You are on page 1of 2

Corporate statement Barclays Capital

Barclays CORETM – opportunities in


trading equity correlation
This article describes how, with its CORE strategy, Barclays Capital is opening up
access to correlation for a wider investment audience

Investing in ‘implicit’ assets such as dividends, volatility and How to trade it


> correlation has become an accepted addition to traditional Both implicit and explicit exposures to correlation are possible. By
investments within a portfolio, such as equities and bonds. While several buying or selling basket options, one can indirectly express a view on the
investment banks provide clients with tools to invest in volatility or correlation of the basket constituents. However, a more direct means is
dividends, including our own benchmark investable Voltaire IndexTM and suggested by the fact that basket (or, in general, index) volatility is linked
Dividend Swap Indices, access to correlation as a specific investment has to constituent stock volatility by the correlation between the stocks.
until recently been limited primarily to hedge funds. In this article we show Indeed, the variance of equity index returns can be expressed statistically
how, with Barclays Correlation Reversion strategy (CORE), Barclays Capital as a sum of variances of constituent stock returns and cross terms
is opening up access to correlation for a wider investment audience. involving the correlation between the stocks:

The case for trading correlation


Correlation between the constituents of stock indexes is a widely
2
σ index = ∑ (w σ ) + ∑ ρ (w σ )(w σ
i i
2
ij i i j j )
i = 1..N i≠ j
monitored quantity, which is traded in the equity derivatives market by a
number of banks and leveraged investors. Being a high-order derivative, Consequently, by trading single stock versus index variance swaps,
correlation, unlike stock prices, tends to exhibit mean reverting behaviour investors can trade correlation. This is encapsulated in our measure of
(see figure 1). Furthermore, it is mathematically constrained above and correlation – what we call imputed correlation – defined as the ratio of
below, allowing easier judgement of whether levels are richer or cheaper index variance to weighted single stock variance:
2
than, say, for stock prices. These properties and a well-established market
σ index
for dispersion trades (where investors take a view on the performance of pimputed =
a basket versus its constituents) have led to correlation trading becoming ∑ wiσ i2
i = 1,.., N
a staple of many leveraged market participants over the past several
years. Indeed, many leveraged investors have a better idea of where they This measure yields a simplified notion of an average level of
would consider selling correlation compared to volatility, which does not correlation between stocks that we consider more intuitive than the
have a maximum limit. involved definition that is derived from the statistical formula.

1 Historical three month imputed correlation, calculated


quarterly (DJ Eurostoxx 50 index) Intuition for expressing a correlation view via volatility

80% Consider an index comprising two equally weighted stocks that are
70%
perfectly (positively) correlated. All else being equal, a decline in the
60%
correlation between them will result in a decline in index volatility, as
50%

40%
the movement of one of the stocks will (partly) negate the movement
30% of the other, dampening the movement of the index. The converse
Source: Barclays Capital

20% works too, i.e., an increase in correlation can result in greater index
10%
volatility, using a similar argument. In simple terms, a short correlation
0%
position can be established by being short index and long single stock
2

95

99

03

07

08
9

0
19

19

19

19

19

19

19

19

20

20

20

20

20

20

20

20

20

volatility (making the trade, initially, volatility neutral), and vice versa.
ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch

ch
ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar

ar
M

Any data on past performance, modelling or back-testing is no indication as to future performance

