Memorandum submitted by Aberdeen Asset
INTRODUCTION TO SPLIT CAPITAL INVESTMENT
TRUSTS AND THE CURRENT ISSUES AFFECTING ZERO HOLDERS
Aberdeen Asset Management PLC is a leading independent
fund management group, which manages over £27.2 billion of
global assets for institutional investors and individuals, from
26 locations around the world. Aberdeen Asset Management manages
three main asset classes: equities (£14.8 billion), fixed
interest securities (£7 billion) and property (£5.4
The Group manages 44 closed end funds/investment
trusts (£6.2 billion) for financial institutions and retail
investors in Australia, Canada, United States and United Kingdom.
As UK investment trusts are listed on the London Stock Exchange,
investment trusts managed by Aberdeen also have non-UK institutional
The funds have a wide range of investment remits
investing in: specific geographic regions (eg USA, Europe), specific
sectors (eg media, technology), specific securities (eg Asian
bonds) or in a combination of these. The Company manages 25 conventional
closed end funds/investment trusts and 19 split capital investment
The main difference between investment trusts
and eg, unit trusts is pricing: the value of the shares of an
investment trust is determined by the market, influenced by the
view of market makers, trading volumes, sentiment, etc. Therefore,
investment trust shares can trade at a price which is higher or
lower than the relative value of the assets of the company.
there are currently 134 splits in
of these, there are 310 share classes
in issue from three main types of share;
total sector assets of £13.2
37 fund management splits; and
ZDP shares were introduced in 1987.
Zeros were widely considered to be a low-risk
investment until relatively recently. In April 2002, Gartmore
Monthly announced it could not redeem its zeros in full on its
Splits cannot be assessed for risk as a class,
but should be assessed upon consideration of various interrelated
factors, such as the level of gearing, the nature and volatility
of the underlying portfolio and the existence of cross-holdings.
These factors are now being considered with the benefit of hindsight.
Fund managers, investment bankers, lawyers, advisers and the listing
authorities all worked to a common assumption that some form of
equity market growth could be expected from equity markets in
each succeeding year, as had been the case for the preceding decades.
Cover and hurdle rate
The "cover" and "hurdle rate"
calculations were the standardised tools used both by analysts
and the AITC to analyse risk attaching to zeros; it was only when
the prospect arose of zeros not paying out that the use of these
calculations came to be considered in any detail. It was not until
early 2002 that the AITC agreed a standard approach to the calculation
and reporting of cover ratios. A 2001 study by Merrill Lynch found:
cover and hurdle rate are, in isolation,
wholly inappropriate measures of risk;
bank debt and increasing portfolio
volatility are major risks facing zeros;
the application of derivative pricing
in the sector is the only form of analysis that provides a rational
assessment of risk and a rational assessment of relative value;
full disclosure (on structure and
portfolio) is essential if the sector is to be fairly priced.
Barbell trusts hold two distinct portfolios
of investments: a low-yielding growth portfolio (typically financed
by debt) and a high-income portfolio (typically financed by equity).
While a good number of barbells are splits, most splits are not
barbells. It is the division of the assets side of the balance
sheet into two separate portfolios, together with a high level
of gearing, which makes a barbell.
The HSBC analysis of March/April 2000 showed
that, relative to the World Index, the splits sector was overweight
in UK equities, thus concentrating portfolio risk. The aim of
many barbells was to diversify this growth risk away from UK FTSE
350 shares into eg US equity, UK smaller companies, technology
shares and Far Eastern shares (see chart on page 12).
Disadvantages of barbells (although these could
affect any split, whether or not it followed the barbell strategy)
many barbells chose to increase gearing
through bank borrowingtherefore they were open to the same
risks as any other highly geared trust;
many barbells were launched since
2000-01, at the beginning of market decline: as a class, therefore,
barbells have reflected the recent decline in the markets generally,
particularly since September 2001; and
the nature of the recent market decline
has been to affect both the equity and bond markets. Consequently,
the perceived advantage of dividing the portfolio to protect against
adverse market conditions has not been realised.
Cross-investment refers to the ownership by
a trust of a portfolio of income shares in other splits to boost
portfolio yield. It is possible for cross-investment to occur
by trusts holding zeros in other trustshowever, this is
less common. The "fund of funds" has long been a feature
of splits, offering diversification, asset allocation and tax
advantages over a single portfolio.
In relation to cross-investment, as at 31 March
only 15.43 per cent of the sector
as a whole was invested in splits;
73.59 per cent of the sector was
invested in equities and bonds; and
a further 7.25 per cent was invested
The effect of market decline on the overall
value of portfolios has been enhanced by applying across the portfolio,
whether it is invested in equities, corporate bonds or other splits.
Allegations of collusion
It makes good investment sense for a manager
of a split to acquire high-yielding income shares, as these fall
squarely within the trust's requirements. One source of such shares
is splits issued by other managers.
The market prices of shares in investment trusts
are independently determined by market makers based upon supply
and demand, following strict rules laid down by the London Stock
Exchange. There are a number of brokerage houses which operate
in the sector and up to seven competing market makers who set
prices for shares in the secondary market. Primary prices, for
new issues, are outlined in the issuing prospectus. These prices
are in no way determined by other managers, so it is highly unlikely
that cross-investment would directly affect the pricing of individual
trusts or the sector overall.
Unlike unit trusts, investment trusts can borrow
money and invest the proceeds. Gearing is an effective management
tool to enhance gains in a rising market and, until recently,
has served to set the investment trust sector apart from other
forms of investments by providing competitive returns to investors.
With a long-term view of rising markets, this remains the case.
Gearing magnifies returns favourably or unfavourably
in rising or falling markets. Over recent months, the markets
have fallen at an unexpected rate and the effect of gearing has
compounded losses in the short to medium term. This is because,
while the percentage loss compared with the total portfolio remains
the same, the costs of borrowing must still be met out of the
remaining asset value.
1. WHAT IS
An investment trust is an investment company
whose shares are listed on the London Stock Exchange.
