There are three effects:

The first is the credit risk in prime mortgages is very small. Prime mortgages tend to have default rates of less than one percent. Ordinarily, mortgages have negligible credit risk to start with.

The second is that sub-primes have high interest rates. This is important because if you get a prepayment, then you can reinvest the money in treasuries, but you lose the high interest rate that the sub-prime mortgage had originally.

Third, I strongly suspect that lots of people were using large amounts of leverage, which made even small differences from expected behavior to be deadly.

Guest: Credit risk scenarios should have been taken into account in the original pricing of the spread and the associated return on capital.

Because prime MBS’s have low credit risk, it’s traditional for the securitizer to assume this risk and so credit risk becomes prepayment risk. It makes it easier for the buyer of the security since it folds credit risk into prepayment risk models. Ordinarily it is a very small correction, and it makes life easier for the buyer.

When there is significant credit risk, it messes things up since it makes it difficult/impossible to distangle things.

Guest: your noted that some i-banks saw this coming, which implies “some” did not. who - setting bear aside — is a likely candidate to be among those who did not?

Mentioning specific companies is a big no-no that will definitely get me in trouble with the my employer and with the SEC.

However, the information for this is pretty public, and you can get a good idea of bank exposure by looking at the SEC filings and seeing what the banks are investing in. The other thing to do is to look at the management team for the banks, google a bit, and see how the people have behaved in the past. If you have a CEO and upper level managers that did well as far as risk managment in previous crises, they are likely to do well again.

My general sense is that as far as investment banks go, lightning has already struck and there isn’t going to be any new bad news as far as the investment banks go.

Finally, unexpected prepayments that are made from a fixed rate mortgage portfolio when interest rates are higher than booked rates actually add value to the return on interest rate risk, because fixed rate assets (ex credit risk) can be replaced at a higher yield (conversely existing lower cost funding can be used to fund new higher rate assets). A prepayment option that is exercised when rates are higher is exercised out of the money. It’s exercise costs nothing to the writer of the option and in fact transfers economic value back to the writer. This is true whether it is a live prepayment or a credit event prepayment - because credit risk is a separate risk from interest rate risk - again, we are talking about credit risk combined with with options on interest rate risk - not options on credit risk. There is no ‘credit risk gamma’ - only interest rate gamma induced by credit risk when assets are fixed rate. Even if there is such a thing as a fixed rate sub-prime ARM, which I don’t think there is, credit event prepayments at higher interest rates would generate profits for interest rate risk and losses for credit risk.

Thinking about it further, gamma risk or negative convexity for a prepayable fixed rate mortgage is exacerbated when rates decline. The value of the mortgage declines from what it otherwise would be, because expected prepayments increase, reducing the mortgage duration. As a result, there is a loss to total value of some capitalized future interest, which is the reason why the value drops.

Moreover, from this new level of assumed duration, unexpected prepayments over and above expected prepayments also decrease the value of the portfolio from what it otherwise would be, because the new assumed duration was still overestimated.

An unexpected default event has the same effect on duration and portfolio value as an unexpected prepayment. But in addition to the contribution through the duration effect (which arguably is a gamma effect) there is the more fundamental contribution of outright credit loss (which is not a gamma effect).

This all assumes we are talking about fixed rate mortgages, which is not the case with conventional ARMS, and I don’t believe it’s the case with sub prime ARMs. If not, gamma is not an issue for either prepayments or credit risk.

i also appreciated your explanation.

your noted that some i-banks saw this coming, which implies “some” did not. who - setting bear aside — is a likely candidate to be among those who did not?

Thanks for taking the time to elaborate. I know it’s a laborious topic.

There are some things you’re saying that I think I understand, and some I don’t.

First, I understand that credit risk and default risk can trigger an effect equivalent to prepayment risk on fixed rate mortgages, and therefore credit and default risk are a source of gamma in attempting to hedge interest rate risk on such mortgages. But I don’t see how this applies to ARMs.

Second, I don’t know why sub-prime mortgages would have any unique credit risk effect as far as default is concerned. By that I mean unique in terms of the type of risk - obviously, the degree of credit risk is higher with sub-prime. But the type of risk and the type of effect on gamma wouldn’t be different than with prime ARM mortgages - just the degree of risk.

Third, I understand that volatile or illiquid markets pose risks of increasing net gamma exposure for delta hedgers.

I’m not too familiar with volatility swaps but I get the idea.

The extra interest charged on the ARM is a charge for the higher credit risk. I can see where credit risk exposure is a contingent function of interest rate risk, but you can’t hedge changing credit risk exposure with interest rate risk instruments unless its done to protect the interest rate sensitivity of the funding. Credit risk scenarios should have been taken into account in the original pricing of the spread and the associated return on capital. So I don’t see how that relates to option gamma in the event of default, unless its exposes the interest rate sensitivity of the portfolio on the funding side. The portfolio is being paid for the credit risk - any credit loss on default shouldn’t be compared to a market reference rate at a different credit risk level. It’s a pure credit loss.

Finally, I may not fully understand the basic interest sensitivity of sub-prime ARM mortgages. I assume it is not fixed rate and that the neutral interest rate sensitivity used in funding them would be short term - the same sort that banks would use to fund prime rate loans.

If interest rate sensitivity is short term, then prepayments have minimal effect on assumed duration and therefore minimal impact in terms of duration mismatch on the asset-liability portfolio (short term funding can also be paid off). This should be true whether prepayments are due to cash repayments or credit risk default.

