Foresight Capital Partners

Last week, Matthew Richards who writes for the FT on closed-end funds contacted me with regard to closed-end funds in general and an article he was preparing that week in particular. The article in question was on the activities of hedge funds who apply the strategy of closed-end fund arbitrage. This weekend Matthew’s article appeared in the FT, and for those who have access to the FT the article was entitled “City prey fight back against the big beasts”, published: November 18 2006, within the FTMoney section. As a direct follow up on this article I wish to make the following additional points:

1) The idea that such arbitrage activity only effect’s weak or failing trusts is just not true. There was the case of Edinburgh Investment Trust which obtain good relative performance in the small cap sector, however the sector as a whole fell out of favor and hence the discount widened to 20%. Recently there has been the case of Resource Investment Trust (REI.L) where the trust has not only obtained good performance but also offers essentially a unique product where existing investors in REI.L will find it difficult to find a suitable replacement. Moreover, in this case the fund management team is in the process of taking legal action against the trust for essentially deformation of reputation. The idea being that if you are running a trust which is wound up then it implies you are not any good. I cannot see this legal action succeeding since the contract between the fund and fund management team is very clear with get out clauses on both sides. However, the activities of arbitrage funds is increasing the risk associated with starting an investment trust (particularly outside of the latest hot sector), and this effect is just not in the interest of retail investors.

2) Matthew also mentioned the use of CFDs which arbs use to hide positions. In addition to this strategy there is also the tactic of obtaining huge amounts of loaned stock just before a vote on a special resolution (usually proposed by the arbitrage outfit) in order to push the resolution through. The point here is that the stock will typically only been loaned for 1-day (over the voting period) where the “real” shareholders are just retail investors which hold the stock in nominee accounts which (1) are not able to vote (due to being held in a nominee account) and (2) have no idea there stock is being loaned. Moreover, the special resolution may not even been in the investors long term interests. However this practice continues because the investment bank which is the agent in loaning the stock can obtain up to 10% over LIBOR, on an asset which it does not even own.

3) Essentially in symmetry to (2) above there is also the issue that within some closed-end funds there are differing classes of investors with differing voting rights. This has been a particular problem for equity warrant holders in Investment Trusts since if the trust is wound up the warrant holders get a particularly bad deal since the LSE uses a arcane formulae which greatly under-values the time value of the warrant. In particular, the arcane formula used by the LSE is not equivalent in any way with the widely used and accepted Black-Scholes option pricing model. What this has meant in practice is that an arbitrary outfit can buy a large amount of the ordinary stock, force a wind up (against the wishes of the non-voting warrant holders) and then not only obtain value from the narrowing discount but also get value from being able to give the warrant holders a bad deal. A good example of this was Bearings Emerging Europe Investment Trust which was restructured in 2002, see:

http://www.trustnet.com/general/news/display-story.asp?id=36978&db=it&txtS=y

There are some other points I would like to make but will not due to the risk of annoying the wrong people. However, if you know me in meat world then I am happy to let you know the gory details in confidence and in person.

As reported by Trustnet, see:

http://www.trustnet.com/general/news/display-story.asp?db=it&id=82081

the JPMorgan Income & Growth split capital trust is going to be rolled over into another split-cap structure. If you are not familiar with split-caps investment trust structures then there is a brief guide on the BBC site at:

http://news.bbc.co.uk/2/hi/business/1613719.stm

Now the significance of this event is that since the split-cap mis-selling scandal there has been no launches of split-caps. Moreover, all split-cap have a fixed redemption date and until now the funds reaching redemption have only offered investors to options of either cash/loan note or to roll the investment into an investment trust vehicle with a “vanilla” capital structure (i.e. only ordinary shares with possibly some gearing).

Anyway, I am glad to see split-caps of an asset class continue since for the more quantitative investor who makes the effort to do the analysis, opportunities do occur from time to time. For example, after the mis-selling scandal the zeros went to extremely low levels and back in the 90s (i.e. pre-internet days) for those who had the patience to get the annual reports and the time to unwind the various cross holdings and debt structures there was great opportunities.

Inflation on the United States of America in US Dollars

Calculator of U.S. Retail Price Inflation (1666-present), Wage Inflation (1914-present), Medical Cost Inflation 1936-present) using Tom’s Inflation Calculator

Links to Inflation Rates in other nation states

UK Inflation in pounds sterling
Japanese Inflation in yen
Euro-zone Inflation in euros
MEASURINGWORTH.COM

To be continued…

Interim dividend was announced at 5.1p this morning which implies a full-year dividend around 14.2p, which raises the divided by 19% putting the shares back onto a 5% yield. However, if you consider for example my original average entry price this year of 220.62p then the yield on this original investment is now running at 6.44%.

The reason I mention this is because the dividend which a stock pays is directly related to the earnings (or more precisely the cash flow), where as the price at which a stock trades is determined by the price market participants are prepared to pay for the present and expected future earnings. In short, the divided is determined by cash in hand, where as the stock price growth is determined by a change in perception which though will be influenced by changes in earnings is not directly related to them. Saying this, typically if the dividend payments of a stock move outside the normal range for that industry (to say 6.44%), and this payout is comfortably covered by the underlying business then one of two things will normally happen: (1) the dividend will eventually be cut because the earnings fall, (2) the share price moves so that the yield becomes more in line with typical market values (say 5%). Hence, if you can eliminate (1) and buy stocks on a high yield for the sector in question then the odds are already leant in your favor.