Generally, an investment trust works by issuing
one or more classes of share that are fully listed on the London
Stock Exchange. The investment company has no business as such
but instead the assets of the company comprise a portfolio of
investments that are actively managed under the direction of the
Board of Directors with a view to providing a return to shareholders.
Investment trust shares are regularly traded
by market makers. Investors have a number of ways of investing
directly or indirectly in an investment trust, of which the main
ones are as follows:
Buy a share as a new issue direct
from the investment trust in accordance with the terms of the
Buy a share after the time of its
issue, for example through a stockbroker.
Buy a share as advised by an adviser
or through the discretion exercised by a fund manager.
Buy a product that itself invests
in, or is based around, shares of an investment trust. Examples
of this include Investment Trust Share and Savings Schemes, and
certain unit trusts.
One of the principal differences between investment
trusts and other forms of investment, such as unit trusts, is
the pricing. With a unit trust, the value of each unit is determined
directly by reference to the total value of the underlying assets
at the time and the number of units then in issue. By contrast,
the value of the shares of an investment trust is determined by
the available market in those shares, influenced by the view of
market makers, trading volumes, sentiment and the state of the
Investment trust shares can therefore trade
at a price that is higher or lower than the relative value of
the assets of the company, having regard to the number of shares
in issue. This is particularly significant for highly geared income
shares whose price in part expresses the expected future dividend
stream offered by the share, as well as its underlying asset value.
2. WHAT IS
A split capital trust is an investment trust
company with more than one class of share (hence "split capital"),
each class having different rights to participate in income or
capital returns. Split capital investment trusts were originally
devised over 100 years ago and operate on the basis that shareholders
look for either income or capital growth and are able to meet
such requirements through choice of the different share classes.
Split capital investment trusts have a fixed life spanon
averge seven yearsat the end of which they are wound up
or rolled over into a new fund.
2.2 Substantial industry
As at December 2000, a Financial Times Supplement
reported that there were over 90 split capital or quasi split
capital trusts in issue. Capitalised at over £12 billion
and comprising some 15 per cent of the investment trust sector.
This represented an increase from £5 billion over just 12
months, while conventional trusts had shrunk by about £3
billion over the same period and were dropping by 10 per cent
These figures have risen still further and there
are currently 134 splits in existence (including umbrella funds
and highly geared funds with only one share type), with 310 share
classes in issue from three main types of share, with total sector
assets of £13.2 billion. Each split has its own independent
manager, and 37 fund management groups manage splits.
Aberdeen is one of the largest managers of investment
trusts in the UK and is the largest manager of splits. Aberdeen
manages 19 splits, all of which have different mandates, structures,
histories and investors. Each split has its own individual website
and can be accessed through Aberdeen's Investment Trust directory
2.3 Stock Market returns over 15 years
3. WHAT IS
A zero dividend preference share or "zero"
is a particular class of share within a split capital investment
trust. Its principal characteristics are:
No dividend is payable at any stage
to the zeroholder. In effect, therefore, the interests of the
zeroholder are principally in capital return rather than income
from the trust.
At the redemption date (the date
specified on the establishment of the split for its winding-up),
the zeroholder is entitled to receive a pre-established redemption
price provided that there are sufficient assets in the trust.
In determining whether there are
sufficient assets to meet the payments due to zeroholders, the
only prior call on the assets of the trust is in favour of any
bank which has lent money to the trust (see section 7 on "Gearing"
Zeros can be contrasted with the other two main
share classes, income and capital shares. As the name suggests,
income shareholders are entitled to a regular dividend payable
from the assets of the trust during its lifetime. Capital shares
do not pay dividends and do not have a predetermined value. Instead,
capital shareholders are entitled to share in the remainder of
the assets of the trust on winding-up once the prior interests
of any lenders, zeroholders and income shareholders have been
met. Consequently, the entitlements on winding-up to the assets
of the trust are in the following order:
(ii) Zero holders;
(iii) Income shareholders; and
(iv) Capital shareholders.
There are also hybrid types of shares such as
Income & Residual capital shares. Individual trusts can have
two or more classes of share in different combinations.
The first split can be charted back as early
as 1873, when a Scots lawyer, William Menzies, considered he had
two types of clients: the risk-tolerant and the risk-averse. The
Scottish American Investment Company Limited ("SAINTS")
was created with two main share classes, ordinary shares and preference
shares. For several years after 1873,the trust sector was made
up of a mixture of legal trusts (like Foreign & Colonial"F&C")
and trust companies (like SAINTS). In 1879, F&C's certificates
were converted into equal quantities of ordinary and preference
shares and adopted a geared capital structure. Some might argue,
therefore, that not only has F&C, the "grandfather of
the industry" been a split capital investment trust for much
of its life, but that the whole trust sector had a substantially
split capital sector for most of its first 60 years.
3.3 Risk profile
Split capital investment trusts have become
an increasingly popular investment trust sector; they offer higher
yields in a low interest rate environment by having different
share classes to offer basically predetermined growth, high capital
growth or high income.
Zeros are used as a quasi-fixed interest product
and until recently, have been considered to be a low-risk investment.
Research during the period supports such views. In Merrill Lynch's
research in March 1999, "Risk Rating Investment Trusts",
zeros were specifically excluded from their analysis. Their risk
category system "loses the very low risk end of the spectrum.
To be fair though, this is only populated only by quasi fixed
interest stocks such as zero coupon preference shares."
The traditional analytical measures of risks
attached to an individual zero are covered in the next section.
Zeros have been used by investors wishing to
obtain a fixed capital gain at a pre-selected date in the future,
for example, in school fees planning. For instance, the Bloomberg
Money Guide of May 2001 included an article in which Barrie Golden,
investment manager at Hargreaves Lansdown, commented on zeros:
"They are the safest split capital investment
as they are the first in line for a slice of the assets at the
winding-up date . . . In terms of risk, they are almost on a par
with building society accounts or UK Government Bonds."
Bill Fowler, senior investment strategist at
Hill Osborne & Co (stockbrokers) in his article of 29 June
1999 entitled "Zero Dividend Preference Shares" concluded:
"No other investments offer the same combination
of high returns, low risk and tax efficiency as zeros, which are
strongly recommended for consideration."
