Portfolio Build
In what follows we outline how one might construct a suite of equity funds built using signals deciphered from micro-sector monthly economic data.
In what follows we outline how one might construct a suite of equity funds built using signals deciphered from micro-sector monthly economic data.
UK – long-short stock specific portfolio building
US – long-short stock specific portfolio building and long-short sector strategy
Japan – long-short stock specific portfolio building and long-short sector strategy
Continental European – long-short stock specific portfolio building and long-short sector strategy
Depending on the regional market in question each fund would involve the construction of a long/short strategy for either defined sectors or specific companies. The data itself reflects monthly information at a level detailed enough to focus on over one hundred specific markets. Examples include dairies, grain millers, steel makers, auto component producers, manufacturers of telecom equipment, food retailers, restaurants and homebuilders. Often the sector detail is such that a clear mapping to a specific stock is possible.
Since the information is confined largely to production, retail, distribution, service and construction sectors we will have to focus on universes within which: these markets are significant; where financials, utilities and pharmaceuticals have modest load-ratios, or alternatively assume full weightings for sectors where we have inadequate data coverage.
In relation to back-testing all but our Japanese quant product has been active since at least 2001. Consequently, whilst we have not been following a model portfolio approach we can point to regular sector and stock successes. These include the rotation out of US supermarket space in 2002 and the rotation into European basic industrials from late 2003. The white-box nature of our data ensures that ‘retrospective back-testing’ can be achieved relatively easily and quickly.
Long-short stock specific building and long-short sector strategy
We now look at the four specific geographic universes from which one could construct a suite of regional quant-based equity long-short funds, based on cash equity or equity derivate products.
UK – long-short stock specific portfolio building
For the UK market the stock universes which are best suited to our work are those which lie below the FTSE100, which is heavily populated by utilities, pharmaceutical and financial stocks. Our argument is that as we work down the capitalisation scale we notice that the company constituents generally tend to be involved in a relatively modest number of market segments. |
Focusing on stocks which act in a small number of markets and tend to be country-centric simplifies our analysis considerably. For instance, Dairy Crest, Cranswick, DS Smith, St Ives, Johnson Press, Marshalls, Travis Perkins, Marconi Corp., TT Group, British Vita, British Polythene, RPC, Corus Group (albeit a FTSE100 constituent), Meggitt, Somerfield, Woolworths, Taylor Woodrow and Persimmon, can all be characterised as operating within only a handful of sectors, frequently just one, and often with a strong UK bias. For such stocks and sector we have reliable and timely monthly data. |
Although we would aim to create a combined long-short approach our focus would be better suited on short strategy. Clearly as we move down the capitalisation scale so liquidity will become an ever more restrictive constraint. We will therefore need to use stock lending and where possible CFDs. Any UK product is likely to be best constructed using a stock selection process rather than a weighted sector strategy. |
For UK-listed companies with a broad geographical spread we can exploit data comparable to that available for the UK in other geographies; the US and Continental Europe being crucial in this regard. For instance Hanson, BPB, Aggregate Industries, IMI, FKI, Unilever, Cadbury Schweppes, ICI, Wolseley, BP, Rexam, GKN, Tomkin, BAE Systems, Cobham and Signet we know that all derive significant earnings from US divisions. |