Serving the Quantitative Finance Community

  • 1
  • 3
  • 4
  • 5
  • 6
  • 7
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 1:20 pm

QuoteOriginally posted by: FredBTThe key issue is that you cannot find an "arbitrage portfolio" ... More precisely, you may obtain a riskyless portfolio, but you cannot do anything with such a "non-tradable" virtual object.what do you mean ? Can you show us an example of this "non-tradable" virtual object ?
knowledge comes, wisdom lingers
 
User avatar
hamster
Posts: 1
Joined: October 12th, 2008, 3:51 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 1:23 pm

yesterday i have read the following paper, based on a simple heterogenous agent model with market makers (be aware of its simplicity; it's rather descriptive theory).Westerhoff 2003look at page 366 for 'hot potato trading' and it's (weak non linear) relationship to trading volume. Ok, it's about fx rates or stocks of foreign money. However, it might be generic for market micro structures. OTC is market making, isn't? One might conclude that increasing trade volume of something (e.g. claims such as swaps) between market makers (e.g. banks) is symptom (not the cause) of 'hot potato trading'. But reading a bit further, the paper reveals that 'panic' (e.g. regarding the claim or its underlying) amplify things even worse although it is not rational (or doesn't make sense). About panic regarding the underlyings' underlyings'... underlying...: A friend of mine (tax advisory) told me that some of his clients ask him to dress down balance sheets (creative downward accounting). I was wondering why these entrepreneurs want to fill bankcrupty? No no (smirk), he told me that some clients are gambling or respectively using a psychological trick. Ok, in some cases, it's bankcruptcy like usual, nothing special. But, some clients trying to induce a sense of urgency for employees in order to reorganise their firm, and in some cases internal management buyout agendas are going on - Threat of bankcruptcy is an elegant bargaining argument, isn't?. Such risk tolerant people make use of the current media panic about "crisis everywhere". The effect is that partner banks of downdressed firms might get shocked (at least partner banks of the partner banks are shocked due to incomplete information), about the all the changing capital structures here and there, although it's fake to some extent (not fully). Therefore, don't trust these media lobbyist talking about armagadon including the death of all debt outstanding - They shout loudly: PANIC PANIC PANIC!!! Everybody will belive it! You see, it works...
 
User avatar
FredBT
Posts: 0
Joined: April 16th, 2008, 8:31 am

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 1:39 pm

QuoteOriginally posted by: daveangelwhat do you mean ? Can you show us an example of this "non-tradable" virtual object ?Well, I mean the following.In equity, you have a derivative on a financial instrument. For example, it can be a call on a stock. The process for building an arbitrage portfolio is as follows:Assume some dynamics - using eg a well-choosen Ito process - for your underlying process, like GBM for stock pricesApply Ito's calculus to obtain the dynamics of your derivativeBecause the derivative and the underlying have the same source of uncertainty, randomness (i.e. noise processes) are the sameConsequently, it should be possible to come up with a linear combination of those such that globally, the randomness vanishesThe above combination yields the desired portfolioThe critical point here is that you can add to your original "derivative portfolio" some "underlyings", I mean you can buy some shares in such a proportion that riskyness vanishes.For a CDS, the underlying process is the default time process. The above development holds whatever is the process, I agree, but it's purely tgheoretical. Practically however, it is not possible to buy "default times", you cannot come up with a risk-free portfolio. Or more precisely you can, but it has no "financial meaning". That was my point...The fact that in the CDS case the arbitrage portfolio is not built of "tradable assets", makes it "virtual"...
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 1:42 pm

QuoteOriginally posted by: FredBTQuoteOriginally posted by: daveangelwhat do you mean ? Can you show us an example of this "non-tradable" virtual object ?Well, I mean the following.In equity, you have a derivative on a financial instrument. For example, it can be a call on a stock. The process for building an arbitrage portfolio is as follows:Assume some dynamics - using eg a well-choosen Ito process - for your underlying process, like GBM for stock pricesApply Ito's calculus to obtain the dynamics of your derivativeBecause the derivative and the underlying have the same source of uncertainty, randomness (i.e. noise processes) are the sameConsequently, it should be possible to come up with a linear combination of those such that globally, the randomness vanishesThe above combination yields the desired portfolioThe critical point here is that you can add to your original "derivative portfolio" some "underlyings", I mean you can buy some shares in such a proportion that riskyness vanishes.For a CDS, the underlying process is the default time process. The above development holds whatever is the process, I agree, but it's purely tgheoretical. Practically however, it is not possible to buy "default times", you cannot come up with a risk-free portfolio. Or more precisely you can, but it has no "financial meaning". That was my point...The fact that in the CDS case the arbitrage portfolio is not built of "tradable assets", makes it "virtual"...I think you are just thowing phrases like "tradable assets" to make you argument sound more intelligent than it is. How about you but a bond issued by IBM and buy protection on that bond ? Why does that not work ?
Last edited by daveangel on February 8th, 2009, 11:00 pm, edited 1 time in total.
knowledge comes, wisdom lingers
 