In May 2000, stockbrokers William De Broe published
an independent report as a basis for discussion on the burgeoning
split capital market. The report's author, John Newlands, who
has also written a history of the investment trust industry, accentuates
the unprecedented popularity of splits and the characteristics
of zeros as a share class.
"The zero dividend preference share, introduced
in 1987, is a success story in itself. Zeros typically offer yields
at a premium over gilts, building society and bank accounts, predetermined
repayment terms, no liability to income tax and low risk."
It has only been relatively recently that market
commentators and analysts have called into question the accepted
risk profile of zeros. This has come about largely as a reaction
to the combination of low interest rates and falling markets,
significantly since September 2001. In April 2002, Gartmore Monthly
announced it could not redeem its zeros in full on its redemption
3.4 Risk profiling of investment trust shares
Despite the endeavours of Merrill Lynch in March
1999, there has been no universal risk rating system of individual
investment trust shares to date. Private investors need to use
their own resources to investigate the profile of their investment
(or rely on analytics produced by the AITC on a monthly basis).
In the case of zero dividend preference shares,
market analysts and commentators alike used similar formulae to
determine the risk profile of individual zero shares, until recently.
These ratios are covered in the next section, however it is worth
alluding to the general market views of the risk profiles of splits
at the time in question.
In considering the risk profile of any product,
there is a fundamental decision to take: is risk related to:
the probability of an event
happening or not happening ie if we consider it extremely unlikely
that a zero will fail to pay out in full then it should be considered
to be low risk; or
the worst possible result if
that event does happen ie if a product would return all capital
in full but provide no return, then it would be viewed as low
risk, compared with a product where the entire capital could be
Kestrel Research, in its Bulletin entitled "Split
Capital Trusts and Zero Funds" dated April 2002 presented
its understanding of risk in this way:
"Myth Number Five: "Low Risk"
means "No Risk"
Consider the world of insurance. Most sensible
insurance companies only accept those risks which they are able
to evaluate. Events which are deemed to have a low (but not zero)
probability of occurring (an earthquake in Tokyo, annuity rates
falling below 10 per cent, repeated mass flooding, the destruction
of the World Trade Centre) are often under-provisioned for. The
logic goes like this: If an event has a very low probability of
occurring, or has never happened in the past, then, for all practical
purposes, we can ignore it. Wrong.
No one can ever fully quantify all the risks
associated with any financial venture for all time. Changes in
interest rates, taxation, inflation, currencies, oil prices, good
health, the birth rate, economics and confidence can all affect
the risks associated with financial assets. In attempting to assess
risks, some people will look to see how often something has happened
in the past. Others prefer to concentrate, not on the probability
of an event occurring, but, on the maximum damage that an event
could inflict, no matter how unlikely.
. . . Zeros are commonly regarded as low risk
investments because, well let's be frank here, up until recently
they always have been. Zeros still remain low risk investments
where the underlying portfolios are blue chip and, crucially,
do not have additional layers of gearing. Our industry faces a
fundamental recurring problem. Innovative structures attract money
and success breeds success. This leads to a progressive adulteration
or creep such that when market circumstances no longer make the
products attractive, certain providers will nevertheless continue
to issue increasingly inferior produce for as long as people are
prepared to buy them."
Many commentators have also come to realise
that splits cannot be assessed for risk as a class. Instead, their
risk assessment should be based on consideration of a number of
inter-related factors, such as the level of gearing, the nature
and volatility of the underlying portfolio and the existence of
cross-holdings. Such measures are however being applied with the
advantage of hindsight. Fund managers, investment bankers, lawyers,
advisers and the listing authorities alike all worked to a common
assumption that some form of equity market growth could be expected
from equity markets in each succeeding year, as had been the case
for the preceding decades.
More recently commentary from Merrill Lynch
has been exploring the weaknesses of the split capital income
share market. For example, Merrill Lynch's consideration of cross-ownership
"The market to date has apparently failed
to interpret cross ownership correctly and as a consequence, the
level of risk inherent in zero dividend shares extends from being
low (some zeros are akin to high-grade bonds) to phenomenally
high (akin to highly geared plays on equity markets). On creation,
the perception was that all zero offerings were low risk investments
(when compared with standard equity risk)."
In a second paper, Merrills summarised the position:
". . . Recent press comment has added to
the uncertainty with sentiments ranging from "zeros are not
risky because no zero has ever defaulted" to "zeros
could become the next mis-selling case". Such extremes in
pricing and opinion are undoubtedly a consequence of the diverse
investment properties offered by different zero dividend shares.
With this in mind we question the suitability of the current analysis
used to value zeros and in particular the use of capital cover
and hurdle rate as the main measurement of risk." 
As the markets have continued to fall and the
split capital investment trust sector suffer in particular as
a consequence, the structure, management and marketing of split
capital trusts has been challenged.
The remaining sections of this paper consider
in more detail some of the principal criticisms that have been
levelled against split capital trusts, in terms of:
The use of "cover"
and "hurdle rate" as exclusive measures of risk (section
The "barbell" strategy
for management of the trust's assets (section 5);
The "magic circle" of
managers and cross-investment (section 6); and
Gearing (section 7).
4. COVER AND
In recent commentary, much has been written
on the subject of "cover" and "hurdle rates"
in connection with the split capital sector. These terms relate
to calculations used to analyse risk attaching to zeros. The purpose
of this section is to consider what these terms mean and how they
have been used both historically and more recently in assessing
the risk profile of this type of investment. The use of "cover"
and "hurdle rate" have been the standardised tools used
by analysts and the AITC alike until very recently.
Cover, or the cover ratio, is a measure of how
well the redemption price of a zero is coveredie the probability
that it will be paid out in full on the redemption date.
Basically what we are trying to achieve is a
ratio of total assets of the fund to the charge of the zeros to
measure how well they are covered. This will normally come up
with a ratio where one expresses that they are exactly covered,
a number less than one that they are not fully covered and greater
than one that they are covered with some degree of surplus.
There are two measures of "cover"
in use. One looks at the impact of cover including debt and all
prior charges and relates to the specific share, the other reflects
the inverse of the hurdle rate.