1008_SP_BarCap.indd 36 2/10/08 10:41:38


The case for short correlation trades variance swaps on the component stocks. Variance swaps are flexible
We have observed that correlation implied by the market has typically investment tools, which do not typically require initial funding. For many
traded at a premium to what is subsequently realised. Options- investors however, it remains difficult to establish a strategy similar to
based trading flows in the market go some way towards explaining Barclays CORETM due to the risk profile of individual variance swaps and
the existence of such a correlation risk premium. Traditional equity associated requirements for risk management tools and infrastructure.
portfolio managers who use equity derivatives generally purchase puts Through the development of our strategy, Barclays Capital has created a
to protect their portfolios or sell calls to generate additional yield. Given means of access to correlation trading for a wider audience of investors,
the ease of use and lower cost, portfolio managers generally purchase effectively overcoming practical obstacles and providing the potential for
index puts. Moreover, they often prefer to leverage their individual good investment performance.
stock knowledge to choose candidates for call overwriting (i.e., selling Historical simulations of the strategy show that it would have
calls on shares they own), something that generates more premium performed well in a variety of market scenarios. The correlation of
than overwriting the index. Taken together with medium-term stock Barclays CORETM to traditional markets such as commodities, bonds or
volatility supply from products like reverse convertibles, these flows equites is relatively low, meaning that it can be a good addition to an
bid up index volatility relative to single stock volatility. Recalling our investment portfolio, providing a source of diversification, along with
definition of correlation, this translates to elevated implied correlation. potentially strong, stable returns.
Also, during market stress periods, spikes in index implied volatility (as Historical simulated performance of Barclays CORETM
investors rush to hedge long exposures) can exacerbate the size of this
structural flow-induced premium, driving implied correlation higher Barclays Bonds Equities Commodties
CORE™ (QW1A (MSCI World (SPGCCITR
still. Historically, leveraged investors have come in at these points to bet

March 18, 2002–June 23, 2008


(50-50) Index) Index) Index)
on realised correlation being lower than what the implied correlation
Annual
suggests via dispersion trades. 25.92% 4.31% 5.57% 18.57%
return

Source: Bloomberg,
Volatility 13.76% 4.17% 20.70% 18.78%
The rationale for a long-short correlation strategy
Over the past year, however, a number of investors in existing trades Sharpe Ratio 1.77 0.31 0.12 0.80
have been caught out by spikes in correlation during market dislocations, Correlation – 1.94% 13.93% 11.47%
begging the question as to whether it can be profitable to enter into
Any data on past performance, modelling or back-testing is no indication as to future performance
long as well as short correlation trades. Based on the historical back-
testing of Barclays CORETM, our new dynamic, fully automated long-short
correlation strategy, we believe the answer is yes. Indeed, our historical
Barclays CORETM forms part of the Barclays Capital Q-Series,
analysis shows the strategy having outperformed one that is only short
a range of quantitative investment products and strategies
correlation.
developed by Barclays Capital’s Equity Derivatives team. The
Q-Series aims to replicate trading strategies and provide access
Introducing Barclays CORETM
to investment opportunities that have historically been employed
Barclays CORETM is a transparent trading strategy that capitalises on the
by a section of the active asset management community.
difference between implied and realised correlation of constituents
These strategies, which might be linked to the performance of
of three major equity indexes: the DJ Eurostoxx 50; the S&P 500; and
econometric models, relative value opportunities or, as in Barclays
the Nikkei 225. Building on the fact that correlation is mathematically
CORETM, implicit assets, are delivered in a transparent manner with
bounded above and below, and that it tends to exhibit strong mean
focus on enhanced liquidity and cost efficiency.
reverting behaviour, Barclays CORETM applies the hypothesis that, by
systematically comparing current implied correlations with historically
realised levels, one can identify whether the current implied levels
are rich and therefore whether it is a good time to go short implied Contact details
correlation. Using this same approach, the strategy also looks for Vladimir Portnykh, PhD
Equity Derivatives Structuring, London
appropriate opportunities to either go long implied correlation or,
t: +44 (0)20 7773 2968
indeed, to stay out of the market. e: Vladimir.Portnykh@barcap.com

Abhinandan Deb
Strategy implementation
Equity Derivatives Research, London
Barclays CORETM makes use of one of the market’s most liquid instruments t: +44 (0)20 7773 2481
– variance swaps – to establish its correlation positions. Variance swaps e: Abhinandan.Deb@barcap.com
allow investors to sell the correlation between index and constituent
www.barcap.com
performance by selling a variance swap on the index and then purchasing

1008_SP_BarCap.indd 37 2/10/08 10:41:40

You might also like