Structured notes and CDs are often vilified as being arcane profit generating schemes for issuers and loss generating money-pits for hapless investors. Having worked as a quant at UBS where I helped structure, model and price these products, my opinion tended more towards the middle. I understood the products and how issuers use arbitrage free pricing models to try to hedge a profit (after giving the broker their commission), but I never had a chance to analyze how well these structures pay off for investors.
So, after leaving the bank, I decided to find out whether there are any good investments in structured products. What I found has made me cautiously optimistic that there are a good number of opportunities out there. I voice this opinion with the large caveat that simply reading a prospectus will not give an investor enough understanding of a product's risks and possible rewards. Some fairly detailed analysis must be done to really understand each product.
Given that need, I plan to write a series of articles detailing how one can perform that analysis to discern value along with giving detailed results for a number of products and, depending on reader interest, also provide spreadsheet examples to help investors get started in their own analysis.
The next section covers a discussion of the valuation of 12 products including commentary related to how the analysis was performed. After that section, a brief discussion of 2 of those products is given in order to demonstrate the need for detailed analysis before investing in a given product.
I have analyzed 12 US equity-based structured CDs and notes that are to be issued in the next couple of weeks. Given current time constraints for performing each product analysis, I sampled products listed on the InCapital and Raymond James websites. As such, the sample is not a representative sample of all the currently to be issued equity structured CDs and notes.
However, the following table shows that a wide range of products is covered. In the types column, CD is self-explanatory, PP is a principal protected note, PPP stands for partially principal protected, and NPP means non-principal protected. Of course only the CDs have their principal protected from losses due to an issuer default, and that only up to the FDIC insured amount of $250k. For notes, the principal is usually unsecured senior debt of the issuer (check the prospectus to make sure for a given product). The CUSIPs are given to allow finding the prospectuses using the SEC's Edgar search engine for the notes while the prospectuses for the CDs can be found on the aforementioned dealer websites. The descriptions are somewhat illuminating but really will not make sense unless one goes through a few prospectuses to understand the payoffs. Payment is fairly self- explanatory as some of the products pay coupons at regular intervals while some only have a payoff at maturity.
All but one of the products' underliers consist of the two major U.S. indices, the Russell 2000 (RTY) and the S&P 500 (SPX). The first CD (structure 1) consists of a basket of 5 equities (AAPL, ABBV, MSFT, T, VLO). The term column is self explanatory except that 3 of the products (4,6,10) can be called early by the issuer but not before the 1st year (1y NC). Also of note, structure 12 has an autocallable feature coinciding with the annual coupon payments. The issuers listed are all among the larger issuers of structured products.
In the table below, I show the results of arbitrage free pricing of the products as of February 12th. As mentioned previously, these prices are around what the issuer would expect it will cost them to hedge the product. I use industry standard models to generate the prices but my valuations almost certainly differ from the actual issuer price as the traders have some freedom in how they mark their implied volatilities. Also, I tend to err towards conservative pricing by adding relatively large issuer credit spreads over the risk free discount curve and by taking the lower end of the listed possible range of payments given in the preliminary prospectuses.
As one can see, all but one of the prices is in the low to mid 90s based on a notional value of 100% and is consistent with the range of anticipated prices listed in the prospectuses. My valuation of structure 10 seems a bit low and I suspect that they are assuming lower implied volatilities than the values I generated from option prices. As one can see, even paying a 2-3% commission to the brokers, the issuer has good room to make a profit.
In the future, one would hope that competition and investor awareness will pressure the issuers to increase the payoffs of these structures and thus drive their prices more towards 100% and increase values for investors. Also included in the table are the annualized volatility of the discounted payments resulting from the pricing simulation along with the payment-weighted duration of pricing simulations. One should compare these with the term column in the previous table. For the callable and autocallable structures the difference can be quite large.
Before delving into results, an explanation of what is meant by historical analysis is required. To perform the historical analysis, a set of historical issue and relative analysis dates is generated. For each product, the payoffs of the structure are calculated as if the structure had been issued at each of these historical issue dates. Historical data of the underliers is then used to determine the payoffs of the structure, and thus build up a past performance record spanning 10 years to ensure coverage of a full economic cycle.