4.3 Hurdle Rate
The hurdle rate, also known as fulcrum point,
measures how much the fund's assets can drop each year, over the
life of the fund and still pay out the redemption price of the
zeros in full.
It is obviously correlated with the cover ratio
and if the cover ratio is greater than one the hurdle rate will
be negative, showing that the fund's assets can drop a certain
amount each year and still be large enough to pay out in full.
Conversely, if the cover ratio is less than
one the hurdle rate will be positive, indicating that the fund's
assets will have to grow by at least that much each year to be
able to pay out the redemption price in full.
4.4 Recent consideration of cover ratios and
Since the failure by certain trusts to pay out
in full to zeroholders on redemption, many analysts have criticised
the use of cover ratios and hurdle rates as measures of risk.
For example, it has been argued by Merrill Lynch that the detailed
methodology used to calculate cover is crucial and without full
disclosure of this to, for example, the investors or their IFAs,
it is difficult to identify the level of risk associated with
"It is evident that mis-pricing in the sector
is inevitable if asset cover and/or hurdle rate is seen as the
sole representation of risk."
Merrill Lynch's conclusions were that:
Cover and hurdle rate are, in isolation,
wholly inappropriate measures of risk.
Bank debt and increasing portfolio
volatility are major risks facing zero investors.
The application of derivative pricing
in the sector is the only form of analysis that provides a rational
assessment of risk and a rational assessment of relative value.
Full disclosure (on structure and
portfolio) is essential if the sector is to be fairly priced.
Merrill Lynch also said:
"The typical ways of looking at zero dividend
shares ie, cover and hurdle rate as a measure of risk and GRY
as a measure of return are hopeless in the face of sizeable cross
ownership. This is more apparent where zeros have been issued
on markets that have since fallen in value (and so the zeros have
no capital cover)."
Merrills' new analytical model attempts to show
the shortcoming of the traditional analytical tools in assessing
the risk in zeros, particularly when uncovered.
However, zeros have been around since 1987 and
until April 2002 all of them paid out in full on the redemption
date. Consequently, questions over cover and hurdle rate have
not been relied upon heavily by investors or IFAs in selecting
zeros above other forms of investment. For example, Bill Fowler,
Senior Investment Strategist at Hill Osborne & Co (stockbrokers)
wrote in June 1999:
". . . Most zeros are adequately or very
well covered, rendering them relatively low risk. Advocates will
proclaim the fact that, since zeros were invented 20 years ago,
there has not been any issue that has not repaid its full price
on redemption. This is an enviable track record . . ."
With such impressive past performance and clear
evidence of high yields in the sector, cover and hurdle rates
were considerably less important to the investor than currently.
This is demonstrated by the fact that until recently, there was
no standard method of calculating the cover ratio or hurdle rate.
Over the last 30 years, various methods have been formulated and
used by City stockbrokers and the AITC and they all aimed to provide
the same simple measurement in slightly different ways.
By way of example, when splits were first introduced,
interest on loans was always paid out of the income account, but
in recent years funds have taken to apportioning part of this
cost to the capital account. This means that the fund's total
assets at redemption will be reduced and hence this effect has
to be taken into account when calculating the cover ratio. The
different approaches for calculating cover and hurdle rates and
the detailed analysis that underlie those calculations meant that
it was extremely difficult for third parties to evaluate these
figures to determine relative risk between different trusts.
In her article "Devil in the detail"
published in the Bloomberg Money Guide in May 2001, Clare Gascoigne
quotes David Crouchen, sales director for Exeter Funds Managers:
"Where can the ordinary member of the public
get access to the kind of detailed research you need to compare,
say, hurdle rates? It's very difficulta lot is only made
available to institutions or specialist independent financial
advisers. The average, general purpose independent financial adviser
won't be receiving research showing the level of gearing. But
you can't make a decision without it."
Moreover, the exclusion of gearing from many
cover and hurdle rate calculations meant that potential investors
could assess the true risk of a particular trust only with the
help of additional information on gearing ratios. Clare Gascoigne
goes on to explain:
``Finding out what level of gearing an individual
investment trust has, however, is difficult for the ordinary investor.
Although the [AITC] publishes gearing levels of ordinary investment
trusts in its monthly statistics pack . . . it does not include
a figure for splits. It does offer that information on its website
. . . but even then only for the last ranking class of share on
Instead, investors must make a judgement based
on a number of variable factors: the asset cover (the higher the
better); the hurdle rate (the lower the better); the number of
years until wind-up (the longer a trust has to go, the more risks
it can take); and the underlying investments of the trust (the
riskier the sector in which it invests, the more dangerous it
It was only once the prospect of zeros not paying
out in full became real that commentators started to analyse the
use of cover and hurdle rates in any detail; it was only earlier
this year that the AITC agreed a standard approach to the calculation
and reporting of cover ratios.
Much has been written recently criticising the
use of cover and hurdle rate as measures of risk for zeros. However,
this analysis has the benefit of hindsight in the context of falling
markets and the failure of some zeros to pay out in full on the
redemption date. Prior to last year, where the markets were strong
and no trust had ever failed to pay out on zeros in full, considerably
less importance was placed by market analysts and financial advisers
on cover and hurdle rate. Instead, many considered that the existence
of high yields and the track record of the sector were sufficient
indication of zeros' low-risk status.
5. THE BARBELL
Most investment trusts, including splits, hold
a single focused ("bullet") portfolio of equity investments.
By contrast, "barbell" trusts hold two distinct portfolios
of investmentsa low yielding growth portfolio and a high
income portfolio. The income portfolio is typically financed by
debt, while the growth portfolio is financed by equity, and this
structure is designed at issue to provide the underlying growth
from the equity portfolio in addition to a high yield.
Several barbell trusts have more than one class
of share capital and are therefore splits but it is technically
incorrect to describe all barbells as splits. Gearing levels in
these funds are typically over 100 per cent, which should be contrasted
with most conventional trusts, where gearing is more commonly
in the region of 10-15 per cent.
