Performance Comparison in CPPI Setup.

Now that we have seen how different hypothetical unlevered strategies can have different expected returns and volatilities, we can consider their performance when levered in the context of Constant proportion portfolio insurances - CPPI. When building a Constant proportion portfolio insurances (CPPI) note, structurers have to make several choices, which drive the performance and the risk of the trade. Naturally, the choice of the investment strategy and the maturity are the main factors. The choice of maturity will determine what fraction of the notional can be invested in the risky strategy: the longer the tenor, the more exposure to the risky strategy. Two other degrees of freedom that are available to structurers are the level of target leverage (TL) and the rebalancing multiplier (RM).
As explained earlier, the rebalancing multiplier determines the frequency of changes in leverage. While a relatively low number is necessary to obtain a dynamic leverage structure, it should not be too low, in order to avoid unnecessary transaction costs. As a rule of thumb, we have chosen RM = TL/5, such that, for a target leverage of 25, the leverage would be adjusted if the portfolio notional deviates from the target notional by an amount greater than 5 times the reserve. Structures with lower leverage would have lower RMs in order to mitigate the impact of the lower volatility. This prevents the frequency of rebalancing from falling too dramatically.
Increasing the target leverage has mixed effects on the strategy (see Figure 9.4). Increased leverage raises the likelihood of very high returns but also the probability of the trade being unwound. For most trading strategies, the shape of the relationship between TL and expected excess spread is concave, which enables the identification of an optimal leverage level.

Table 9.6 shows the performance of these same trading strategies when levered and in CPPI format. All returns here should be understood as annual excess return over LIBOR for the entire Constant proportion portfolio insurances (CPPI) note—not for the risky strategy only. Not surprisingly, the pattern of results for the CPPI structures follows those of the same strategies when unlevered. The best two performers in terms of information ratio are the long 10-year and the long/short. The equity versus credit trade and the long 5-year also produce excess returns but suffer from higher volatility. The CMCDS trades return sub-LIBOR performance, as expected. The last two rows of the table show the optimal target leverage for each strategy and the performance using historical spread changes over the past 10 years, assuming 2005 starting levels. Using these historical and simulated World Finances numbers, the long strategies fare best and all the other strategies have disappointing results.
Other Strategies
The list of strategies we have reviewed in this study is by no means exhaustive. Among the others we could have considered are tranche-based strategies and managed strategies.
Tranche-Based Strategies Must Consider Spreads, Defaults, and Correlation The diffi¬culty we see with tranche-based strategies is that one needs not only to assess the possible changes in spreads and the number of defaults during the life of the trade, but also to consider future changes in correlations. Correlation changes can lead to significant marked- to-market swings even when spread changes in the underlying credits remain moderate. Figure 9.5 shows how, at the beginning of May 2005, a fall in correlations led to strong underperformance of the equity tranche and the outperformance of the mezzanine 6 to 9 percent tranche, which rallied despite a sell-off in the index. A Constant proportion portfolio insurances (CPPI) referencing a long equity versus short mezzanine strategy on iTraxx or CDX would have suffered significant marked-to-market losses and, depending on its leverage, could even have been unwound (see Figure 9.5 for a graph illustrating changes in correlations and resulting changes in anche spreads). Tranche strategies (as well as CMCDS strategies) would also suffer from higher transaction costs than would straightforward long or long/short CDS strategies.
Further Levels of Sophistication Can Involve Noncredit Markets More sophisticated strategies combining credit and foreign exchange or credit and interest rates, possibly with an inflation hedge component, can also be devised. As in the case of managed trades, investors should carefully consider whether the added cost and lower liquidity of such sophisticated trades are adequately compensated for by either enhanced expected returns or lower risk.
Keeping It Simple Looks Best to Us By analyzing five different strategies over many default and spread scenarios, we have shown that simple strategies (long or long/short) tend to perform best across our forward-looking simulations. Had we included transaction costs, which tend to be higher with more complex strategies, we expect we would have found that the results favored the simpler strategies even more.
We highlight the crucial importance of the choice of strategy and of the target leverage for the performance of the Constant proportion portfolio insurances (CPPI) structure. While sophisticated strategies including tranches or CMCDSs may be grabbing the headlines, we think simple strategies are actually more appropriate for a CPPI setup.