What is the point of innovative financial instruments ?
Robert Waldmann
claims that, after the invention of the mortgage based security, which clearly served to diversify risk, there have been three main purposes : weakening the effects of prudential regulations, weakening the effects of prudential charters and making balance sheets look better.
I remain very ignorant about banking and real world finance. Some time ago, a commenter noted that while at first I said I was winging it I seemed much more confident and asked if I had learned a lot or if I was winging it louder. I am winging it louder.
I don’t know what innovative financial instruments have been invented. I tend to assume that the purpose of some is tax avoidance. For all I know, some are used to share risk, and might actually be socially useful.
A very short list of the financial innovations and their purposes follows after the jump
CDO.
Why pool and tranche ? A common argument is that different investing entities have different risk preference, so it is useful to provide them with tranches of different safety. This doesn’t follow at all. In a simple model with no transactions costs and no non-traded assets, all agents buy all risky assets. The more risk averse invest more in the safe asset (treasury inflation protected securities or TIPS). There is no agent who wants a slightly risky AAA tranche but not a mezzanine or equity tranche. If all agents buy equal amounts of all tranches, there is no point in tranching. An intermediary is only needed to pool if the final investors would have to buy odd lots to diversify their portfolio. Individuals did not invest much in CDOs, they were bought by institutional investors who could have diversified on their own. (Source: https://bit-profit.io/)
However, for the purposes of banks’ capital controls, what matters is the amount of AAA, Aaa etc, and not the risk. A special purpose entity which pooled, tranched and sold to a bank reduced the risk adjusted assets of the bank. The purpose of relaxing capital controls was clearly served.
Also some institutional investors have charters which require them to invest only in investment grade debt instruments. CDOs gave them a way to bear the risk of, say, defaults on liars loans, without breaking this rule. CDOs helped neutralize the prudential charter.
I think it must have been about the rating and not overall porfolio risk (value at risk).
Interest rate swaps
Can be reproduced with long and short, spot and forward positions in bonds. However, total assets and total debt are extremely different if the same transaction has the form of an interest rate swap. I think they clearly existed to change the leverage reported on balance sheets without changing any obligations from one party to the other.
Trust-preferred Securities
See post below. They are a way to list debt as equity.
CDS.
A CDS can be reproduced with a REPO account. Consider firm A, firm B and AIG. Let’s go way back when AIG was AAA and assume that firm A is rated A.
First story, firm B buys a debt instrument from firm A and a CDS on the debt from AIG-FP. Firm B will be paid in full unless A and AIG both go bankrupt.
Second story Firm B buys firm A’s debt. Firm B opens a repo account at AIG-FP. It holds AIG bonds and a short position in Firm A bonds in it’s repo account. Firm B is paid in full unless Firm A and AIG both go bankrupt. Since AIG debt is rated safer as firm A debt, firm B has to send money to AIG to keep the repo account in the black and open. It is just like a CDS.
However, there are two differences. First some regulator might not count the repo account as making the debt safe – they definitely counted the CDS as making the debt safe. Second, and much more important, AIG has to issue a bond to make the second scheme work. To insure as much debt as it insured with CDSs AIG would have to issue $ 3 trillion in bonds. After doing that, AIG wouldn’t be rated AAA. The point of CDS is to make liabilities look small. They appear on the balance sheet at market value not at notional value.
AIG couldn’t make $3 trillion in debt instruments safe, yet, given accounting standards, debt rating practices, and Basel I capital controls, it was considered to have done so.
This made its balance sheet look better reasuring counterparties and top management.
Case_Shiller index bonds. Their purpose is neither to evade rules and regulations nor to buff balance sheets. They are useful to local governments trying to hedge risk. They were a total flop with trading volume around zero. They are the exception that proves the rule. Given the stated purpose and justification for financial innovation, they should have been a huge hit. They were a total wiff. This tends to make me more confident that the stated purpose and justification for financial innovation has little to do with reality.
I repeat I am ignorant. There may be other purposes to innovative financial instruments. Obviously I have only discussed the very few new instruments of which I am not completely ignorant. If anyone can describe an instrument which is used for a purpose other than the three I stressed (and, you know, is actually used unlike Case_Shiller bonds) please tell me in comments.
I think you’ve hit the key points.