A good summary of what makes a barbell trust
can be found in Dr Andy Adams' and Robin Angus' article about
barbell trusts in April 2001 Professional Investor Magazine:
"While a good number of barbells are splits,
most splits are not barbells. It is the division of the assets
side of the balance sheet into two separate portfolios, together
with a high level of gearing, which makes a barbell."
The article went on to say:
"In our view the "barbell" concept
is fatally flawed. But the split capital concept has proved its
worth over many years. The FSA must not confuse barbells with
splits. If additional regulation is required, it should be applied
uniformly across the entire UK closed end fund industry."
In its publication entitled "Barbells unbalanced"
of July 2001, Cazenove reported that £9.2 billion of total
assets had been raised by highly geared and split IPOs since the
beginning of 1999. More significantly, during 2000 and 2001, of
the £7.1 billion of total assets raised, barbells accounted
for £4.2 billion, or 59 per cent.
5.2 Advantages of the barbell strategy
There is an element to which the barbell strategy
is a natural development from the structure of splits. As an investment
trust, splits require a regular income in order to satisfy the
income shareholders, while also maintaining good capital growth
in order to meet the entitlements of zero holders and the expectations
of the capital shareholders. To this extent, therefore, it is
necessary for any manager of a split's portfolio to balance the
portfolio between income producing investments and those that
will generate capital growth. Given that the directors of the
trust must adopt an investment strategy that is to the benefit
of its shareholders generally, having express regard to the dual
nature of the split's profile is clearly appropriate.
Much commentary on "barbells" has
recently centred on fall in the income portfolios, many comprising
corporate bonds and split capital income shares. It is important
not to underestimate the expectations of the underlying equity
asset base that was expected to offer the growth component from
The chart above, produced by HSBC in March/April
2000, demonstrates the weighting of the split capital sector relative
to other equity sectors worldwide. Relative to the World Index,
the splits sector was overweight in the UK equities, thus concentrating
portfolio risk. The aim of many barbells was to diversify this
growth risk away from UK FTSE 350 shares into different equity
asset classes, such as US equity, UK smaller companies, technology
shares and Far Eastern shares for example.
The majority of fund managers and investment
banks during the 1999 and 2000 period were confident of continuing
returns from equity markets.
Whether the manager chooses to manage a single
portfolio balanced between income and capital growth or expressly
divides the portfolio into two notional asset pools, the overall
investment strategy would be similar. The expectation would be
to invest the portfolio prudently to achieve both income and capital
growth with the added security that if one part of the portfolio
suffers a fall in markets, the other part may hold up the trust
5.3 Disadvantages of the barbell strategy
The disadvantages of barbells are generally
related to a number of factors that could affect any split, whether
or not it followed the so-called barbell strategy.
First, many barbell trusts requiring restructuring
have chosen to increase gearing through bank borrowing. As highly
geared trusts, therefore, barbell trusts are open to the same
risks as any other highly-geared trust. This is discuss further
at section 7 below. Moreover, if barbells hold cross-holdings
in similarly highly-geared trusts, the risk of enhanced losses
is greater. The effect of cross-holdings is considered at section
Secondly, barbells have increased in number
dramatically since 2000-01, as noted above. Consequently, they
have suffered the unfortunate timing of having been launched at
the beginning of market decline, a decline which has been more
marked than, possibly, could have been expected, particularly
since September 2001. Barbells have as a class, therefore, merely
evidenced the recent decline in the markets that has affected
all forms of market investment.
Thirdly, in some cases barbells have suffered
from the market decline to a greater extent than other investments.
For example, where a barbell trust invests in the technology sector,
it might be an appropriate management strategy to invest in direct
equities to provide capital growth, while also investing in bonds
to provide income. However, the nature of the recent market decline
has been to affect both the equity and bond markets. Consequently,
the perceived advantage of dividing the portfolio so that one
part might support the other in adverse market conditions has
The barbell strategy has become increasingly
popular for the management of split capital trusts over the last
few years and should be viewed as an appropriate strategy to match
the requirement of the split capital trust to balance both capital
growth and income.
However, the timing of this increased interest
in barbells has coincided with a general decline in all major
equity and bond markets. Consequently, it would be too easy to
align the current market losses with the barbell strategy itself.
In reality, the losses suffered by barbell trusts are, as with
any split capital trust, caused by the inter-relationship of various
factors such as gearing on falling asset classes and to a lesser
extent portfolios in geared investment themselves: split capital
income shares or "cross investment".
6. THE MAGIC
6.1 Origin of the expression, the "magic
The phrase "magic circle" is generally
taken to refer to split managers that invest in other investment
companies. 10 years ago, there were far fewer fund managers managing
splits and "fund of fund" splits predominated. However,
during the low yield environment of the late 1990s, the split
capital sub-sector grew quickly from being a small constituent
part of the investment trust market with relatively few participants,
to a large part of the investment trust sector (see section 2.2
Cross-investment is used in this paper to refer
to the ownership by a trust of a portfolio of income shares in
other splits to boost portfolio yield. It is possible for cross-investment
to occur by trusts holding zeros in other trusts. However, this
is less common.
While the term may suggest that trust cross-investment
only occurs where, for example, trust A owns shares in trust B
and trust B owns shares in trust A, this is a much narrower definition
that may confuse certain of the issues for co-investment generally.
Of course, in practice, it may be the case that
the narrower form of cross-investment does occur. For instance,
it is quite possible for the following scenario to come about:
In this example, it could be said that Trust
A, in effect, owns its own shares. However, the number of intermediate
owners means that:
The manager of Trust A is highly
unlikely to know that there is this form of cross-investment;
It would be extremely difficult to
"engineer" this form of cross-investment deliberately
to affect the split capital sector; and
The possible disadvantage of Trust
A's portfolio of owning its own shares is significantly diluted
by the potential benefits of exposure to the performance of the
underlying investments in Trust B (in which Trust A's manager
has knowingly invested) and, indirectly, those of Trusts C and
"Cross ownership" and investment trust
companies investing in income shares or other investment trusts
have become incorrectly intertwined in the recent debate. "Fund
of funds" has long been a feature of the split capital market,
offering diversification, asset allocation and tax advantages
over a single portfolio.
