The Australian
Imputation Tax System
has undergone major changes over the years. Under the Simplified Tax System (STS) introduced from 1
July 2002, the franked dividend and credit incomes earned
by companies as distributions from other companies have
been formally reported and can henceforth be directly observed.
At the same time, a regime for consolidated reporting for
corporate groups was introduced. The financial year
2003 allowed for transition arrangements. The ATO
has had a lot of trouble deciding on the appropriate data
for the period 2001-2003. The past data has been revised
numerous times, both up and down.
For the period
2004-2012 company net tax payments were $486 billion.
Some $146 billion was added to the combined Franking Account
Balances (FAB) of companies for this period. This is
30% of the tax payments. The net credits issued to all
shareholders were $337 billion which represents 70% of the
company tax paid (according to the FAB data). This
is an estimate for the national access fraction. ATO
financial data also show $321 billion of credits were
distributed as fully franked dividends. Of these, $91
billion were reported as franking credit income by
companies. And of these $91 billion, an estimated $8
billion were received and redeemed by companies within the
superannuation businesses of Life Offices, leaving a net
$83 billion of credits recycled within companies. So the
net credits distributed by companies were $237 billion and
from these $148 billion have been redeemed (including the $8 billion
by Life Offices):
>> $94 billion
redeemed by personal taxpayers
>> $42 billion redeemed by super funds
>>
$4 billion redeemed by charities and other endorsed
organisations.
Some 28% of the distributed credits
($89 billion of the gross amount of $321 billion) are not
recorded after being issued. This is approximately the
proportion of all Australian equities (both public and
private) held by foreigners.
But it also includes any amount not yet filed by domestic
taxpayers.
Over the period of 2004-12, the ATO
reports that companies had taxable income of $2,008
billion and the ATO determined that they were liable
company tax of $486 billion, leaving an after-tax income
of $1,522 billion. From this amount they distributed
dividends totaling $840 billion cash and $321 billion of
credits (89.1% of dividends were franked, 10.9% were
unfranked). The aggregate ATO data imply that the
system-wide effective company tax rate was 24.2%, the
profit payout ratio as dividends was 55.2% and the
distribution fraction of tax as franking credits was
66.0%. But this effective tax rate is false because it
includes franking credits as company income but deducts
them as tax offsets before net tax is calculated. The credit income
claimed as a tax offset is 4% of total taxable
income of $2,008 billion so the average effective tax rate
in Australia is much closer to 28%.
About 62% of
distributed credits are redeemed – the redemption
proportion of net credits distributed to all shareholders
outside of companies as shareholders. This is the
national redemption fraction but it is not a reliable
estimate because ...
There is an unreconciled $100 billion of credits
difference between tax data, FAB data and financial data
for the period 2004 to 2012. The FAB data ought to
be net data as credits paid by one company to
another cause a debit in the payer's FAB and a credit in
the receiver's FAB, resulting in no net change to the
collective FAB. The dividend data ought to be
gross data as companies report their dividend
payments regardless of whether or not it is paid to
another company (outside of their group). But the
FAB data implies that a net $337 billion of
credits were paid and the dividend data is that a gross $321
billion of credits were paid. Combining these data
with the inter-corporate flow of $83 billion results in
the unexplained $100 billion of credits. We explore
possible answers to this conundrum.
What is a
robust estimate is the national average gamma of 31%.
This is based on the total company tax paid and on the
total credit utilisation by taxpayers. It is not
dependent on the interim data of how much is retained in
the FAB and how much was distributed as franked dividends. |
|
Back to Top |
The Australian Market Risk Premium has slowly changed over recent decades. We estimate the MRP
using data from 1875 to 2013 but concentrate on the data
from 1929 and for this data we estimate the MRP in three different ways. First
as historical averages of differences between market
returns and government bond yields (with various
sub-analyses such as real vs nominal and averages
calculated over time spans of different lengths), then as
the dividend discount model applied to historical dividend
yields since 1929 (making allowances for the introduction
of imputation tax) and finally as forward estimates based
on analysts' forecasts of future corporate income.
We demonstrate the simple logic that reconciles the two
approaches and indicate the required growth rate to be
used in the dividend growth version.
We find broad agreement among all
three estimates whereby the MRP gradually declined up to
2005. Analysts’ implied MRP averaged 5.9% pa prior
to the GFC which was very close to the 130 year historical
average MRP of 5.7% pa. During the GFC, implied
yields were very high and the commensurate risk premiums
were very high. After the GFC, implied risk premiums
temporarily dropped back to 7% pa and then began to climb
again. Our
latest implied risk premium estimate is 8.8% as at June 2013.
Historical averages as estimates
of the MRP have the perverse property of large and
widespread falls in stock prices due to rises in risk
aversion, such as the GFC period, cause a decrease
in the MRP estimate because the averages now include large
negative market returns. Implied estimates are
forward looking and should not suffer from this problem.
We allow for missing return estimates from the S&P/ASX indices due to credits being ignored.
We observe a very highly statistically significant "kink"
in the ASX indices pre- and post-imputation (rsq > 99%)
and estimate the ASX indices are missing returns of 54 bp per annum
due to franking credits which are priced into stocks. We look for reasons why the MRP might have changed. |
|
Back to Top |
Reducing the rate of company tax
and eliminating imputation tax.