Most financial derivatives created in the past 30 or so years (as opposed to the futures/forwards on commodities and warrants/options on stocks that have been around nearly forever) exist for one of two reasons – 1. hiding leverage and 2. regulatory evasion. The regulations being evaded in 2 generally pertain to capital adequacy, so you might want to view that as just a subset of 1.
CDOs prioritize risk such as default risk and prepayment risk. Insurance companies tend to want predictible payments and CDO tranches go a long way in making payments predictible. Since they are so useful its doubtful that the new finance bill will do anything serious to impede their use.
For a numerical example assume there is a 50 percent chance of prepayment for each mortgage security in a pool of 100. Now you could divide that pool into the first 75 prepayments in one tranche and the second tranche being assigned the last 25 mortages to repay early. Notice specific mortgages are are not assigned to either tranche. Rather the first tranche has to absorb the first 75 replayments, thus guarding the the premium AAA tranche. With a 50 percent chance of any one security paying off using a binomial distribution there is only about a 1 in 3.5 million chance that the AAA tranche will have any early repayments. The same holds for default. And you can make a more complex model of risk and the essential element of creating a priority sheme that shields the AAA tranche still holds.
I think you’re missing something pretty significant, which is what interest is – compensation for time. CDO tranching changed the preference of timing, not just order in the capital structure.
“There may be other purposes to innovative financial instruments”
4th purpose is so we can give loans to deadbeats, and some people will make lots of money off it.
“If anyone can describe an instrument which is used for a purpose other than the three I stressed (and, you know, is actually used unlike Case_Shiller bonds) please tell me in comments.”
I don’t like innovative financial instruments, so I haven’t looked.
Cardiff you have not been keeping up with my posts. I was shocked and horrified at the sort of ignorance about basic probability that you just demonstrated. In your example, you assumed that prepayment on each security is an independent random variable.
In fact, with a 50% chance of any one security pre-paying, the probability of prepayment in the AAA tranche ranges from 0% to at least 50%. If prepayment on each security is perfectly correlated with pre-payment on each other security, the correct calculation is one in 2 not one in 3.5 million. You did not state any assumption about independence. You made a mathematical mistake.
A binomial distribution describes the distribution of the sum (or average) of draws from *Independent* binomials with the same probability. Your error demonstrates that you don’t understand anything about probability and statistics. My advice, avoid any interaction with finance. You don’t have the tools.
Now think of the real world. In fact, default and prepayment on different mortgages was highly correlated. Same for pools of mortgages. It was idiocy like your gross error which caused people to grossly misprice CDOs.
As to a predictable flow, you seem to be saying that insurance companies do not like risk. OK. So they should buy equal amounts of each tranche of each CDO. I briefly explained that this is a standard result in basic microeconomics.
Since you are completely mathematically illiterate, as your gross massive error demonstrats, I don’t expect you to understand (it has to do with this funny thing called a first order condition).
I don’t expect to be able to convince someone who doesn’t know anything about probability. I do suggest you study up a bit before making a fool of yourself again.
m.jed long ago there were debt instruments with a variety of maturities. I can make pretty much any maturity I want using only Treasuries.
The standard result holds for instruments with different and stochastic maturities.
I’d love to see how someone like Delong would respond to this…
Robert,
Nice speech but you missed my key point. My point is that the priority scheme in a CDO is what makes it usefull in shielding the A+ tranche from the various type of risk.
Moreover, its clear that I was only using the binomial distribution with independence as a way to generate a simple numerical example.
You should have also paid more attention to this sentence
And you can make a more complex model of risk and the essential element of creating a priority scheme that shields the AAA tranche still holds.
This mean that I’m only providing a simple model which should make sense to you since I’m pointing out the priority scheme and not doing an exercise in probability.
And then you also said this:
You made a mathematical mistake.
I made no mathematical mistakes. What I did was to create a list in microsoft excel from 0 to 100 and then computed the probability of each outcome using Excel’s binomial distribution function: BINOMDIST(A3,100,0.5,FALSE). For the top 25 in the tranche to be affecte this means that the bottom 75 had to default. Summing probabilities for the having 75 to 100 events out of 100 sums to 2.81814E-07. 1/2.81814E-07 equals 3,548,438 which rounds to 3.5 million. So what I said stands.
DeLong is a liberal but he’s not a financial illiterate. I suspect he knows why CDO tranches are useful.
See Frank Partnoy’s Infectious Greed for a long series of variations on these basic themes:
– 1. hiding leverage to making beta look like alpha, or to take risks the tail of which will be borne by someone else.
