is five percent, the latter being the rate on ten year bonds. The
5
This rate is inclusive of medicare levy of one and a half percentage points.
7
Table 1: Parameters used for AustCo
On the basis of the parameters provided in table 1, the net present value (NPV) of
Yield Y 0.91
Tax X 0.41
Dividends D 0.19
The above parameters and rate of return are taken as the baseline case in all of the
numerical simulations that follow. We now let AustCo compete with other
8
institutional investors in a global company which provides a constant pre-tax return to
investment.
(AustCo), Japan (JapCo), and the USA (USCo). Let the residents of each of these
countries hold shares in their respective resident institutional investors only. For
institutions, this imposition will be relaxed later. Each of these investors demands a
GlobeCo
The question posed is as to whether the three institutional investors can provide a
common after-tax rate of return to shareholders in the three countries given that each
of these countries has a unique combination of corporate and personal income tax
9
rates. The only parameter of those listed in table 1 that the company has a choice over
table 3 below, is in the affirmative; the investment institutions can provide a common
after tax rate of return in the face of different income tax rates by choosing the portion
of income that is distributed as dividends. The table below shows the computations
Note: X/Y is the effective tax rate when stocks in NPV terms are used in the computation; full-franking
Given that AustCo distributes 40 percent of its income as dividends, the Australian
are able to match this return to their respective resident shareholders by distributing
against the remainder which is held for future capitalisation. Given the pre-tax return
on investment of 10 percent being higher than the discount rate of five, a lower US
personal income tax rate induces larger capital growth and hence a larger effective tax
rate when calculated in net present value terms. The far right-hand-side column of
table 3 shows that the effective tax rate when stocks of yield and tax collected in net
10
present value (NPV) terms for the US being about three percentage points higher than
We consider the extremes, the tax rates which will force the institutions to distribute
all and nil of its income, respectively, as dividends for the after-tax return reported in
table 2 above. In other words, we investigate the extremes of the tax rates from those
reported in table 1 which will give the after-tax return of 50 percent reported in tables
2 and 3 above. As a controlled experiment, we change only the tax rate on personal
income and consider the two polar cases of complete and zero franking.
Case δ τ µ τ µ
Notes: The calculated figures are given in bold italics; and the baseline parameters and after-tax return
The results in table 4 show that personal tax rates can vary widely with arbitrage in
after-tax returns; the extent of this variation is greater with, as against without
franking of corporate taxes. Personal income tax rates have an important role in
franking, a personal tax rate of 23.3 percent results in none of the corporate income
being distributed as dividends; on the other extreme a personal tax rate of 71.7 percent
induces all of the corporate income to be distributed as dividends. A personal tax rate
11
in-between these two extremes is consistent with the attainment of after-tax arbitrage
of corporate investment when complete franking is allowed. The case without any
franking is different to the extent that now a much narrower personal tax spread is
taxes paid enables wider disparities in rates of personal income taxes across
jurisdictions.
The importance of the results in table 4 is illustrated using the above calculations in
respect of income tax rates in Australia. For the parameters given in Table 1, with
franking Australia does not have to enter into harmonisation of personal tax rates with
34 of the 36 countries listed in Table 1; the exceptions being Switzerland and Hong
Kong who have personal marginal tax rates of 11.5 and 20 percent respectively. In
contrast, in the case of zero franking, tax rates in 27 of the 36 countries are not
consistent with arbitrage of after-tax returns from splitting corporate income between
We next question the plausibility of the assumptions made in this analysis. All of the
foregoing analysis has assumed that companies can choose the fraction of income
they distribute as dividends, the remainder of which is then held for capital growth
and future distribution. This is not strictly true since the dividend policy is driven by
a host of considerations (see Hakansson, 1982 and Black, 1976 on the dividend
puzzle), true to the extent that now there are several forms of equity and financing
options that enable this choice. 6 We have also assumed that firms access capital from
the global pool with the requirement that after-tax rate of return to investors is
12
equalised. This is an extreme position but the proliferation of global finance
that investors be distinguished in respect of the tax rates they face and the ensuing
this assumption to diminishing returns will bring about arbitrage in after-tax returns
There are three policy implications from the above analysis. First, with franking of
corporate taxes, there is little need for harmonisation of tax rates across jurisdictions;
there are only a few exceptions where this is necessary. Second, reductions in the top
marginal personal income tax rate will induce greater retention of corporate income
for future capitalisation; as an example, our simulations suggest that reducing top
Australia — will drop dividend distributions from 40 to less than 18 percent as well as
reduce total income tax collections but this will be at the gain of corporate savings.
Third, in a world with complete franking of corporate income, corporate tax rates
matter only at the margin and as such there is little point in harmonising these rates
across countries.
6
See Deutch and Rumble (1998) on availability of innovative financial products. Hakansson (1982)
13
5. Conclusion
This paper presents three results. First, complete franking of corporate income taxes
paid does not provide immunity from double taxation of income. Second, there is
little need for harmonisation of tax rates on personal income to bring about arbitrage
in after-tax returns; this being particularly true when franking of corporate income is
permitted. Third, a marginal tax rate on personal income being higher than on
corporate income induces lower corporate savings when the corporate sector
The analysis has assumed that statutory rates are complied with and that arbitrage in
after-tax return holds. The extent to which these impositions hold are empirical issues
requiring separate attention. This together with the theoretical propositions from the
above analysis has to be confronted with real data, this is the next stage of this
research.
and Black (1976) note that dividend policy is driven by a host of considerations, tax rates being only
one of these.
14
Table 1: Income tax rates across countries (percent)
15
References
Athukoral, P and S Chand 2000. “Trade orientation and productivity gains from
5-6.
Sydney.
Hakansson, N H 1982. “Dividend policy and valuation: Theory and tests”, The
Miller, M H and F Modigliani 1961. “Dividend policy, growth, and the valuation of
16