Log in

No account? Create an account
Previous Entry Share Next Entry
Mortgages in theory, part 6 of N
Let's talk about bonds -- which Mortgage Backed Securities are one form of -- and the bond market.

A bond is a way of borrowing more money than any one lender is willing to lend. If it costs $100M to build a new road, or a new factory, or to buy a new piece of equipment, most companies or governments don't have that cash lying around, and most banks are mindful of the adage "If you borrow a million from a bank and can't pay it back, you are in trouble. If you borrow a billion from a bank and can't pay it back, the bank is in trouble". So the company or government involves issues bonds, allowing a large number of people to lend a little money each and thus they can raise the $100M they need.

Each bond in a bond issue is an identical obligation from the company, and are designed to be "negotiable instruments", i.e., they specify an unconditional obligation to pay the bond holder fixed sums of money based on a prewritten schedule, and they allow the bond holder to sell the bond to others, thus transferring the rights to the payments to the new holder. The bond market allows investors to buy and sell bonds they own, which in turn makes bond investors more willing to invest, since they know they can sell them later if they need the money for some other purpose.

As an example bond, here's the beginning of the text on a railroad bond issued in May 1865 (picture at link):

The City & County of San Francisco in the State of California will pay to the Western Pacific Rail Road Company or to the holder hereof ONE THOUSAND DOLLs thirty years from the date hereof with interest thereon at the rate of SEVEN/ percent per Annum pay-able semi-annually on the First days of May and November of each year upon interest Cou-pons hereto attached, both principal and interest payable at the City and County of San Francisco in United States Gold Coin Dollar for Dollar.

This bond demonstrates all of the features of a "fixed-rate coupon bond": It states an amount to be paid ($1000) when the bond "matures" (1 May 1895), and a fixed interest rate (7% per year) payable in semi-annual "coupons" (each worth $35, when due) until maturity. Although most, if not all, bonds traded these days no longer have paper certificates, you can see in the picture of this bond that there really were coupons attached to the bottom of the bond, suitable for presentation and redemption. Most bonds in the US are issued in denominations of $1000.

For fixed-rate coupon bonds (which are probably the most common type of bond), the most important factors in determining the value of the bond are its maturity date, it's interest rate, and the "quality" of the issuer -- essentially, the credit rating of the issuer. All other things being equal bonds from higher rated issuers are going to be worth more than bonds from lower rated issuers.

In general, the value of a bond can be expressed either as a price (how much someone would be willing to pay now for the bond) or as a yield (what is the effective interest rate of the bond, based on its current price). The calculation for the yield can be complex (not to mention that there are several different types of yield, each with their own methods of calculation) but the basic idea tends to be to compute a "present value" for the bond, and divide that by the price. A high price gives a low yield, and vice versa.

The calculation for "present value" involves the time to maturity, the interest rate, and a subjective "discount factor" based on what the investor sees as alternatives to the bond and a perception of what's going on (economically, or with inflation, etc) in the future, among other factors. Since all of these factors tend to be external to the bond, a bond trader will tend to use one discount factor for computing the present value on all his bonds. Since all of these factors tend to be time-varying, the present value on his bonds will tend to rise and fall as well. The yield on the bonds would also rise and fall if the price of the bond didn't fall and rise to compensate. Roughly speaking, the interest rate on new bonds (and thus the target yield of existing similar bonds) is determined by the discount factor of the day, so yields The price of a bond tends towards the maturity value when the bond is approaching maturity (nobody is going to sell a bond for $950 or buy it for $1050 if it's going to mature and pay out $1000 in a month).

Because the price of a bond on the bond market is rarely the same as the price of a newly issued bond, the yield is a major factor in comparing two bonds: A newly issued bond with an interest rate of 7% is roughly equivalent to an existing bond by a similar company with a yield of 7%, even if the interest rate on the existing bond is higher or lower. This tends to force bonds that are similar to each other in terms of risk and maturity to have the same yield: if two similar bonds differ in yield, investors will sell the lower-yielding bond (lowering its price, raising its yield) and buy the higher-yielding bond (raising its price, lowering its yield) until the two bonds have the same yield.

When comparing two bonds of similar maturities at a given point in time, there are a lot of similarities: Both bonds are subject to the same overall economic and market forces, both bonds have the same amount of time to pay out interest, both bonds have to compete against the same investment alternatives. The major difference between the bonds is the perceived risk of failure of the issuer. Investors want to be compensated for higher risk. A bond with a 4% chance of default would have to have a yield of at least 7.3% in order to match the expected return of a risk-free bond with a 3% yield. As such, if it had a lower yield, investors would sell it until its yield rose enough.

So bond yield is very much related to investment risk: the higher the yield, the higher the risk. And two similar bonds with the same risk profile will tend to have the same yield. Issuer credit ratings are a measure of risk. So bonds with the same credit rating will tend to have the same yield, and higher rated bonds will tend to have higher prices/lower yields than bonds with a lower rating. This also implies that when the credit rating on a bond or an issuer changes, the bond is no longer selling at the right price for its new rating, and will have a sharp price fluctuation to bring it in line with its new risk class.

Bond ratings are, unfortunately, not clear-cut, easy to work with numbers. No ratings agency is going to come out and say they believe this bond has a 4% chance of default, and that bond has a 1% chance. Rather, they give letter-grades, like AAA, or CC, with no stated correspondence between rating and chance of default beyond the rather general idea that the better the rating, the lower the risk of default. Generally, however, ratings are divided into two broad categories: "investment grade" and "high yield", or roughly low risk and high risk. I'm fairly certain that the dividing line is somewhat arbitrary, and there probably isn't much difference in risk between the lowest-rated investment grade bonds and the highest-rated high yield bonds.

But there's a major catch: Many "institutional" investors, portfolio managers representing incredibly large pools of (usually) other people's money, are constrained to only buy "safe" investments, meaning investment grade. And because of the side of their portfolios, they only buy and sell in huge quantities. What this means for a bond is that when the bond gets re-rated across the line from investment grade to high yield, massive amounts of the bond can be dumped on the market by institutional investors unloading their positions, and the price plummets.

The flip side is that most of the money is from institutional investors, so folks creating bonds want very much to get their bonds rated investment grade, so (a) the price will be high, and (b) there will be a lot of institutional investment money available to buy them. One of the major things which happened in the mid-2000's was investment banks found ways to structure bonds so that they looked like they were good investments in order to get the coveted investment grade ratings.

The crash happened, in part, when bondholders started to realize that the bonds they held were much riskier than they thought, and everybody wanted to sell but no one was buying.

  • 1
I agree with everything you say here. How often does that happen?

I dunno, we rarely comment in each others journals. I guess rampant disagreement has gone unnoticed.

Really good info.

Thanks for sharing...

  • 1