Nigel Sidebottom of BFS in the Crédit
Lyonnais Investment Trust Yearbook explains some of the problems:
"Press reporting on the split sector has
implied that an investment in another investment company is in
itself a cross holding. This is of course nonsense. If Fund A
invests in Fund B, then it has made an investment in another investment
company. That investment only becomes a cross holding if Fund
B then invests back into Fund A, either directly, or through a
series of other investments, eg Fund B invests in Fund C, which
invests in Fund D, which invest in Fund A. A trust with high exposure
to other trusts may have no cross holdings."
The impact of cross-holdings to recent falls
and the potential for a spiral has to be qualified and this is
in fact noted in the original July 2001 Cazenove report. Sidebottom
"This impact, in many cases, will not be
as severe as many might think. To examine the potential effect,
let us assume Trust A and Trust B are both conventional trusts
with no gearing and their shares trading at NAV. They each have
a 10 per cent cross holding in the other. Trust A's share price
falls 10 per cent. This will cause the NAV of trust B to fall
1 per cent. Assuming Trust B's share price follows its NAV the
resulting 1 per cent decline in share price will cause a further
0.1 per cent fall in A's NAV (in addition to the 10 per cent decline
that triggered it). Assuming Trust B's share price follows its
NAV the resulting 1 per cent decline in share price will cause
a further 0.1 per cent fall in A's NAV (in addition to the 10
per cent decline that triggered it.) We can conduct a second iteration,
which will reduce A's NAV by another 0.001 per cent. By the third
iteration the impact is de minimis."
6.3 Advantages of investment in the split
As discussed above, the nature of a split trust
means that a manager has to adopt a strategy that produces both
capital growth and income. Against this background, the income
shares of other splits offer an appropriate risk profile to achieve
income for the trust as part of its overall portfolio. In fact,
there are relatively few investments that are able to offer the
manager the same or an equivalent combination of income and risk
as offered by the split capital sector. One possible alternative
is a corporate bond, which already forms a part of many split
portfolios. However, while offering equivalent yields to the split
income shares, coporate bonds do not offer capital growth in the
secondary market and lay open the portfolio to their own default
Recognising the split capital sector as an investment
type in its own right therefore has served trusts well over the
years. For example, in the mid 1990s, split capital funds of funds
were a highly successful investment and supported dividend payouts.
6.4 Why has cross-investment been criticised?
Brewin Dolphin Securities wrote as recently
as 19 March 2002, "Cross-holdings, by which we mean trust
shares held by other trusts, existed in July 2001, but had been
there for years. No law or fundamental economic reason restricts
the holding of such portfolios." (Brewin Dolphin, among others,
analysed the effect of cross-holdings in June 1997 and again in
In reality, as at 31 March 2002 only 15.43 per
cent of the sector as a whole was invested in split capital investment
trusts. 73.59 per cent of the sector was invested in equities
and bonds, and a further 7.25 per cent was invested in cash.
With the decline in markets generally and the
more frequent default of split trusts as a consequence, split
portfolios invested in other splits have suffered alongside other
investors. Consequently, the effect of market decline on the overall
value of the portfolio has been enhanced by applying across the
portfolio, whether it is invested in equities, corporate bonds
or other splits. As Merrill Lynch has explained:
"Cross ownership imposed through traditional
split capital structures gives rise to acute gearing and increases
the embedded fund expenses at the outset. Zeros issued on such
portfolios are unlikely to be low risk investments. Split capital
structures with high levels of cross ownership can only repay
zero dividend holders when stock markets rise dramatically, despite
the fact that the zeros may be covered at the outset."
In addition, it is suggested by commentators
that cross-investment concentrates market weaknesses in the splits
sector such that splits invested in other splits are affected
to a greater extent by issues such as portfolio diversification
and liquidity and dividend cuts. Considering these in more detail:
Increases in cross ownership tend
to increase the level of correlation between the different constituent
investments of the split's portfolio. Increasing correlation increases
the risk in a portfolio (by reducing diversification), as all
the constituents tend more to move in the same direction, exaggerating
any market moves.
It is thus argued that cross-ownership
increases both the volatility and the correlation of the underlying
portfolio. Consequently, the market decline experienced more recently
has had a more marked effect on splits where high levels of cross-invesment
With high levels of gearing in a
falling market, any income achieved by a trust will most likely
be used in paying borrowing costs, reducing the amount available
to income shareholders. If, as a result, the investee split elects
to make a dividend cut, this can have a marked effect on the value
of the investing trust. When faced with a dividend cut (either
through the core portfolio or cross-owned income shares) the portfolio
manager can do one of two things:
(i) Reflect the cut through their own
income shares (thus reducing the yield).
(ii) Attempt to maintain the dividend
and fund the cut by increasing the portfolio yield.
The course of action taken by a fund manager
is fundamental to the security of a zero dividend share. In summary,
in (i) there will be little effect on zeros, provided that any
effect on the net asset value is not so marked as to affect cover
levels. In (ii), it will almost certainly have a negative impact
on the security of the zero shares. The portfolio will need to
be restructured to increase the portfolio yield which may involve
converting capital into income.
In both of these examples, it is not the fact
of cross-investment that leads to decline in values of the investing
split. Instead, it is the manager of the portfolio who has to
determine, in line with the strategy set by the directors of the
trust, the appropriate action to take. In the first case, a portfolio
with high levels of correlation and poor diversification will
suffer just as badly (or worse) in a falling market than a split
that has a moderate level of cross-investment balanced with a
diverse portfolio with low correlation. Similarly, in the second
case, the manager of the split may elect to follow option (i)
by implementing its own dividend cut, with minimal effect on the
levels of capital available to meet the claims by zero holders.
One further criticism of cross-investment has
been the allegation that managers have colluded deliberately to
influence market prices in the splits sector.
6.5 Allegations of collusion
As explained above, it makes good investment
sense for a manager of a split to acquire high yielding income
shares, as these fall squarely within the trust's requirements.
One source of such shares is splits issued by other managers.