But if sub-prime ARMS have embedded fixed rate risk, then clearly gamma enters the equation. Perhaps they do and I’ve missed this.

That’s one source of negative gamma. There is another. When interest rates rise, people with poor credit with adjustable mortgage rate mortgages can no longer pay them, and they default. This triggers a prepayment. When you have a prepayment, the value of the mortgage drops. If you graph the curve of the value of the mortgage versus interest rates, you get a frown, which is the dreaded negative convexity/gamma curve.

Guest: Generic ARMs are not fixed rate, so there is no fixed rate risk or related optionality risk in respect of prepayment or default on the underlying.

Yes there is. If a poor credit refuses to pay, then this triggers a default event. Subprime ARM’s have a “hidden” option which you need to include in your prepayment models.

It gets worse. If a person with poor credit defaults, you get cash which you can reinvest at current interest rates. The trouble is that the interest rate for the ARM was probably much higher than the market interest rate since you were charging more for poor credit. You lose that extra interest if the person defaults.

And it gets even worse…….

You can deal with gamma risk either by constantly buying and selling securities to make sure that you have a balanced hedge or by buying and selling volatility swaps which allow you balance the frown shaped payoff curve with another curve. Neither works well in the type of market environment we have now. You can no longer easily buy and sell securities which means you can’t rebalance your portfolio easily. In a market crisis, then the equations that govern option prices for exotics are likely to change radically so balancing a portfolio with those is likely to stop working.

But there is some good news:

1) One bit of good news is that *some* investment banks and hedge funds saw this coming and didn’t have big or any positions in subprimes.

2) The other bit of good news is that this is a “slow motion train wreck.” The thing that you want to avoid is a cycle in which high interest rates -> defaults -> people don’t lend -> higher interest rates -> more defaults -> people don’t lend

If that starts to happen you could blow apart the economy. Fortunately, the thing is happening slow enough so that the central banks can react and pump in liquidity to keep the cycle for happening. The situation is unfolding on a day by day basis. It’s not minute-by-minute like the 1987 crash or hour-by-hour like the LTCM or Mexican peso devaluation situation.

The analogy that I had about not forcing a heart attack victim to run to the emergency room is an apt one. Maybe the heart attack victim was remiss at smoking, eating bad food, and not exercising enough. Maybe the second you patch them up, they’ll leave and start doing bad things which will get them in the hospital again.

However, to try to “teach the victim a lesson” when they are on the stretcher is a horrible thing to do. You’ll almost certainly kill the patient.

Thanks for the explanation. But I don’t think so.

Negative gamma or negative convexity results from accelerating prepayment expectations on fixed rate mortgages as market interest rates decline. It’s essentially a put option on an underlying notional liability whose value increases as rates decline. Such an option is a function of fixed rates.

Generic ARMs are not fixed rate, so there is no fixed rate risk or related optionality risk in respect of prepayment or default on the underlying.

ARMs permit conversion to fixed rate mortgages when rates decline, but at the prevailing market rate. An option with a strike price equal to the prevailing market rate has no option value and no gamma.

Default risk is no different in its optionality consequences than prepayment risk with respect to interest rate risk on underlying assets that aren’t fixed rate.

The higher standard of living in the West is a result of a couple of centuries of capital investment, most of it manufacturing.

The most robust qualitative economic expansion in recorded from poverty to wealth was of course the United States from about 1875 to 1930. This was on the back of technological-oriented manufacturing, a strong internal market, and some external trade barriers.

The next one is happening now in China.

It is well known that rent-seekers (I guess this is what a “platform company” can personally make more profit, but a competitive economy producing products probably makes overall more wealth for owners and workers, even if the profits of the individual firms are smaller. The ‘lord’ of a banana republic can get really rich from political connections but the

This is a composition fallacy, that maximally profitable individual firms lead to greatest net prosperity.

The point is that manufacturing extends its influence of wealth, multiplier effect, to a far larger extent than rent-seeking platform companies. Manufacturing faces more competition and requires more capital investment. The factory owner might not make as much easy money—but far more people, over the long run, will benefit.

China knows this. The US used to know it. There is a difference between laissez-faire {where rent-seeking can flourish} and capitalism. Remember that Adam Smith was a liberal and wanted to change the system.

I disagree. It has everything to do with adjustable mortgages. Gamma is a measure how the value of a security changes in response to an external change in something like interest rates. If you have positive gamma (the curve looks like a smile), then a major jump in interest rates will make you money. If you have a negative gamma (the curve looks like a frown) then a major jump in interest rates will lose you money.

If you have negative gamma then you have to either buy and sell securities to move the frown around so that you are always on top of it. Or you can buy volatility options to also stay on top of the frown. Now if you are in a situation where the markets stop working and you can’t buy and sell to keep on top of the frown, then……..

Now in the case of subprime ARM’s. Gamma is negative. If interest rates rise, people default or refinance, this causes the price of the security to drop. It drops more than prime mortgages because there is a larger spread between the interest rate of the mortgage and the discount rate.

If people were indeed trading subprime ARM’s with the belief that gamma has nothing to do with ARM’s, this would go a long way to explain how we got into this mess.

I have no reference at all. The number is quite immeasurable. I am just guessing. It could be $65, $265, or $26.50. It is probably greater than $6.50. But $650 it is definitely not.

When Nixon delinked the dollar from gold, most “reputable” economists thought that the gold price would dive in just this way and for just this reason, because gold’s role in the monetary system was over. They were right that B follows from A, but they were wrong about A.