– 2. scamming patsies (counterparties, customers, employers, investors, clients), into accepting risk they don’t understand, sometimes by causing bosses, regulators or rating agencies to misclassify it.
– 3. legal arbitrage, which is a polite word for a deeply anti-social practice: evading legal or contractual regulatory regimes — including safety, disclosure, pricing, executive incentives and tax — by creating an unregulated, or untaxed, or undisclosed, or contractually permitted equivalent to a regulated, taxed, disclosable or barred transaction.
The likelihood of finding a legitimate purpose is thin:
-Insurance is impossible, since diversified portfolios are already diversified and have therefore already eliminated insurable risk.
-Socially useful and efficient risk shifting is highly unlikely, since the originators of the risk will always have more information than the recipients, and, again, all are already diversified. Far more likely is recipients that have mispriced or recipients that expect to default or shift the risk further on.
-financial products are commodities, so competition should assure that any clearly legal and socially useful product does not generate significant profit; the fact that financial intermediaries make significant profits suggests something is wrong.
-Risk reduction by liquidity is theoretically suspect, since heavy trading creates new risk: markets are complex, self-referential systems with an ever present possibility of phase-shift to turbulence. Once liquidity permits traders to trade based on their beliefs about other traders’ beliefs, risk vastly increases.
Cedric,
I don’t like innovative financial instruments, so I haven’t looked.
Do you have life insurance? it so do you know what your policy has invested in? If you have one then your life may in fact be insured by some “shitty” CDO.
Daniel,
Why is it that all the liberal go on auto pilot and refuse to acknowledge the obvious benefits of CDO tranches that prioritize how bears the risk first. First and last makes a difference.
Yes, but the demand for mortgage or automobile loans typically don’t have people seeking 2 or 3 year maturities/payoff schedules. Since insurance companies also typically try to match fund and hold to maturity, buying a lot of “off-the-run” maturing MBS doesn’t make much sense either. Finally, insurance would be significantly more expensive if only funded by treasuries. Your simplifying, no-transaction costs, no-trading assets models may indicate that they could indeed exist and remain competitive against actual companies – in which case, I suggest you start one and compete against actual companies.
Actually, I cancelled my term policy in 2006. But it seems like I’m an owner of AIG without any insurance now. I think that is an “externality” that needs fixing.
Now that I know what CDOs are, I will be on the vigilant lookout to make sure the lower traunches don’t sneak up on me from somewhere. Proponents always talk about how great the AAA traunche is, but now the rest of the crap is out there somewhere. Unless it all found it’s way to Maiden Lane.
I assume you meant to say that “WHO bears the risk first” instead of “HOW bears the risk first.”
You probably should have said “First and last MAY make a difference.” Whether it does or not depends on whether the underlying assumption regarding correlation is correct. Priority is meaningless if you screw the pooch on that assumption. Just ask the people holding that worthless AAA CDO paper. Unfortunately for them, they paid a premium to buy that paper because they believed, incorrectly, that it was an iron law that priority in a CDO always matters.
One lesson from the Goldman Sachs Abacus deal is that Paulson obviously believed that the ratings agency models’ assumptions regarding correlation were wrong and could be gamed. Another lesson may be that Paulson was betting on the purchasers believing, mistakenly, that the risks of default and prepayment on each loan were independent, i.e., that there was no correlation.
Daniel,
Why is it that all the liberal go on auto pilot and refuse to acknowledge the obvious benefits of CDO tranches that prioritize who bears the risk first. First and last makes a difference.
fixed typo
Tao,
I think it’s more likely that Paulson just made an overall story that the real estate market was going to tank and positioned himself to take advantage of it. This is what hedge funds do.
Ya, of course telling Goldman that you would like the pool to consist of NINJA loans with South Central Los Angeles zip codes, when the ave selling price there hit $400K, would be a nice touch too.
Paulson wasn’t just placing a bet on the real estate market or random mortgages. He actually selected the mortgages that backed the Abacus synthetic CDO, for example, excluding loans based on underlying quality.
See, http://www.businessweek.com/news/2010-04-16/wells-fargo-bonds-rejected-by-paulson-for-abacus-deal-sec-says.html
Tao,
So do you think the two counter parties in this deal, one German, and once Dutch bank got bamboozed. My theory is that the Obama administration really did not give a hoot about these two banks but jimmied up the issue using their people at the SEC to push a finding just in time for a showboat hearing the just as the financial legislation was making its way through congress.