Of the $486 billion in company tax
paid over the period 1 July 2003 to 30 June 2012, $148
billion was utilised by taxpayers as pre-payment of their
tax liabilities or redeemed. Hence, ultimately some
31% of the headline company tax collection was not company
tax but a pre-payment of other taxes as Australia has
this
partly integrated tax system between company tax and
personal tax (including retirees on a zero rate and
superannuation funds on a 15% rate). If the rate of
company tax was reduced by 31% and imputation tax
abolished, then ostensibly the net revenue to government
would be unaltered (with a whole lot of caveats about
investor behaviour). There would be major winners
and losers from this proposition. We explore this in
a two step process - first cutting the rate company tax
from 30% to 21% under the imputation tax system and then
the second step of eliminating imputation tax at this 21%
rate. In this manner we can attribute the impacts on
the various parties due to each of the two steps. We
first use Telstra Corporation Ltd results for 2014 to
demonstrate the two steps in a concrete example and to
demonstrate the various data and concepts that need
addressing. |
|
Back to Top |
| The
market average value of a credit = 51% of its face value. There are three milestones in the
life of franking credits; they are created when company
tax is paid, they are distributed along with dividends and
they are redeemed when shareholders claim them against
personal tax liabilities. When a stock goes
ex-dividend it also goes ex-franking credit. The
ex-dividend price of a fully franked stocks falls further
than an equivalent unfranked stock. This gives us a
means of estimating the market value of franking
credits. Tax statistics give us the ultimate
redemption value of a franking credit. . We estimate the
drop-off values across big cap, mid cap and small cap
stocks and for dividend yields above and below 2% per
event. The small cap results are unsurprisingly very
erratic. The big cap results give us our estimate of
51% of face value. We can see no
variation in the market value of credits over time but the
changes in the company tax rate have changed the amount of
credits available per dividend. |
|
Back to Top |
| The Australian stock market index has changed many times over the last 130 years (1875-2004) which reflects the changing composition of the Australian economy. We review the changes over time of the major sectors within the ASX (now S&P/ASX) indices. These changes are also very strongly reflected in the volatility of the indices whereby various waves of volatility have passed through the ASX indices. The last major one corresponded with the rise and decline of the resource sector within the ASX market. The decline in that volatility is still happening today in that we are still experiencing the post-boom decline in volatility. We are seeing a similar decline in GDP volatility, both in Australia and the USA. This hints at a declining future market risk premia. |
|
|
Back to Top
WACC formulae can be dangerous to use in some cases and they remain a mystery to many people. We introduce a simple logic for describing the various WACC formulae and then we put them to the test for robustness. There are a couple that should always be avoided because they consistently give under-estimates for valuation. They also give the widest spread in valuation outcomes so we strongly suggest that these particular WACC formulae not be used in valuation exercises. |
|
Back to Top |
| The cost of capital to government is the private sector's pre-tax cost. There is no cheaper capital available to government nor is there any valuation wedge between private and government ownership. Tax equivalent payments, which under government ownership take the form of cash, under private ownership become franking credits, whose market value depends on the pay-out policy of the firm concerned and its ability to distribute franking credits. Government-owned business will make correct investment decisions if they use the private sector post-tax cash flow and the private sector post-tax cost of capital. In this way, they will get consistency with their pre-tax cost of capital and their pre-tax cash flows. As most returns are observed and measured in the post-tax environment, it is preferable to do all the valuations on a post-tax basis as this will avoid making assumptions about converting observed post-tax data into unobservable pre-tax data. |
|
Back to Top |
| Errors in valuation calculations are not at all uncommon and some have very large consequences - running into hundreds of millions of dollars! This set of linked papers is a collection of errors we have seen over the years (names of individuals and events withheld). It includes a brief description of each error and example calculations. Contributions (acknowledged or anonymous) are welcome. |
|
Back to Top |
Book vs Market errors occur when a valuation Vanilla WACC valuation is conducted with the debt amount determined by the capital expense (the book value) and not the market value of the asset. These will only be the same when the NPV =0. For NPV > 0 projects, it will erroneously understate the debt amount, understate the debt tax shield and overstate the debt cover ratio.
|
|
Back to Top |
Inflation must be taken out of the nominal WACC after forming the nominal costs of capital and then applying the Fisher equation. Errors are made when forming a “real” WACC from real costs of capital. These errors seem more common in the perpetuity valuation of residual or “tail” cash flows that are often implicitly real.
|
|
Back to Top |
Tax shields on interest payments must exist before they can have any value. Usually the tax shield on debt interest is readily available and this is consistent with the classical WACC method which values the shield as if it were immediately available. But, if the tax shield on debt is not available for some years due to little or no tax being paid, then the standard WACC approach is in error.
|
|
Back to Top |
Debt interest tax shields (ITS) should be valued using the cost of equity, not the cost of debt. The tax shield on debt is a cash flow to shareholders , not debt holders. The note on this issue demonstrates that the value of the ITS is the difference of two values, both calculated using the cost of equity but in the special case of a perpetuity the difference reduces to a debt value capitalised at the cost of debt.
|
|
Back to Top |
