January 25, 2009

If your ideas is so good, why aren't you rich already ?

If this works then why are you telling us about it and not doing it yourself?
There are many responses, including Wilmott's response:

1. I don't have the ability to do it myself, this is my marketing pitch, want to back me?

2. Ideas are cheap, I know which ones are good or bad but not everyone can tell the difference

3. Do you know all the barriers to entry? $1million in lawyers fees to set up a fund, months of software writing, years of knocking on doors trying to raise money. Forget it!

4. This is a great idea, but I've got better

5. I don't want to spend the rest of my life doing this, even if it is profitable, variety is the spice of life

6. I did, and now I've retired or, more simply, I've got enough money already

7. My lawyer/doctor/wife says I mustn't

Continue reading "If your ideas is so good, why aren't you rich already ?" »

October 30, 2006

Tranched mortgage pools

The fun is tranching pools of mortgages into different securities.
You take a package of mortgages, preferably from different areas
of the country, and then you create different tranches with
different credit qualities, and then one zero coupon that bears
all the residual risk.

The first tranch, naturally, gets all the guarenteed income stream
(mortgage payments) and bears *no* (or actaully, very little)
prepayment risk (it is good to maturity).

The next tranch gets income stream, and bears some prepayment
risk (if there are a lot of prepayments, it gets a swath if the other
tranches are fully repaid), and on down the line.

The last tranch before the zero gets income, but bears risk
if the income falls short (mortgages default), and also bears
*the most* risk for prepayment (it has a call option owned
by the mortgage holders, they can repay the loan if the
interest rate changes). If you own this last tranch, you have
lots of duration and gamma risk, whereas if you own the
first tranch you have a very different profile.

The 'residual zero' tranch is practically binary.

Either it pays off, or it defaults and gets the (last) bit of
recovery value in liquidation. Lots of folks use equity
models to calculate expected return on these.

Freddie and Fannie do this, but also Morgan Stanley and Goldman.
If you want to play, you can call them and they can cobble toghether
a structured deal that will match pretty much any flavor you want.
Some of their customers are developers who are highly exposed
to one geographic area (say, Toll Brothers) and need hedging.

A good intro book is Collateralized Mortgage Obligations
by Chuck Ramsey and Frank Fabozzi
. It goes beyond CMOs and
talks about a lot of the risk horizons and what traders of these
do and play with. Lots of former interest rate traders apparently
do well in this field.

As far as point three: you are talking 'real options' theory here,
and you have to look for a lot of asset value delta to overcome
the (identified) barriers to exit and entry and transactions costs.
I have yet to see a good book on the real options of real estate,
but the ones that come close deal with mineral rights and land
and mostly were developed for energy exploration. Not really
'commercial' or 'residential' real options.

Continue reading "Tranched mortgage pools" »

June 8, 2006

Wilmotters' Quant Quantitative Finance notes

Paul Wilmott, Emanuel Derman and Domicic Dominic Connor (DCDC)
write Wilmotters: Paul, Emanuel, Dominic.

July 15, 2005

OAS measure of yield

OAS measure of yield has been introduced to accurately price callable
bonds but is also used now as a measure for bullets' yield.

1. For bullets, it is more accurate than yield to maturity (YTM) as

a. You use implied forward rates instead of the yield to maturity as
a reinvestment rate.

b. You discount using the zero cpn curve instead of the YTM
Even more, you calibrate your forward rates so that the PV of yor
coupons match the market values.

2. For callable bonds and MBS, the YTM measure also assumes holding
till maturity which is obviously inaccurate so the OAS uses binomial
tree which takes into account the contingency of the future coupons.

The OAS is a constant spread to the whole discount curve you use
( e.g. treasury) which would result in a PV equal to the market price.

Continue reading "OAS measure of yield" »