Saturday, October 3, 2009

Analysis of Prior Question about Goldman Sach's Floaters

Added to Calculations the benefit of being paid par value after original post:

This post addresses the question raised in Saturday's post: GOLDMAN SACHS FLOATERS If I wanted to invest $5,000 in just one floater issued by Goldman Sachs, or tied by a swap agreement to a GS bond, which one of the following six would I buy now:

265 shares of GSPRD at $18.89 GS.PRD Stock Quote - Goldman Sachs Group Inc
336 shares of PYT at $14.88 PYT Stock Quote - Pplus Tr GSC 2 CT FL RT

This post assumes familiarity with the terms of each of these securities.

I can only discuss how I would approach the question, and others could easily reach a different conclusion. I am going to assume that the average 3 month Treasury Bill rate and 3 month Libor rates for the next 25 years will be 4.5% and 5% respectively.

Before starting the game, I am going to assess a few rewards and penalties.


I am going to add .25% to the GJS average per annum rate and .15% to PYT and GYB representing the value to me of their superior claim to dividends. Since GJS is tied to a senior bond, I am giving it a little extra. This is sort of an insurance question. What is it worth to me to have more protection in the event of bankruptcy by owning GJS compared to the junior bond or an equity preferred, with the later two most likely being rendered worthless in a bankruptcy whereas a senior bond holder may recover something. 

The second component is what is the value of a cumulative dividend or a payment which can not be deferred? The GJS owner does not have to worry about a deferral, just bankruptcy. The owner of equity preferred has to be concerned about an elimination of a non-cumulative dividend. The junior bond holder is in between, with GS having the right to defer payment for up to five years. 

Another consideration is that the most likely scenario would be similar to Lehman, where all security holders lose their distributions at the same time, that is, going from paying everybody to paying nobody in one fell swoop. But that is not the only scenario. It is possible that at some future date in the next 25 years, GS may eliminate its common and equity preferred dividends, and pay all its bond holders, or defer the junior bond payments and continue to pay the senior bond owners. So this is just a question that each individual has to answer for themselves. I am saying that the junior bond will be given a credit of .15% and the senior bond .25%.

2. Penalties: Now, I have another problem. The annual average interest rate that I want to use is based on historical rates which include some high Libor and T Bill numbers. But, the synthetics have a maximum rate. GJS tops out when the 3 month T Bill hits 6.6%. PYT will top out at 7.15% 3 month Libor. GYB tops out at 7.4% 3 month Libor. I could run a precise calculation to assess this penalty but I do not want to spend the time. I would do that calculation by, for example, substituting 6.6% for the monthly 3 month Treasury Bill rate whenever the historical monthly rate exceeded that number to come up with a new average just for GJS. Instead, I am just going to eyeball and guess. GJS is going to penalized .3% , PYT .25% and GYB .2%. for their respective maximums.

Now, I am going to assess a special penalty against GJS for having such a low minimum which comes into play now. That penalty is .1%

So, GJS comes out of this process with a net penalty of .15%
PYT has a net penalty of .1%
GYB has a net penalty of .05%

I am not going to assess rewards and penalties based on the different tax rates. I will just assume that the securities are all bought in a retirement account where the tax issues disappear.

What I am now going to do is calculate the average annual interest payment and the overall 25 year payments for each security with all of the assumptions outlined above, actual results will of course vary (calculations assume an average of 4.5% on the 3 month T Bill and a 5% 3 month LIBOR, adjusted for the penalties):

GSPRA .75% + 5%=.0575% x. $25= $1.4375 x. 249 shares= $359 in 1 year, $8,950 in 25 years

GSPRC .75% +5.%=.0575% x. $25=$1.4375 x. 254 shares=$365 in 1 year, $9,125 in 25 years

GSPRD .67% +5%= .0567% x. $25=$1.4175 x 265 shares=$376 in 1 year, $9400 in 25 years

GJS .9 + 4.35%= .0525% x 25=$1.3125 x 385 shares=$505 in 1 year, $12,625 in 25 years

PYT: .85% + .049%=.0575% x. 25=$1.4375 x. 336 shares=$483 in 1 year, $12,075 in 25 years

GYB .85% + .0495%=.0583% x. 25=$1.4575 x. 321 shares= $468 in one year, $11,700 in 25 years.

I was somewhat surprised by this result and would have picked PYT 1st, GYB 2nd and GJS 3rd before starting this process. I knew that the equity preferred floaters would bring in the rear.

I may need to check these calculations later.

