Is manager selection top-down or bottom-up, and how many funds are in the portfolio?
Our approach is purely bottom up. We do not asset allocate across strategies as we have no evidence that such a subjective and directional approach adds value consistently. Our research highlights the difficulty of imposing a directional approach. If an investment manager is successful calling the future direction of a particular financial asset class, it is practically impossible to determine which fund will best perform under this scenario.
The portfolio currently holds 92 funds. The minimum number is 35, there is no maximum. Research into alternative investments suggests once diversification has been achieved across strategies', holding five to 10 hedge funds eliminates 75% percent of the specific risk in the portfolio. While we would hold with the rationale behind this deduction, we feel many important elements of risk have been ignored. Such research supposes the only risks to which a portfolio of hedge funds is exposed to are investment risks. Research by Capco has shown that the only 38% percent of hedge-fund failures were caused by investment risks alone, whereas 50% were caused by operational risks, 6% by business risks and the remainder were caused by a combination of the three. When one considers that hedge fund attrition rates vary between 2.2% and 12.3% annually, these figures have a profound implication for the structure of a fund of hedge funds (FoHF). Controlling investment risk alone will not effectively control total risk in the portfolio; an increased level of diversification is required.
Do you demand a minimum track record before investing in an underlying fund?
We are only interested in investing in funds delivering either substantially higher levels of return with a commensurate level of risk or those that deliver levels of return with substantially lower level of risk. This means we need a real track record before we can invest. Our policy is generally to avoid start up funds until a track record has been established, unless of course they are managers well known to us, operating under stable operational conditions. However, as all investment decisions are made on a qualitative basis, we will invest with managers with no less than a six-month track record where we have no prior knowledge. This period is too short to derive any quantitative assessment of performance, but it does demonstrate the operational capability of a start up and hence enables our investment on a practical basis. The fund is managed on a 'core and nursery' basis, with well-known managers with whom we have established strong relationships (the core) representing 60%-70% of assets. The remaining portfolio consists of a larger number of smaller or newer managers (the nursery).
When a new fund is introduced as a nursery fund (following extensive due diligence and risk committee approval), the holding is monitored over six months for consistency and profile of risk/return. If the numbers are as anticipated, exposure is raised towards the optimal weighting.
The fund aims to produce 7% yield annually. When is it decided whether or not to pay, and when is the payment made?
Dividends are declared twice yearly (using the January and July month-end prices) from capital gains realised from the investment portfolio. The investment manager cannot erode capital to meet dividend payments. Payments are made in March and September. There have been three dividend periods since inception. The fund paid out dividends in July 2003 and January 2004. In the period between 31 January, 2004, and 31 July, 2004, the fund lost 1.5% (in dollar terms) due to difficult market conditions in the first half of 2004, particularly April and May. As a result, the fund directors decided not to declare a dividend for that period. When the fund has paid out less than the 3.5% target in the first half of the year, this is taken into consideration in the second-half dividend, again according to directors' discretion. In the period 30 July, 2004, to 31 December, 2004, the fund has so far earned 8.21% (dollar terms). Since it did not achieve the 3.5% target in the first half of 2004, it is the directors' intention to make up the difference with the second-half dividend, fulfilling the 7% per annum objective.
What was the philosophy behind launching Alternative Income?
At the tail end of 2002 bond yields had reached generational low levels and any sign of interest-rate rises threatened to depress the fixed-income market. The Forsyth Alternative Income Fund was designed to provide an absolute return alternative to fixed income investing, with an added benefit of high income generation as well as preservation of capital gains.
Forsyth must balance healthy fund growth through performance with income. Some may argue paying the 7% could stunt the fund's AUM growth. How are these two reconciled?
Healthy fund growth is not threatened as the fund cannot erode capital to meet dividend payments. AUM growth merely slows by paying out the gains rather than compounding them. Additionally, we have created a non income-paying 'roll up' share class for those investors purely seeking capital gains
Does the income requirement steer Forsyth toward or away from certain strategies?
The dividend requirement has no bearing on portfolio construction and strategy allocation. Our objective to provide consistent risk-adjusted returns limits the portfolio's exposure to high volatility strategies. The fixed income focus of the portfolio prohibits investment in long/short equity funds.
The fund by its nature is heavily skewed towards fixed-income markets, but this focus is diversified across 10 distinct subgroups of strategies, broken down further geographically, thus limiting the portfolio's vulnerability to systemic shock. Within each strategy classification, managers can have very different approaches to making money in their own asset class, as well as being on different sides of the same trade. To try to capture the different forms of market risk, funds are analysed using regression methods to try to assess their exposure to a series of indices that represent different forms of risk. This analysis is carried out on a rolling basis to try to capture the changing nature of risks within the hedge fund arena. We accept the accusation that these descriptions will never be fully up to date. That is effectively why there are so many constraints placed upon the portfolio, to try to counteract the lack of confidence in any one of the classifications. To try to produce diversification within the portfolio, funds are grouped into blocks that are either positively or negatively related to the risk factors in question, and then limits are placed on maximum weightings on each directional exposure. It should be noted that, unlike the strategy classifications, these exposures are not mutually exclusive and funds may appear in several exposure groupings. The risk factors in question can then be aggregated into macro classifications that describe the broad approach to generating return. These classifications are: liquidity; volatility (related to implied or achieved volatility); and risk (directional market plays). Here minimum weightings are used to try to spread the sources of risk for the funds.
