Trading

The Need for a "Hurdle"

It is not enough for one issue to be marginally better than another to justify a trade; there must be sufficient prospect of excess returns that may be earned by execution of the trade to justify the certain costs associated with it. These costs may be summarized as

• Commissions
• Possible Capital Gains Taxes
• Settlement fees
• The bid-offer spread
• Possible market impact

Therefore, a minimum valuation pick-up is required when comparing two issues for possible swaps and this minimum is varied by HIMIPref™ in accordance with the particulars of the transaction. It should be noted that the requirement for a minimum pick-up helps lessen the chance of whipsaws, which are a feature of most practical implementations of quantitative strategies.

Risks Based on Single Issue Characteristics

A number of measures have been derived for use within HIMIPref™ that, while logically being of importance when considering a trade, cannot logically be related to expected future returns (which are discussed in the Overview of Valuation). Examples of such factors, referred to as "Penalties" are:

• Option Doubt: Valuation is reliant largely upon the analysis of expected future cash flows, but how reliable are these projections? If there is a high degree of variance in the characteristics of the instrument depending upon the potential exercise of embedded options, we should be more hesitant to rely on this analysis.
• Pseudo-Convexity: To what extent will the characteristics of the instrument change when the price is varied? A negative pseudo-convexity implies that the instrument will become less attractive with any change in price and therefore should be avoided to some extent as it might change from a "buy" to a "sell" with a relatively small change in price.

Risks based on the characteristics of a single issue are, by definition, calculable in isolation and may therefore be reported on the Report Summary

Risks Based on Issue Pairs

Risks of this type are reported as part of the reporting package with respect to trading analysis. They may be summarized as follows:

• Risks based on Yield Curve Characteristics : these risks are treated differently in the "Issue Method" and "Portfolio Method" of trade analysis and are discussed below
• Risk due to Valuation Difference Heterogeniety : This risk is based on the understanding that Valuation is the sum of many independent elements, which do not all necessarily agree - one example is whether one should prefer the higher-yielding member of a pair, or whether the potential for capital gain outweighs this consideration if the lower yielder is trading above its average value. Such a conundrum is an example of Valuation Difference Heterogeniety and the hurdle rate for trade recommendation is increased according to the degree of heterogeniety in the valuation elements.
• Risk due to Over-reliance on Few Factors : The valuation elements are examined pairwise between the putative sale instrument and the purchase. If any individual element exceeds a threshold, the amount above this threshold is effectively reduced in the ensuing analysis.

Risks Based on Portfolio Holdings

When trading for a portfolio of shares, it is necessary to account for the other positions before executing a trade. There are two adjustments made to the hurdle rate based on position:

• Value-Size Adjustment: As the position becomes larger relative to the average trading volume of the instrument, the more difficult it will be to dispose of and capitalize on transient market pricing anomalies. Hence, an adjustment is made to the hurdle rate such that larger positions will be traded with an increased reliance on yield factors, rather than potential capital gains
• Holdings Adjustment: If more issues are held than is considered optimal for the portfolio, trades into new positions will not be forbidden, but will be required to pass a stricter test of presumed profitability in order to be executed.

Risks Based on Yield-Curve Factors

The yield curve premia discussed in the Overview of the Yield Curve Calculation were chosen as being factors which a rational, prudent and informed investor could wish to use in order to determine whether or not to proceed with an investment; it is hypothesized that instruments sharing a like profile in terms of these factors will, to a greater or lesser extent, have returns that are more similar than those with which the characteristics are not shared.

Hence, the concept of "Risk Space" was developed, with the idea that risk could be represented by an n-dimensional space, with the probability of like future returns being estimated by the "Risk Distance" between two points representing these issues in Risk Space.

Each of the elements giving rise to Yield Curve Premia is assigned a dimension in Risk Space and the calculation of distance in any given dimension is optimized in the course of simulations. For simulations using the issue method, the distance between the two issues is used to adjust the hurdle rate; when the portfolio method is used, the relevent measurements are of the portfolio's position in Risk Space relative to the Index, before and after the trade being examined.