This is exactly what has happened in the case of BT.A, and provides an illustration why a dividend investing approach can offer tremendous risk/reward investment opportunities. Back in the Spring BT was trading at around 10 times earnings and yielding 5%. In the Spring BT management made it clear that they wished to move towards a divided payout of 2/3 of earnings which lead further support to the dividend and lead investors to believe that management believed a 2/3 payout to be a sustainable level for the underlying business. Now under such a situation and assuming the underlying business is healthy (please see earlier blog posts for this) the odds of a 25%+ return over the next year are very much stacked in the investors favor.

For example with my BT investment this year, there has been some multiple expansion from around 10 to 11, resulting in a 10% capital return before any move in earnings, the earnings have moved forward a healthy 20%, resulting in a total capital return of around 30%, while all the time I have been earning a 5% yield. Resulting in around a 35% return so far this year, even without any multiple expansion a total return of 25% would still have been achieved.

The final point I wish to make with such an approach is that in it important not only to analysis the health of the underlying business (i.e. the growth in earnings) but also have an insight into the way in which stocks multiples vary according to market cycles and themes. Since the large fluctuations in multiples which stocks trade has an outsized effect on performance. For example back in 2000, BT was on a multiply of over 40, and tobacco stocks such as BAT was on a single digit multiple; this year BAT has been on around 16 times earning where as BT has been as low as 9.5. Therefore, the effect of such changes in multiples over this period ignoring the effect of changes in earnings would be for BT to fall over 75%, and for BAT to rally over 60%. Therefore, understanding the dynamics of changes in multiplies is essential to realizing equity like returns (which in my mind our 25%+ annual returns). To be honest, for me the judgment of market themes and associated changes in multiples is more of an art than a science. Coming to a view on market themes does not follow as a deductive process but more of a slow realization, or evolving intuition of the underlying dynamics. However saying this, the development of this intuition is feed from undertaking daily research of the market which is generally undertaking as a formal maybe even scientific activity.

Maybe the market is 90% efficient, maybe even 99%, but personally the exact level of the inefficiency really is of no consequence because as far as I am concerned I only look at the in-efficiencies.

Yesterday, a conversation prompted me to recall an bit of quantitative research I did in 2001 regarding Correlation Trading of Energy products. This morning I dug this work up and provide it here for who ever may be interested. The idea behind this system is very much based on the fundamentals of the energy market where energy products are continuously shipped between various locations because of arbitration opportunities in the physical market.

Overview

Though by the application of our approach the generation of pure alpha in an absolute sense for trading parties may not be permissible, within the context of a large financial institution with significant customer flow within relevant markets the possibility of the generation of (meaningful) alpha may be possible from the context of absolute returns and over-all cost of capital reductions for the institution as a whole.

With regard to cost-of-capital regulations relating to product correlation we sight the FSA documentation, CM: Section 3: Page 2 of 5, 3.3.1, part 4, Date Issued 30.9.98:

Capital charges for each commodity are to be calculated separately, except under either of the following provisions, where positions may be treated as if they are in the same commodity:

1. positions in different sub-categories of commodities in cases where the sub-categories are
   deliverable against each other; or

2. positions in commodities which are close substitutes for each other and whose price movements
   over a minimum period of one year can be shown to exhibit a stable and reliable correlation of at
   least 0.9.
   1. The correlation referred to in (2) above relates to price movements and not to prices.

With reference to this guideline we investigate to what degree we can conclude that give products exhibit `reliable and stable correlation’. We argue further that any products found to exhibit reliable and stable correlation can be handled in the corresponding manner laid out within FSA regulation with regards to `cost-of-capital’ requirements.

Quantitative Trading and IT

Ideally one would wish to investigate the potential Correlations between as many energy products as possible however in order to do so requires such parties to have a reliable enterprise level IT infrastructure. In short, Excel will just not cope which such a task and I would suggest either using EJB/Java with DBMS (as I did) or SQL Server 2005 with ideally the business logic (i.e. .NET DLL) run from within the database. The reason why Excel will not cope is because with just 20 energy products there are 190 energy product pairs in which the Correlation properties should be investigated, if you have 40 energy products then there are 780 pairs, and for 60 products 1770 pairs. So the point I am trying to make is that in this instance the application of enterprise software platforms with some basic quantitative techniques can lead to a powerful and robust systematic trading technique. In short, IT can give you an edge.

Correlation Report and Basel II

The report from my original investigation can be downloaded from the following link:

Correlation Report (PDF File)

I never really took this approach too seriously from a trading point of view but was interested to see the possibilities in terms of cost-of-capital reductions and internal inter-book (or fund) systematic trading. The system value lies as much in the nature of the FSA regulation procedures as the underlying fundamental rationale for such an approach. Now with Basel II on the horizon, see:

http://www.bis.org/publ/bcbsca.htm

there will be end-less similar opportunities produced by this new regulatory framework. In short, Basel II being a set of guidelines rather than hard rules, will allow the smart and motivated banker to produce quantitative research motivated by the Basel II guidelines, which assuming the research is sufficient robust to get past the regulator will allow the host institution to reduce their cost of capital, feeding directly through to the bottom line.

Our Software

Sorry for the plug but my software firm (WebCab Components) offers all the Correlation functionality discussed above along with general linear/non-linear regression analysis, and ANNOVA type analysis which I imagine would be the functionality particularly relevant to cost of capital related analysis vis-a-vie Basel II. The relevant product would be one of the 4 editions of WebCab Probability and Statistics corresponding to the particular development environment you are using namely: J2SE (plain Java), J2EE, .NET (also contains COM, XML Web service implementations), or Delphi. See WebCab Compoennts home page, or following direct links to product desciption pages: J2SE, J2EE, .NET, Delphi. If you decide to use our software then you can even have the source files of my Correlation Report to re-use for your own reporting purposes.