User avatar
FredBT
Posts: 0
Joined: April 16th, 2008, 8:31 am

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 2:04 pm

QuoteOriginally posted by: daveangelHow about you but a bond issued by IBM and buy protection on that bond ? Why does that not work ?Well, my humble opinion is that the bond gives you information about default probabilities, but unfortunately nothing regarding exact fair value. Suppose for example you have a CDS on a company C maturing at T, and a risk-free rate curve. We consider a contract with a fixed recovery rate.1) suppose we read in my crystal ball that C will default at t*. Then, you know exactly, i.e. without any indeterminacy, the PV of your position. More precisely, you know the LGD (defined to be zero if t*>T) together with the amount and the scheduling of the premium to be paid. After proper discounting, you come up with one single price.2) I must admit that if you give me bond prices instead of t*, I could potentially come up with default probabilities, but I could not come up with a value of my position being not subjected to any discussion. Maybe it is possible, but I fear then to not know the procedure. I think there is still room for randomness when doing this way (simply because you need probabilities). Technically, I suppose we can claim that the value of your position is not "Bond"-measurable, but well "Default time"-measurable...
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 2:08 pm

QuoteOriginally posted by: FredBTQuoteOriginally posted by: daveangelHow about you but a bond issued by IBM and buy protection on that bond ? Why does that not work ?Well, my humble opinion is that the bond gives you information about default probabilities, but unfortunately nothing regarding exact fair value. Suppose for example you have a CDS on a company C maturing at T, and a risk-free rate curve. We consider a contract with a fixed recovery rate.1) suppose we read in my crystal ball that C will default at t*. Then, you know exactly, i.e. without any indeterminacy, the PV of your position. More precisely, you know the LGD (defined to be zero if t*>T) together with the amount and the scheduling of the premium to be paid. After proper discounting, you come up with one single price.2) I must admit that if you give me bond prices instead of t*, I could potentially come up with default probabilities, but I could not come up with a value of my position being not subjected to any discussion. Maybe it is possible, but I fear then to not know the procedure. I think there is still room for randomness when doing this way (simply because you need probabilities). Technically, I suppose we can claim that the value of your position is not "Bond"-measurable, but well "Default time"-measurable...I still don't get your point. Sorry I am thick. Bond is trading at par implied yield is T+75. it has 5 years to run. you can buy protection on the bond at 50bps pa. whats the problem with the arb ? You clip 25bps per annum. If it defaults you deliver the bond and get par. If it doesn't you get par.
Last edited by daveangel on February 8th, 2009, 11:00 pm, edited 1 time in total.
knowledge comes, wisdom lingers
 
User avatar
FredBT
Posts: 0
Joined: April 16th, 2008, 8:31 am

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 2:22 pm

Well, I fear I don't get your point either ... I simplly emphasized the fact that you could not build a risk-free portfolio having an economic value, because this would require the combination of a credit derivative together with its underlying, and in the case of a CDS, the underlying does not have any financial value. How does that observation contradict your thoughts ? I never said that arbitrage was not possible...By the way, I've just seen your assasine sentence I think you are just thowing phrases like "tradable assets" to make you argument sound more intelligent than it is. "Dear mate, thanks for this. Although I must say I don't see the interest of writing such kind of things. If you're just considering my intervention as being "two cents", why not stop discussing ? I think this is not very constructive, nor respectful. But I leave that up to you.
Last edited by FredBT on February 8th, 2009, 11:00 pm, edited 1 time in total.
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 2:28 pm

QuoteOriginally posted by: FredBTWell, I fear I don't get your point either ... I simplly emphasized the fact that you could not build a risk-free portfolio having an economic value, because this would require the combination of a credit derivative together with its underlying, and in the case of a CDS, the underlying does not have any financial value. How does that observation contradict your thoughts ?I have just given you an example where a portfolio is risk free and has financial value to the tune of 25bps per annum. this happens every day in the market. QuoteBy the way, I've just seen your assasine sentence I think you are just thowing phrases like "tradable assets" to make you argument sound more intelligent than it is. "I think you mean asinine ...The reason why I want to continue the "discussion" is because I want to learn something. Please enlighten me with a concrete example thanks "mate"
Last edited by daveangel on February 8th, 2009, 11:00 pm, edited 1 time in total.
knowledge comes, wisdom lingers
 