Added 8:32 P.M 10/3

I neglected to calculate the additional amounts that will be paid at maturity for GYB, PYT and GJS, assuming GS survives to pay par value:
GJS= $4,620 ($12 spread to par value x 385 shares=$4620)

PYT = $3,400


This increases the GJS advantage. The equity preferred stocks have no maturity date and are selling closer to par value anyway. I currently own PYT, GYB, GJS and GSPRA.


  1. Thank you for a very useful analysis. GJS wins my favor on another count: it trades at a slightly higher volume than PYT.

    I own 22 shares of PYT. I placed a limit order for 80 shares and it was partially filled by Scottrade, who claimed that was the best they could do on that particular trading day. I paid the $7 commission but got fewer shares than I wanted. If I wanted more shares I would have to wait for another day and pay another commission, or place an order for 100 shares and specify "all or none" (which is not an option on odd lot sizes.) It was a good lesson for me.

    I was wondering what your general strategy is for buying or selling thinly traded issues. Do you ever use market orders?

  2. Cathie: The analysis on the GS floaters is price sensitive on the initial buys. So a different result might be reached by buying GJS at say $13.75 rather than $13 and PYT at $14, though I have not run that calculation. Another important factor is the assumption about future LIBOR and T Bill rates. Generally, if the average short rate for the next 25 years is higher than the numbers that I used, that would increase the advantage of GJS, whereas a much lower number,say 3% or 3.5%, would cause the GJS advantage to shrink or even disappear on a current yield basis. Another factor is that I view a .85% spread over 3 month Libor to be better than a .9% spread over the 3 month T Bill, so it would not take a large price narrowing between PYT and GJS to tilt the advantage toward PYT.

    Also, there is a current yield difference that makes GJS look worse, something like a 2% current yield at the $13 price, due to the extremely low T Bill rates now and no minimum guarantee (other than what is in effect created by a float .9% over a zero T Bill rate). So any buyer would have to be willing to suffer that hopefully temporary yield disadvantage for at least several months, maybe longer. The advantage of the GJS profit at maturity would remain irrespective of the future course of the Libor and T Bill rates, assuming the lower purchase price, since the profit at maturity is a known constant dependent only on the original price paid for the shares and the survival of Goldman to 2033 or 2034.

    I had a limit sell order for 50 shares filled the other day with just 1 share. I then had to sell the other 49 with a separate market order, and pay two commissions. That is one of the problems with odd lot orders. I have also had a odd lot market order for PJS fill a quarter or so above the highest trade that day. That is another problem with them. Both happened when the orders were routed through Knight. 95% of the time I hit the bid or ask price last displayed when the odd lot order was placed.

    I use market orders and limit orders. Sometimes I mark the limit order all or none, but this makes it more difficult to fill on a lightly traded issue like these trust certificates. Some of my best fills on the TCs have been GTC limit orders well below the market. I had 100 hit on JZE at the absolute low of $12.5 last fall. If I am buying a stock like GE or PG, I will sometimes use a market order since the spread between the bid and ask is just a penny. Or, I will sometimes place a day limit order for those kind of securities a quarter of more below the current trading activity. So a lot depends on the security, the volume of trading, the bid/ask spread, and the movement up and down on a particular day.

    On Friday, I bought 100 GJS with a market order. There was at least 500 on the bid at $12.95 and 500 or more on the ask at $13. If I placed a limit order, I would not have been able to buy the shares at $12.95. I could have put a limit order at $13, probably would have gotten the shares without risking paying more with a sudden price move in the ask price which has happened to me. Soon after I bought, the bid moved up to $13 and the ask price jumped to around $13.45. I would not enter a market order with a spread greater than 10 cents.

  3. Thank you very much.

    I'm assuming the survival of Goldman Sachs and its affiliate, the US Treasury (ha ha), but not necessarily my personal survival, until at least 2034.

  4. Cathie: Some would say Goldman Sachs and its affiliate, the U.S. government, but I would count on a least another 30 years of personal survival. On the bright side, assuming GS remains prosperous, a holder of GJS or PYT will not have to wait until 2033 or 2034 to receive par value, as the current discount to par value will start to narrow in the coming years and those securities may consequently be near par value within four or five years of maturity, say 2029, a mere 20 years from now, which is the case for GJJ now which matures in late 2013. I noticed the web site today about GS: Someone posted a link to my blog there, and I suspected that those who frequent that site may not possess a favorable view of GS, something about 666 in the web address tipped me off.