Cardiff,
You’re changing the topic completely, but I’ll bite.
Whether or not the two counter parties got “bamboozled” is secondary. The issue, in my mind, is Goldman Sachs’ behavior and that of the financial industry as a whole. The financial industry and our economy can only work if the institutions that act as intermediaries in our financial system can be trusted. Indeed, the only issue under the Securities Exchange Act is Goldman Sachs’ behavior, not its consequences.
This approach to regulating behavior to ensure a baseline of trustworthiness is also found in ethical rules for lawyers and accountants.
If the two counter parties did, in fact, reasonably rely upon Goldman Sachs’ representations to their detriment and those representations were knowingly false or misleading, then that would give rise to a separate cause of action by those parties for civil fraud (i.e., “bamboozled”). We’ll see what happens there. (Yes, I spent a chunk of my life as a lawyer.)
Regarding the timing of the complaint and whether the Obama administration politicized the decision to press the claim, I just don’t know. A couple of weeks ago, when some theorized that GS and Wall Street intentionally caused the “flash crash” to send a message to D.C., I felt the same way. Then I read up on how the Second Bank of the United States intentionally caused a recession to retaliate for Jackson’s withdrawal of deposits, and it certainly became plausible. Now, I feel the same way about the timing of the complaint: it’s plausible.
Personally, I think Obama, like Bush before him, is Wall Street’s best friend. He talks a good liberal game, but he actually acts conservatively. The disconnect between what he says and what he does is shocking. That’s one reason I believe that the “persecution” of GS and Wall Street starts and ends with this tiny little law suit that motivated liberals and conservatives alike. My bet is that Obama and Wall Street have choreographed this little Kabuki to keep the little people at each other’s throats instead of waking up to the reality that there’s only one ideology driving both parties, and that’s neoliberalism, aka corporate feudalism. GS has nothing to fear from this lawsuit. It’s a blip, part of the cost of being allowed to rip us off on a daily basis, er, I mean the cost of doing business.
One could actually argue that the complaint was timed to help water down the financial reform legislation, and it was certainly watered down at the administration’s insistence.
Not sure your analysis of interest rate swaps works – ‘can be reproduced’ is very different from ‘can be efficiently reproduced’, and at the extreme, becomes ‘can be theoretically but not practically reproduced.’ A sufficiently liquid market in bonds may not exist to allow for getting the same effect as an interest rate swap (especially in small amounts), or at least, not without very high costs.
‘ I think they clearly existed to change the leverage reported on balance sheets without changing any obligations from one party to the other.’ This does not seem to allow for companies/banks actually hedging, which does actually happen.
1. I think it was pretty obvious, even to Cardiff himself, that he was assuming independence of rvs. Waldman’s reaction seems a bit too harsh, even by his own standards.
2. “ It was idiocy like your gross error which caused people to grossly misprice CDOs. ” People pricing CDOs did not make these simplifying assumptions. It wasn’t such idiocy that caused mispricing, but pressure from traders and counterparts. Many quants were fully aware of the limits of their models. If you know otherwise, prove it.
3. The thesis of the post is however uncontroversial. Both equity and credit derivatives can be used to increase leverage and do balance sheet engineering.
4. Which finally reminds me why I now visit Angry Bear so rarely: it’s either obvious or arbitrary, while managing at the same time to be rude. And its authors are winging it.
Cardiff you made a mathematical mistake. Excel’s binomial distribution is based on the assumption that each of the 100 variables is independent. You did not state that assumption so you can’t make that assumption. It is exactly a gross extreme and reckless assumptions about correlation (less gross extreme and reckless than yours) which caused gross overpricing of CDOs.
You obviously don’t understand the assumptions behind Excel’s function BINOMDIST. Even if you had stated your assumptions, you should have understood that your calculation is not relevant to pricing CDOs in the real world. However, you are incapable of understanding basic probability and statistics.
I don’t have time for remedial math lessons. Please learn some math or stop commenting on math.
gappy I like used the word “like.” I did not write or imply that the error of CDO mispricers was as extreme as Cardiff’s. To say one thing is like another is not to say they are the same. It’s like the work like means something. I think it is clear that I contrasted the idiocy of the Gaussian copula with the gross idiocy displayed by Cardiff.