User avatar
HOOK
Posts: 0
Joined: October 10th, 2008, 5:15 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 3:33 pm

I´m sorry to bother you guys again on this.People say CDS are priced on the risk neutral measure. I can understand what the theoretical hedge (portfolio of the bond + risk free) would be like, provided a recovery rate is given. How can we hedge this recovery rate so that we are in fact pricing CDS in the risk neutral mesaure, free from market prices of risk, etc?No human being so far could answer that question...maybe an angel ?
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 3:46 pm

QuoteOriginally posted by: HOOKI´m sorry to bother you guys again on this.People say CDS are priced on the risk neutral measure. I can understand what the theoretical hedge (portfolio of the bond + risk free) would be like, provided a recovery rate is given. How can we hedge this recovery rate so that we are in fact pricing CDS in the risk neutral mesaure, free from market prices of risk, etc?No human being so far could answer that question...maybe an angel ?the reason no human being can answer the question is probably because it is not clearly set out.I think your assumption is wrong - you don't need to know the recovery rate to create a risk free portfolio of bon dand CDS. What is wrong with buying a bond, swapping its cash flows into floating rate and buying protection on the issuer that coincides with the maturity of the bond ? Why is that not a risk free portfolio ? Why do we need to know the recovery rate ?
knowledge comes, wisdom lingers
 
User avatar
Traden4Alpha
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 4:35 pm

QuoteOriginally posted by: HOOKQuoteOriginally posted by: Traden4AlphaFiat currencies, fractional reserve banking, and most derivative contracts are quasi-stable. They are stable over some range of economic dynamics, but go horribly pear-shaped if pushed too far.That´s a very insightfull approach ! What do you think could make one derivative more stable than another ? Clearing houses? Feasiability to hedge ?Very good question..... Yes, clearinghouses do help by diffusing the counterparty risks, improving price discovery, increasing transparency on to the total state of the market, and providing a low-friction venue for price-stabilizing market-makers. But clearinghouses are not guarantor of stability (and can promote instability!).Hedging helps, but only if it can be strictly static (e.g., a "hedge-and-forget" strategy). Dynamic hedging depends far too much on the kindness of strangers and the pleasantries of well-functioning markets to provide liquidity and hedging instruments at a fair price.Another might be reserve regulations that record these instruments in terms worst-case value (e.g., liability value = the expectation of greatest of the likely payout and asset value = expectation of the least of the likely payoffs). Too many derivatives contain hidden leverage in the form of the span of outcomes relative to the nominal or "expected" value. Some of this might also be covered by the clearinghouse regulations, which would (we hope) have appropriate margin requirements. In contrast, I don't think that the Basel II-style asset categories work too well because it's too easy for high quality assets to suddenly become low quality assets and require an increase is reserves at the wrong time (i.e., the asset risk categories may be procyclical, not anticyclical).What fascinates me about the current crisis is that it represents the failure of a widely-used model of risk more than the failure of any given institution. If only one bank, even a behemoth, had over-leveraged and lent far too much money to uncreditworthy consumers, the system could have handled the problem. As a first approximation, if the market share of liabilities of the largest bank times one minus the worst-case recovery is smaller than the reserve ratios used by all other banks, then the financial system can survive the failure of the largest bank. Strictly speaking, none of the banks in the world was really "too big to fail." But, to the extent that so many banks overleveraged and held similar portfolios of toxic debt (e.g., large quantities of consumer debt instruments), then the failure of one bank becomes much more than a source of direct losses (i.e., unrecovered assets), it also becomes an indirect source of losses by repricing the assets of all the banks that hold similar portfolios. Thus the un-accounted correlation risks internal to consumer debt instruments were converted into un-accounted correlation risks in the larger banking system.That's why exchanges can actually increase instability. On the one hand, exchanges diffuse concentrated risks to reduce the probability that a counterparty failure begets trading parter failures. On the other hand, exchanges increase the correlation risk by allowing everyone to buy a piece of a toxic asset class. Clearly, exchanges aren't a prerequisite for the over-diffusion of risk -- that European banks took such large hits from American MBS if proof of that. But exchanges do increase the spread of instruments (especially instruments associated with a price bubble) by broadening access of capital to the freely exchanged instruments. Had MBS been actively traded on a public exchange, we would have seen even more money flowing into subprime mortgages, a delay to the peak valuation of housing, and an even deeper crash than the one we are experiencing now.The deeper lesson is that correlation risk cannot be avoided as long as we allow institutions to hold mutually correlated portfolios of assets. To the extent that most banks hold similar portfolios, then most banks will fail under the same conditions. (the same logic applies to hedge funds in which the forces liquidation of a "bad" HF causes trouble for all other "good" HF that hold similar portfolios)But the deepest lesson is that all attempts to stabilize a system tend to have two effects. First, attempting to stabilize the system does stabilize the system by expanding the range of scenarios under which the system remains controlled. Second, the experience of a stable system leads people to increase leverage, risk taking, and increases pressure to relax the stabilizing measures. People cannot help but interpret a period of stability to mean that we've had excessive conservatism and that it's OK to borrow a little more or push a little harder to get returns. Add L-2 norm risk models (e.g., our friend, the Normal distribution) and we create a system that shows increasingly modest risk with high probability and increasingly severe risk with low probability. That is, the pencil tip grows larger, but people compensate by getting/creating larger pencils.It's not clear to me how or even if we should try to control this phenomenon of boom and bust cycles. Ultimately, I'm ambivalent about bubbles (and crashes) due to their role in both creating beneficial new industries and in clearing flawed models, but that's a discussion for another thread.
Last edited by Traden4Alpha on February 8th, 2009, 11:00 pm, edited 1 time in total.
 