The relatively modest historical scale of the sector inevitably
meant that there were relatively few managers involved ten years
ago, and so the range of splits available for investment was more
limited than it is today. However, to categorise this as something
collusive or improper is quite wrong.
Moreover, the market prices of shares in investment
trusts are independently determined by market makers in the light
of supply and demand and following strict rules laid down by the
London Stock Exchange. There are a number of brokerage houses
that operate in the sector and up to seven competing market-makers
that set prices for shares in the secondary market. Primary prices,
for new issues, are outlined in the issuing prospectus. These
prices are in no way determined by other managers and, therefore,
it is highly unlikely that cross-investment would directly affect
the pricing of individual trusts or the sector overall.
Cross-investment by split managers can have
a tangible benefit to the strength of the portfolio and income
shares should be viewed as an investment category in their own
right, forming part of a balanced portfolio requiring both income
and capital growth. The historical success of funds of funds in
the splits sector demonstrates that appropriately managed cross-investment
in strong market conditions can benefit the investor.
However, the potential side-effects of cross-investment
in terms of enhancing the effect of gearing, increased correlation
and loss of portfolio liquidity have been experienced more recently
in worsening market conditions. Again, therefore, it is not necessarily
appropriate to criticise cross-investment per se, but rather
to recognise the interrelation of various factors that have affected
the sector in recent months.
Unlike unit trusts, investment trusts can borrow
money and invest the proceeds. This is known as financial gearing.
Typically the level of gearing of a particular trust is described
as a ratiogearing factor of 120 means that a trust with
equity of £100 million has £20 million debt (bank borrowings).
Launches over the past couple of years typically had heavy gearing,
for example £100 borrowing for every £100 raised. But
while this is high, most homeowners aged 26 have a 5 per cent
equity component and 95 per cent debtthe mortgagein
their house so it is not automatically appropriate to equate gearing
with risk or as a contraindication to retail investors.
Gearing is also a term used to describe the
structure of a split capital trust. In a simple trust, typically
each class of share is entitled to either all the income from
the underlying portfolio or all the capital growth. If each class
of share were issued in a 50:50 proportion, then the income share
could be said to have double gearing. This use of the term gearing
is generally referred to as structural gearing.
7.2 How gearing works in investors' interest
The ability to gear can be beneficial when markets
are rising as gearing can be used to enhance gains. This is because
a manager can borrow funds from the bank at, say, 4 per cent and
use the proceeds to purchase stock that will yield say 6 per cent.
This is clearly beneficial to investors as it increases their
exposure to equities and hence the potential return. This is also
a tactic that has served investment trusts well up to nine or
so months ago. As a share class, while zero holders are only entitled
to a fixed return and would not, therefore, benefit directly from
any gains enhanced through gearing, the overall return on the
portfolio is equally important to zero holders. This is because
by maximising returns, the manager also increases the cover ratio
or hurdle rate for the zeros, which will directly affect the price
of zeros traded in the secondary market. This has been widely
recognised and endorsed:
"Daniel Godfrey, the director general at
the AITC, said the ability to gear the fund improved performance.
`If the markets are going up over time, which they tend to do,
gearing will add value. A recent report by Cazenove estimated
that gearing added 1 per cent a year on average to the value of
investment trusts,' he said."
While an academic financial commentator remarked:
"The first purpose of this paper is to use
split capital funds in order to examine if the `financial engineering'
of a company's liabilities increases their value relative to the
market value of their assets. We find that 11 per cent of extra
value is generated. Perhaps this is not altogether surprising,
as otherwise there would be no reason to develop such complicated
structures in the first place. The second purpose of the paper
is to pinpoint more precisely from where the extra value comes:
is the result of having debt, separate shares for the dividend
stream or a date on which the company will be wound up. It transpires
that the prime/score split into dividend shares and capital shares
increases fund value significantly, but only by an average of
1.3 per cent. A much larger gain in value of 5.6 per cent comes
from levering the fund's returns wih debt: zero dividend preference
shares do this in a tax advantaged way and so are found to be
particularly good at adding value."
However as we have experienced recently, gearing
can also enhance losses when a sector is falling. As Kestrel Research
reported in a Research Bulletin of April 2000:
"while cross share holdings (where trusts
effectively own a portion of their own equity) do add to the effective
gearing of a geared trust, they should not in themselves lead
to a downward spiral in net asset value in the event of a fall
in the underlying assets. It is the gearing upon gearing seen
in barbell trusts in their income portfolios rather than the cross
share holding risk per se which is linked to concerns about
a systemic collapse."
7.3 Why has the impact of gearing become apparent
Gearing is an effective management tool to enhance
gains in a rising market and, until recently, has served to set
the investment trust sector apart from other forms of investments
by providing competitive returns to investors. With a long term
view of rising markets, this remains the case. As Rolly Crawford
stated in February 2002:
"And gearing has the obvious answer to enhancing
returns in rising markets . . . An example of that, going back
to the history of Second Alliance between 1950 and 1965, the NAV
rose eight times, which was twice as fast as the rise in gross
assets. And in times of falling markets, clearly gearing can be
very damaging and in the aftermath of the crash in 1929-32, the
NAV of Second Alliance fell by around 80 per cent . . . High-yielding
stocks suffered very badly during the recession of the early 90s
and split NAVs fell as well. The NAV of Scottish National capital
shares was negative at the time for many months, though they were
actually finally repaid at over 104p a few years later . . . Gearing
is the greatest single factor contributing to the loss of value
in the sector. Geared funds fall rapidly in falling markets, whether
it is today or [the crash of] 1987 or 1980." 
Over recent months, particularly since September
2001, the markets have fallen at an unexpected rate and the effect
of gearing has compounded losses in the short to medium term.
This is because, while the percentage loss compared with the total
portfolio remains the same, the cost of borrowing must still be
met out of the remaining asset value. In effect, therefore, the
cost of borrowing accounts for a higher proportion of the remaining
asset value, thus reducing the amount available to shareholders.
An additional factor that has brought the existence
of gearing to the fore is in the accounting treatment of the cost
of borrowing. It is the responsibility of the directors of each
trust to determine the trust's policy on the allocation of expenses
either against the capital or revenue account. In simple terms,
deducting the cost of borrowing exclusively from the revenue account
benefits zero holders and capital shareholders, while deduction
from capital account benefits income shareholders.