The thesis of the post is that the only purpose of derivatives developed since the original MBS is to increase leverage and do balance sheet engineering. The claim that they are used to do that is obvious. The claim that they have no other purpose is not. You do not contest my claim, but rudely say it is obvious while mistating it.
Then you call me rude. You can be as rude as you like, but please comment on the post as written rather than rewriting it so it is obvious and then saying it was obvious.
I actually did’t recommend buying only treasuries. I just said I can make any maturity I want using treasuries. I advocated (here at angrybear) for example buying lots of corporate bonds at random. I believe that such a portfolio has given higher returns than treasuries over all horizons ever. My belief is based on a guess about recovery ratios in the case of a Moody’s default event.
I’d guess that my avoid CDOs approach would have done fine compared to my competitores. I have no inclination to run an insurance company. Also, I’m sure you are aware of the issue of the minimum efficient scale which creates barriers to entry. A better investment strategy is not enough for an insurance start up to be successful.
I ask for information, when was the most recent insurance start up which competed with existing insurance companies ?
Cardiff, you now add rudeness to your inumeracy. Even if I grant your assumption that all random variables are independent, you have not explained why tranches are useful. Yes senior tranches are safer than junior tranches. So ? According to the EMH, they are redundant securities given the existence of treasuries. If I want to buy the risky tranches, I can open a Repo account, go long a pool of RMBS and short treasuries. If you want safety you can buy treasuries. We have no need for CDOs.
Note that no one has come up with a coherent response to my post. I don’t think I’m a financial illiterate.
Cardiff, I explained in my original post and in a comment up-thread why CDO tranches do not have obvious benefits.
I guess you win. I didn’t mention point 2. I absolutely agree that more liquid markets, that is thicker markets with higher trading volume are often more risky. I think this is a simple empirical fact. I think the refusal to accept the possibility is schizzo finance — if markets were efficient, thicker markets wouldn’t be riskier. Also financial operators care a whole lot about market thickness, because their schemes to make huge risk adjusted profits don’t work if markets are thin so they drive prices against themselves.
So we have a great desire for thickness based on the combined assumptions that markets are efficient and that they aren’t.
I was tempted to post a longer reply, but time is a scarce resource, and it would be pointless anyway. Better to set P(visit angry bear)=0. A mutually beneficially agreement.
I considered the actual hedging when I argued that innovative instruments could be reproduced given old fashioned instruments. The same hedging can be achieved, it just looks different on the balance sheet.
I don’t see how liquidity is a problem with constructing a swap with long and short spot and forward positions. I am discussing the same deal between the same parties. The only difference is what is written in the contract. It matters a great deal for measured leverage, but not at all for cash flows from one party to the other. I’d say firm A offers firm B a package deal which is just like an interest rate swap. They haven’t invented a new instrument. I don’t see why transactions costs are any different. As far as I can tell, the only difference is gross assets on Firm A and Firm B’s balance sheets.
Hence my claim.
Robert,
Cardiff you made a mathematical mistake. Excel’s binomial distribution is based on the assumption that each of the 100 variables is independent.
This is not a mathematical mistake. For a harvard PhD you display an alarming degree of ignorance on how CDOs work and a total lack of curiosity to fix this deficiency with reseach.
Again, the part that you don’t understand is that a CDO is a priority scheme with the lower tranches protecting the higher ones.
Since you could subsititute the most sophisticated model in the world to forecast default and I could easily incorporate it an make the my point just as well it show your critisism is without merit.
Robert,
I’m sure I do a lot more probabilistic modeling then you do so I’m getting a little sick of your unwarrented arrogance. I know all about independence, the lack of it, and using copulas to generate random variable from bivariate and higher order multivariate distributions.
You know you can spot people with low a skill level by the trivial points that they think they are sticking you with. Coberly does this a lot. I would expect better from you given your credentials. But so far, no luck.
Robert,
I’m sure I do a lot more probabilistic modeling then you do so I’m getting a little sick of your unwarrented arrogance. I know all about independence, the lack of it, and using copulas to generate random variable from bivariate and higher order multivariate distributions.
You know you can spot people with a low skill level by the trivial points that they think they are sticking you with. Coberly does this a lot. I would expect better from you given your credentials. But so far, no luck.
Cardiff, you assumed independence without stating the assumption, That is a mistake. I have no idea how you got the idea that I don’t understand that the most senior tranche of a CDO is safer than each underlying asset. That is, indeed, obvious. I don’t see how you could have gotten the impression that I don’t understand that from my post.