User avatar
HOOK
Posts: 0
Joined: October 10th, 2008, 5:15 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 5:43 pm

QuoteOriginally posted by: Traden4AlphaIt's not clear to me how or even if we should try to control this phenomenon of boom and bust cycles. Ultimately, I'm ambivalent about bubbles (and crashes) due to their role in both creating beneficial new industries and in clearing flawed models, but that's a discussion for another thread.I look forward to this Shumpeterian/Darwinist thread.
 
User avatar
HOOK
Posts: 0
Joined: October 10th, 2008, 5:15 pm

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 6:45 pm

QuoteOriginally posted by: daveangel<brI think your assumption is wrong - you don't need to know the recovery rate to create a risk free portfolio of bon dand CDS. What is wrong with buying a bond, swapping its cash flows into floating rate and buying protection on the issuer that coincides with the maturity of the bond ? Why is that not a risk free portfolio ? Why do we need to know the recovery rate ?Indeed you wouldn´t need it. The portfolio would be risk free. Well Dave, seems I was lot on some cumbersome mathematics and lost the perspective."I once was lost, but now I´m found, Was blind but now I see"Cheers
 
User avatar
freddiemac
Posts: 7
Joined: July 17th, 2006, 8:29 am

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 7:25 pm

QuoteOriginally posted by: HOOKQuoteOriginally posted by: daveangel<brI think your assumption is wrong - you don't need to know the recovery rate to create a risk free portfolio of bon dand CDS. What is wrong with buying a bond, swapping its cash flows into floating rate and buying protection on the issuer that coincides with the maturity of the bond ? Why is that not a risk free portfolio ? Why do we need to know the recovery rate ?Indeed you wouldn´t need it. The portfolio would be risk free. Well Dave, seems I was lot on some cumbersome mathematics and lost the perspective."I once was lost, but now I´m found, Was blind but now I see"CheersIt is true that you don't need to know the recovery rate in order to create a risk free portfolio given that the bond is trading at par. For a bond trading away from par you need to make assumptions about the recovery rate in order to be able to determine your hedge ratio and you must also choose between hedging default risk or spread risk.
 
User avatar
PlasticSaber
Topic Author
Posts: 0
Joined: April 28th, 2007, 8:17 am

Taleb wanted to get CDS banned. Why??

February 9th, 2009, 8:46 pm

QuoteOriginally posted by: daveangelQuote I think your assumption is wrong - you don't need to know the recovery rate to create a risk free portfolio of bon dand CDS. What is wrong with buying a bond, swapping its cash flows into floating rate and buying protection on the issuer that coincides with the maturity of the bond ? Why is that not a risk free portfolio ? Why do we need to know the recovery rate ?Daveangel is largely right. We don't need to know a recovery rate to create a basis package. It is very common to see bond trading cheap in relation to CDS since everyone was hoarding cash at around Oct to Nov 08 . Say a bond is trading at 40c of par and you can buy CDS protection for 45% upfront (with no running coupon to simplify the discussion). For a 8.5M initial cash outlay (buy 10M notional of bond for 4M, 10M notional for 4.5M), you will get 10M back when the bond matures. If the bond defaults, you can just choose physical delivery of the bond along with the CDS. You will get 10M back no matter what the recovery is. You can see a 1.5M guaranteed gain (unless the CDS counterparty goes down first then the corporate bond issuer files for bankruptcy right after ...)However, when we look closer, we would notice that such a basis package is a default-timing play. The PV of this 1.5M gain would be higher if the default event happens right away. Wait a minute, how do we estimate the expected default time? We need the CDS curve and the recovery rate. It is how the recovery rate crawls back to the picture. But I consider the effect of this secondary...