The CAPM (Capital Asset Pricing Model) is then used to determine a spread over the risk free rate at the product's calculated duration. Also, as the structure is subject to issuer default, a credit spread is added. This cumulative spread can then be added to the risk free rate curve to generate a structure discount rate curve. This issuer discount curve allows the calculation of historical structure values at all the historical analysis dates. And finally, these historical structure values can be averaged to determine the average historical value of a given product as well as other standard investment metrics such as an annual return and Sharpe ratio.
The results of the historical analysis are given in the table below. As the CAPM has been used to determine the discounting, a value of 100% implies that the return of the structure is adequate to justify the level of risk as indicated by the structure volatility. So values above 100% indicate superior returns to that of simply investing in a combination of the market and a risk-free bond of the same duration as the product. In this context, the "market" is the S&P 500 or Russell 2000 as appropriately indicated by the underliers for the given structured product.
As one can see, all but one of the products has a historical relative value of above 100% with 4 relatively close to 100%, leaving 7 (structures 2,4,6,7,9,10,12) with values superior to the market portfolio. The calculated product returns are also quite good.
Using historical averages for underlier returns and volatilities as well as the previously determined structure discount curves, the projected performance of these structures has been calculated and is given in the table below. In comparison to their historical values, the expected performances of the structures in this scenario are not as good. The one real exception is the relative value of structure 4, but comparison of the duration indicates a much larger duration for the scenario simulation which, given the above market performance, naturally leads to higher values as the product duration increases while the return remains nearly constant.
Insights into the valuation of structures 3 and 12:
While the relative performance tables above are useful suggestions as to which products may be good investments, to truly understand how the product may perform I believe that one of the most useful things one can do is plot the historical relative values as a function of their historical analysis dates along with the performance of the underlier.
The graph below shows the historical performance of structure 3 as a function of the historical analysis date. This Cliquet structure adds the quarterly performance of the S&P 500 subject to a maximum of 4% per quarter over 7 years. If this final sum is above 8%, then the sum plus initial notional is paid at maturity; otherwise 8% plus notional is paid. As the figure indicates, a relatively large number of positive quarterly returns must occur in order for the structure to pay the above minimum return. For this product, the sum of returns above 8% only really began around 2008. Prior to that the product always paid the minimum which indicates how rare it may be to have an above market value for this product.
Structure 12 is another interesting example. The structure is autocallable once per year over a period of 4 years if the lesser performing of the S&P 500 and the Russell 2000 is above it's original trade date value. If it is called at the end of the 1st year, it pays notional plus 12%; the plus amount increases to 24%, 36% and 48% on the 2nd, 3rd and 4th years respectively. If it's not called at the end of the 4th year and the lesser performing of the 2 indices is greater than 70% of its original value, then it pays notional plus 10%. Otherwise it pays the lesser performing of the 2 indices relative to their original value. Thus one could possibly lose all of his original principal.
As the graph shows, over 10 years of issue dates ranging from 2002 up to 2012, the loss of original principal never occurs, as neither index is below 70% of its original trade date value after 4 years. Thus, the structure always has a relative value above 100% of notional. That's not bad but definitely not a guarantee of future 4 year performance as indicated by the scenario valuation of structure 12.
Another point of interest is that the 5 distinct cumulative payoffs indicate the 5 different payoff situations that occurred -- the lowest value being when the structure was not called at the end of 4 years and paid notional plus 10%. The other 4 levels indicate when the structure was called after the 1st, 2nd, 3rd and 4th years. Thus, as one can see, the structure would mostly have been called at the end of the 1st year.
On a relative value basis, there do appear to be a surprising number of possibly good investments in this sample of structured products. Of course, given their complexity, detailed analysis must be done to really understand how the structures will perform relative to how their underliers perform. In future articles, I plan to provide in depth analysis of specific products as well as give more valuation results of currently to be issued products on a regular basis.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.