Brokered CDs have this in common with a Cadillac automobile: the minute you've bought it, much of its market value is gone, irretrievably. This loss is completely distinct from any losses or gains that are due to interest-rate fluctuations. Here I present a method of estimating the magnitude of this (non-interest-related) loss at any one time.
To begin, here is a summary of findings from data available to me: while brokered-CDs (unlike their bank-bought cousins) are touted as marketable at any time, their market values are in fact substantially below those to be expected from the prevailing interest rates at any given moment. Second, there is strong evidence that CDs issued by lower-ranked banks suffer significantly sharper losses in the secondary market than issues from higher-ranked banks.
There is a straightforward procedure of assessing the magnitude of these non-interest effects: the widely available algorithm for finding the "price" of a bond. (See Appendix A for a description of the procedure). This "price" is the amount for which an older issue would sell if it were valued according to currently operative interest rates; in other words, it is the price that an investor would pay to obtain an interest rate equal to that of an original-issue security of identical characteristics. I will call it a correct price to distinguish it from the actual market value, the price at which this older issue actually brings on the secondary market. This actual price is furnished by brokers to their customers on a regular basis for each security that is held in a customer's portfolio.
Now the ratio between the actual and the correct, the A/C ratio, may be used as a measure of resale value of a bond or CD. As one would expect, Treasury issues have an A/C of just about 1.00, but bank-issued CDs have A/Cs that are lower, sometimes much lower.
Unfortunately, the market-side data needed to perform the calculations of A/C ratios are not publicly available in the case of brokered CDs. The secondary market in these securities is opaque; transactions are not reported in the press or anywhere else. If you wish to sell your CD on this market, you call your broker, who calls "someone" but won't say whom, and then comes back with a price for you. There are businesses that price these issues, i.e. give estimates of market value, but they do not report these data to the public. Brokerage firms receive such estimates and use them to give market values in monthly statements to their customers. In other words, an individual customer can know the (estimated) market value of his CD at any one time, but he does not know the value of other issues.
Since I own four different CD issues, and also a number of Treasury issues, I have been able to make a modest analysis of the second-market behavior of these securities. (See Table I). But it is a very modest analysis. To make a more conclusive study, I would need a great deal more data. My brokerage firm has these data, but it has steadfastly refused to divulge them, even during discovery proceedings in an NASD arbitration case.
Table I summarizes my results. Since financial markets fluctuate from day to day, these results can show only a static snapshot of what is actually a dynamic movement of prices. The date chosen is November 7, 2003. There is no reason to suppose that this date was in any way atypical.
The last column, Resale Value, is the most crucial. It shows that these already-issued CDs all sell below prices that initial-issue CDs would bring. This is in contrast to the Treasury issue shown in the table, which sells at just about the price of an initial issue.
The first column in the table contains a brief description of each issue. (The second of these issues, a Treasury note, is included for purposes of comparison and contrast to the CDs).
Bank ratings, column two, come from BauerFinancial. Bauer "recommends" banks of four or more stars. A five-star rating, the maximum, is here imputed for the Treasury issue.
Correct Price, column three, is computed as shown in Appendix A.
Actual Price, column four, was obtained from my broker's statement, on line, dated November 7.
Current rates for Treasury Bills, Bonds, and Notes are listed in the financial pages of the New York Times daily.
Unless they are held to maturity, brokered-CDs have the following disadvantages:
1. While a secondary market exists, it seems to guarantee a loss to anyone who sells these instruments. My broker (Vanguard) did not disclose this feature when I bought such issues.
2. The quality of the issuing bank is a crucial factor in the secondary market of CDs. My broker neither disclosed this fact to me, nor, by his own admission, considers it necessary to inform his customers of the credit worthiness of the issuing banks.
Finally, it may be asked why credit worthiness should influence the market in government (FDIC) insured investments. Upon reflection, the answer is not difficult to fathom. FDIC insurance is limited to $100,000 per investor at any one banking institution. But this is a paltry sum for the large investors who seem to dominate the secondary market. For such players, these CDs are essentially uninsured instruments and are priced in accordance with the credit standing of its issuers.
"Bond price" is one of the built-in functions in the HP-12C Financial Calculator, among other such machines. The owner's handbook for HP 12C, in turn, makes reference to "Standard Securities Calculation Methods" by Spence, Graudenz, and Lynch.
The procedure uses the following parameters:
For any date D between date of issue and maturity:
I = prevailing yield to maturity on date D; that is to say, roughly, what a new security bought on this day would yield. In effect, this value is an estimate of the new-issue value of a security on day D
PMT = coupon rate
D = current date
M = maturity date
With these four values plugged into the algorithm, the machine will yield "Price". This "price" must be understood to mean a theoretical price, that is the price the issue would bring if, and only if, only if, only interest rate fluctuations had effected the price. All prices are based on a par value of 100.
An important ingredient in this calculation is a determination of the current interest rates of CDs, i.e. the "I". Such rates are not directly available because, unlike Treasury issues, there is no publicly available information on interest rates applicable to CDs that mature throughout the year. For that reason, I estimated these values as follows: For any given CD, I determined its yield Y at day of its issue. I next determined the yield X of a Treasury of identical maturity that was sold at or close to the same date. I then multiplied the ratio Y/X by the yield of a Treasury selling at date D.
A ten-year CD is purchased on January 15, 2003, maturing on January 15, 2013. It has a coupon rate of 4.15%, payable semi-annually. On November 7, 2003, this kind of security, with the same maturity date, would yield 4.35%. What "should" the original security bring on the secondary market ?
Here I=4.35, PMT=4.15%, D=November 7, 2003, and M=January 15, 2013. Plugging these values into the machine resulted in Price=98.49.
|Bank Rating||Correct Price (C)||Actual Price A)|| Resale Value
10 years. 4.15%
| US Treas.
10 years. 3.625%
| Capital One
5 years, 5%
10 years, 4.1%
| Home Ln Indl
five years, 4.75%
All price values as of November 7, 2003
Correlations: Bank Rating vs. Resale Value
CDs only, without Treasury, r=.944
Brokered CDs: Caveat Emptor
Seniors are the most vulnerable
Click here to go to home page of Werner Cohn
Have you heard about my Early Companions ?