With regard to serial correlation, within our portfolio construction process we accept the possibility it exists, particularly within the fixed-income markets and, as such, we attempt to identify funds that exhibit serial correlation and correct for this. On the majority of occasions, the effect of serial correlation is to present much lower statistical volatility than the economic reality. Our process employs an 'un-smoothing' mechanism to adjust for this statistical error and we use these adjusted volatilities in our risk estimates.
How has strategy allocation changed over the past year?
Our strategy allocation is not run on a directional basis; therefore all changes to asset allocation have resulted from reducing or increasing exposure to specific managers. The key changes have been reducing the portfolio's convertible arbitrage holdings, which reflect difficult times since April, leading us to eliminate some under-performing funds from the portfolio. This has enabled us to allocate additional cash to multi-strategy and fixed-income arbitrage holdings that have shown accelerated risk-adjusted return generation.
What proportion of the fund is not in hedge funds?
While we reserve the flexibility to invest in long-only fixed-income orientated funds we have yet to utilise this ability.
What has been the historic volatility and returns of the fund, and its targeted volatility and returns?
Volatility has ranged from 2.78% at the fund's first anniversary to 4.63% in November 2004. Therefore the fund has successfully achieved both its targeted return (of 7% per annum) and its targeted level of 4% volatility. We generally concentrate on targeting levels of risk rather than return. Return estimates tend to be very unstable when calculated from past data or forecasts, and, as such, lead to very high and often avoidable levels of portfolio turnover. However, estimates of risk tend to be much more stable through time, and provide a more useful approach to achieving a fund's goals.
Which other funds does Forsyth run, and what has been their performance since their launches?
• Forsyth Diversity Fund (launched 1 November, 2002): 23.40% (figures to 30 November, 2004, in dollar terms).
• Forsyth Leveraged Diversity Fund (launched 31 December, 2003): 6.30% (figures to 30 November, 2004, in dollar terms).
• Forsyth Equi-Beta Fund (launched 31 December, 2000): 11.40% (figures to 14 January, 2005, in dollar terms).
What are sell signals for underlying managers, and how many managers have you replaced in the fund since inception?
Generally we will sell a fund if and when:
• it has suffered three successive monthly draw downs, implying to us the manager's strategy is not absolute return;
• its risk return profile has significantly shifted and no longer meets our investment criteria;
• our qualitative research raises concerns that cannot be addressed.
Once we sell a fund, it falls out of our investable universe, unless there is a strategy or manager change - in which case the fund is treated as a potential new investment. We have replaced 11 managers since inception.
What other risk systems/allocation limits do you have to control the riskiness of the portfolio?
The data within the hedge fund world, particularly at an individual on a fund by fund level, is, at best, weak. We are confident we have the techniques to appraise a hedge-fund manager's skill, but by the time we have collected enough data to provide high levels of confidence (that is, statistically significant) in that manager and his skills, the fund is generally shut! Therefore, selection of hedge funds needs to be performed almost entirely qualitatively. The control of their specific risk will be is performed in the same way.
The fund selectors rate each fund within its own asset class by their confidence in the manager's ability. Limits are then placed upon maximum weightings a fund in each rating can have. A small minimum restriction is placed upon all funds in the portfolio (to ensure their inclusion in the optimised portfolio), but the minimum for a highly rated fund is not greater than that of a fund with a lower rating because this would over-emphasize a mean-variance inefficient fund on the basis of confidence. Other than name diversification, the most obvious form of risk to control is that of strategy risk. The control process here is simple: limit the amount that can be held in each strategy. The exact nature of such limits needs to be discussed on a fund by fund basis, depending on one's overall objective. This is a simple form of a forced naïve diversification but it has been shown to achieve superior results over pure mean-variance optimisation (due to the non-normal distribution of returns in many strategies). Constraints are also imposed as to the underlying liquidity within the fund, enabling overall liquidity for investors. Overall, we have tried to create a portfolio construction process usable in a practical sense and which produces a well-diversified portfolio, capable of meeting its objective without being torpedoed.
What is your view on leverage of underlying managers, and is the FoHF leveraged?
We have no objection to using leverage by one of our underlying managers, but we place a particular emphasis upon reviewing and understanding terms under which leverage is provided to the manager. We have found in most cases the bank/prime broker "pulling the plug" on lending has been the main cause of fund failure. We carry out due diligence checks on the administrator, prime broker and/or other service providers to the target investee funds if we have not already undertaken this work for a prior investment. Forsyth Alternative Income Fund utilises its facility to use leverage of up to 100% of NAV for investment purposes, as well as to provide twice-monthly subscriptions and redemptions.
Rossen Djounov is managing director, institutional, and a principal of Forsyth Partners. A Bulgarian national, he graduated from the Technical University of Sofia with degrees in electronics and international trade relations and has an MBA in financial management from Exeter University. He also holds the IMC. Before joining Forsyth Partners in 1995, Djounov was responsible for FX and equity trading for First Private Bank in Sofia. Previously, he was involved in project appraisal, fund management and brokerage for investment fund Nova Bulgaria. At Forsyth Partners, he is an integral part of the investment management and research process, heading up the FoHF business.