International accounting standards have changed
such that investment trust boards have to reconsider the allocation
of borrowing costs. Whereas the cost of borrowing had previously
been deducted from income account in the majority of trusts, there
has developed an increasing practice to deduct from capital. Each
Board has to make this decision based on a number of assumptions,
including the future prospects of the markets, where is was anticipated
the trust would obtain its returns. In practice, the effect of
this decision has been to reduce the level of cover for zero holders.
Gearing magnifies returns favourably or unfavourably
in rising or falling markets. All split capital closed end funds
are geared: either structurally or financially. The adverse impact
of gearing on certain share classes in falling markets is not
a new phenomenon. However the nature of gearing, and particularly
now how flexible bank borrowings proved to be in adverse markets,
is a topic of renewed interest.
8. PRESS COMMENTARY
8.1 The press coverage of splits has historically
been positive and enthusiastic in relation to investments in zero
dividend preference shares, and began to turn sour only in mid-2001.
A historical survey of the financial pages presents the following:
"Only if your chosen trust puts in a truly
disastrous performance will the zero holders' return be at risk."
"Zero dividend preference shares in a split
capital investment trust are another safe option. These are shares
which pay no dividends, have a finite life and a fixed yield.
Consequently the return is guaranteed." Daily Telegraph
"The risk of less back at maturity than
you invested in a zero is very small. Stock markets would have
to fall through the floor." Daily Mail
"The risk on this is pretty low. A good
idea for risk averse investors, for school fees planning and those
approaching retirement." Daily Mail
"Historically, the volatility characteristics
of portfolios of zeros have been very similar to gilts and significantly
lower than UK equities . . . Indeed, since the first zero was
issued in 1987, no zero has yet failed to be repaid in full."
"Zeros currently display more safety features
than a Volvo." Money Marketing
"It is now widely accepted that clients
who require low risk should have some exposure to zeros."
"A meltdown of the splits sector remains
a remote possibility, and so long as investors choose wisely,
zeros will continue to be a low-risk way of wringing out a better
return from your money than if you put it on deposit with the
building society." The Times
Then by the middle of 2001, more negative coverage
began to appear in the financial pages, for example:
"Many investors . . . have been seduced
by what they have been told is another safe investment. Zeros
have been around for two decades and are often billed as a halfway
house between gilts and equities . . . However, share prices of
some zeros have fallen sharplyin some cases as much as
25 per centon fears that if stock markets do not recover,
for the first time in their history, some may not have enough
assets to repay their zero shareholders. As a result adviser R
J Temple is advising its clients that zeros have moved up the
risk scale. It rates them as a bigger risk than corporate bonds
. . . `We feel that many investors may yet be unpleasantly surprised
by what they have bought. A lot of the perceived protections in
many of these investments are more illusory than real,' warns
Temple." Financial Times
"But how safe are [zeros] now that markets
are tumbling and split capital trusts are being forced to increase
their gearing all the time? Well the answer is, not as safe as
they once were . . . for the first time, there is a real danger
that a zero may fail to return to investors the fixed sum it said
it would . . . Aberdeen's run of form looks like it may be coming
to an end. Two months ago it was forced to suspend the dividend
of its European Growth & Income Trust which has run into dire
straits after borrowing heavily to invest in telecoms bonds, and
its technology fund, long a top-performing unit trust, has also
suffered. There is a chance it may meet the repayment but it's
not big. It would be best to steer well clear of Broadgate Trust
and its zeros." Business a.m.
The next day, terrorist attacks on the World
Trade Center precipitated even greater troubles in the already
falling world markets. By October 2001, Brewin Dolphin Securities
commented on the press coverage of the split sector thus:
"Manhattan on 11 September produced the
tail spin and panic selling by private clients after an event
which has no precedent and which was a complete surprise. The
flames of greater client anxiety were then fanned by ignorant
but often wilfully misleading press comment which has continued
from August until [19 March 2002] when the Financial Times
once more promoted bearish prognostications about the split sector
against the more reassuring evidence . . . which is becoming apparent
today . . . sensational and ill-informed press comment has been
largely responsible for this situation and prolonged it to a stage
where small shareholders have lost a lot of money quite needlessly."
Aberdeen Asset Management PLC is concerned about
the level of confusion surrounding the debate on splits and is
naturally anxious about the level of falling returns for many
We are working closely with trade bodies and
FSA alike to help restore investor confidence.
Aberdeen Asset Management has proposed a number
of initiatives for the sector going forwards, including recommending
increased transparency of reporting and a new risk rating system.
It has also taken the lead in waiving fees and proposing other
recovery initiatives for the sector.
We would be pleased to supply any further information
as well as Aberdeen's briefing document of 15 April 2002 and a
monthly sector monitor. Aberdeen has established a consumer website
at www.aberdeen-asset.com/splits where investors are able to source
additional information about splits managed by the Aberdeen Asset
Management Group. Aberdeen initiated and sponsors www.splitsonline.co.uk,
a dedicated information site updated daily.
1 Merrill Lynch, 2001, Zero Dividend Preference Shares:
The developing marketplace-a comparative analysis. Back
Merrill Lynch, 2001, Zero Dividend Preference Shares: Understanding
the risks and how to price them. Back
Merrill Lynch, 2001, Zero Dividend Preference Shares: Understanding
the risks and how to price them. Back
Merrill Lynch, 2001, Zero Dividend Preference Shares: The developing
marketplace-a comparative analysis. Back
Merrill Lynch, 2001, Zero Dividend Preference Shares: The developing
marketplace-a comparative analysis. Back
Daniel Godfrey quoted in Investment Adviser FT Supplement
December 2001 at pp 16 and 19. Back
First written in October 2000, updated 2001, by Mr Gemmill of
City University. Back
Address by Rolly Crawford, an analyst at ABN Amro Hoare Govett,
at "The Regulation of Split Capital Closed End Funds",
a one-day conference on 1 February 2002. Back