My argument that, in a simple model with no non traded assets, all agents should buy equal amounts of all tranches, is entirely consistent with that property of CDOs, which, I agree is obvious.
I did not assert that your mathematical error is, in any way, relevant to my post. It is a mathematical error which I find shocking and appalling (I wrote that when I wrote a post about the error as made by Alex Tabarrok search the archives for Tabarrok and waldmann and you willf ind the post). I said your error was an error, because that is what I think. Note Tabarrok was not criticizing me, yet I posted about the error (identical to yours) because I was sincerely shocked and appalled that anyone would make such an error.
Now I will grant, for the sake of argument, both that you stated your assumption of independence and that it is interesting to talk about CDOs with that assumption. In that case, in a simple model with no non traded assets and no transaction costs, all agents will buy equal proportions of each tranche.
That is to say your error is not related to my response to your criticism. If I grant you your unstated assumption, I still find your belief that you have criticized my post mistifying. I never said anything about the riskiness of different tranches of CDOs except that none provided a perfectly safe real return. My argument is a valid argument about CDOs made of underlying assets with independent returns.
You might find it odd that an extremely risk averse agent will invest in the equity tranche. The mathematical result (which is just a first order condition) is that an extremely risk averse agent will buy very little of a risky asset (where risk is correctly measured by comovement with the agent”s consumption with, with no non traded assets will be determined by the return on the portolio). This applies proportionally for all risky assets.
This is a very basic result which I was taught in the second economics course I ever took. increased risk aversion causes a proportional decline in investment in all risky assets, both those which are very risky and those which are very slightly risky. This is true, even if the only risk is inflation as nominal returns are completely safe, so it definitely applies to RMBS based CDO tranches including the most senior tranche.
Please re-read my post and try to understand why you thought I claimed that senior tranches must be as risky as each of the underlying assets, or the pool of underlying assets or whatever you think I wrote. Since I have never written or thought either of those things, I have no idea what you believe I wrote. In any case, I didn’t write it.
Indeed. Anyway, I would recommend googling “gaussian copula” I think you will find ample evidence for my claim that an error along the same lines as that made by Cardiff was actually important. I don’t claim it is proof, but I can say it wasn’t my idea that CDO mispricing was importantly affected by false assumptions about multivariate distributions (not that they were independent but that the probability of default of two securities could be described by a multivariate probit with some estimated correlation which is, itself, a very strong assumption which has been rejected by the data by now).
I note that you made a harsh claim about my post. I assert that your paraphrase is inaccurate. Now you say you won’t waste time either retracting it or defending it. I think you should either stand by what you wrote or retract it.
Cardiff, you assumed that random variables were independent without stating the assumption. This is not a trivial point. Correlation is central to porfolio theory and you neglected it. It’s nothing personal. I have a full post expressing horror at the same error made by Alex Tabarrok.
OK so you know about copulas. Tell me, why is it a terrible mistake to use a gaussian copula when modelling default ? What is wrong with using data on CDS prices to estimate the correlation of a bivariate gaussian copula ?
Why is it so hard for you to admit that you made a mistake ? Assuming independence without stating the assumption is a mathematical mistake. You have gone from saying you didn’t make a mistake to saying it is trivial and that my pointing it out shows that I have a low skill level. This is strange ?
Off the top of my head no life insurers come to mind, but Property & Casualty companies pop up after catastrophe events destroy industry capital with quite a bit of frequency. There exitst a “Bermuda Class of:” 2005 (post Katrina, Wilma and Rita); 2001 (post 9/11); 1992 (post Andrew); and 1985/86 (casualty reserve crisis). Commercial P&C is distributed through brokers and with corporate clients, so distribution isn’t a significant barrier to entry which is a big difference compared with life insurers. There are reinsurers of life insurers that are also created, for similar reasons. Warren Buffet recent started a bond insurer de novo as well.
I’d guess that an avoid CDOs insurer would have survived and put *your*company in better stead post crisis, however, it would have been starved of capital from 2003-2007, as competitors offered more compelling products to its clients.
I’d also point out, that unless I’m mistaken, while 1 life insurer (Lincoln) and 2 multi-line insurers (Hartford and AIG) received TARP, not a single insurer went under and as we all know, it wasn’t AIG’s insurance operations that were the cause of its downfall.