ask Dcifrths...well, anything...about finance/mkts/ports that is.

[CrushinFelt]

You could probably just think about it incrementally. You would need to produce an alpha that is greater than the interest rate for the borrowing to make sense. Basically if you understand futures faily well it is a much simpler thing to do. Margin requirements are not usually much more than 10% of the position so you'd have that 90% remaining to try an obtain your alpha.

[DcifrThs]

SC

you said:

[ QUOTE ]
Lastly, do you agree that economists (that was my major) are "non experts" as described in the Black Swan and basically worthless, particularly as it relates to prediction (I subscribe to this belief, FWIW).

[/ QUOTE ]

i've finally read enough of black swan to feel like i can answer this question.

first i guess it depends what you consider an economist. while at the DoL i was titled an "economist" but i was tasked with the analysis of health care & pension coverage of US citizens. i wasn't trying to predict anything.

some economists are only concerned with small areas that belong to mediocristan (# of immigrants in a given year, or annual increase in labor force) that can be predicted fairly well.

others, fall victim to their own conceit (As NNT predicts) in trying to predict the level of the DOW or the price of gold or the level of inflation years out on the curve.

those guys & gals are doomed to repeat the mistakes of their predecessors. but it isn't their fault imo. the demand for their (histerically wrong) predictions seems to only INCREASE!!

so they are worthless as far as prediction is concerned.

further, i think you just fell victim to the platonization of the problem you stated (and one of the main things NNT tries to move the reader away from). you said you believe economists are worthless as predictors...but there are economists who are great at predicting their small area as their area may not be in extremistan (at the risk of falling victim to catagorizing things myself...but NNT tries to at least separate the types of these things: those in which one massive extreme CAN overtake all other measurements-extremistan- and those in which one massive extreme is just a drop in the bucket and makes a very small difference overall-mediocristan)

finally, the definition of "expert" above may be too narrow. i'm sure all economists are "experts" by the definition ascribed to them by their employers, but fail in predictions (they bring revenues & respect or whatever to their employers though their products may fall short). one such person i feel is horrible at this area is "legendary" bill gross. he is a "bond bull" that very recently turned "bearish." he turned "bearish" only AFTER the biggest upward swing in 10yr yields that many like MS, JPM, my former employer, and thus myself have thought very likely for some time.

but the mkt still thinks bill gross is an expert b/c he runs a profitable and highly desired company, or whatever it is (PIMCO). oh...and when bill gross turned bearish, yields pared back by 10 bps (though i still agree w/ his current position that yields are generally now still headed up).

anyways, enough about micro rambling. i hope i asnweres your question.

thanks SC,
Barron

[RicoTubbs]

Barron,
For whatever reason, you shut down a thread immediately after I asked you a question, so I'll repeat the question here:

"where did you get your information about LTCM? You make several very specific statements (they didn't consider the possibility of a 20% loss, they didn't use monte carlo simulations to stress test their models, they made predictions with 100% precision in their shareholder letters) that sound unlikely to me. I’ve read a bit about LTCM and would appreciate any references you have that would support your claims. [I don’t mean this as a dick-waving challenge – I would genuinely like to read any sources that you have.]


This is the thread:
http://forumserver.twoplustwo.com/showfl...=0#Post10936451

[DcifrThs]

[ QUOTE ]
Barron,
For whatever reason, you shut down a thread immediately after I asked you a question, so I'll repeat the question here:

"where did you get your information about LTCM? You make several very specific statements (they didn't consider the possibility of a 20% loss, they didn't use monte carlo simulations to stress test their models, they made predictions with 100% precision in their shareholder letters) that sound unlikely to me. I’ve read a bit about LTCM and would appreciate any references you have that would support your claims. [I don’t mean this as a dick-waving challenge – I would genuinely like to read any sources that you have.]


This is the thread:
http://forumserver.twoplustwo.com/showfl...=0#Post10936451

[/ QUOTE ]

i'm sorry.

i missed that post and thought i answered the last one.

i'll repost & answer it here:

you wrote:

[ QUOTE ]
I was trying to make more of a philosophical point, but failed, I think. You brought up Taleb, which allows me to restate my point, maybe more clearly.

The problem I have with Taleb's writing, albeit based only on Fooled by Randomness - I haven't read Black Swan, is that he seems to pound the table to point out that:

1. People are overconfident about their ability to understand statistics and don't have the ability to estimate or appreciate the true distribution/probability of events. Alternatively, people fail to appreciate the degree to which unexplainable "randomness" exists in data.

[/ QUOTE ]

not really. the main point is that people fail to appreciate the degree to which wild randomness FAILS to exist in the data

[ QUOTE ]

2. People underestimate the likelihood of rare events.

These two points seem contradictory to me, at least in the way that Taleb presented them. By simultaneously claiming that the true distribution of some outcome isn't knowable and that people underestimate the extent of rare events, his book came off as almost a religious text where Taleb himself had a supernatural ability to understand the true distribution of events.

Why, for example, couldn't it be the case that people tend to overestimate the occurrence of rare events? As a concrete example, Steve Levitt would say that parents overestimate the probability of guns causing harm to their children.

Unless Taleb knows the actual distribution, which is unlikely given his stance that such things are unknowable, who is he to say whether people are over- or under- estimating the probability of some event. That, for me, was the fatal flaw with “Fooled by Randomness” – he never gave a fundamentally sound reason for why he believed that people underestimate rare events.

[/ QUOTE ]

he follows up very well in The Black Swan. he doesn't have a problem with people who know they don't know the exactitudes of a distribution. he DOES have a problem with people who either a) don't know they don't know and use the bell curve, or b) people who know they don't know AND STILL USE the bell curve. the latter proffer the reason "well some approximation is better than no approximation." that logic is very bad and clearly shown by saying in response "well, i know how a molecules act in a solid state so i'll use them to aproximate how molecules act in a gaseous state."

the two casese are not contradictory imo.

you can know about the distribution of events (i.e. what i was searching for w/ mandelbrotian -fractal- distributions where the randomness scales) but not know it exactly. that is far better than knowing you don't know something but using a terrible approximation for reality anyways.

so your problem is that he says people underestimate the likelihood of rare events AND that they don't know what the distribution is. but he knows what distribution THEY FAVOR and THAT DISTRIBUTION vastly underestimates the probability of rare events (the tails).

further, just for clarification, he never says that the distribution "isn't knowable." he says that it "isn't gaussian."

confirmation is virtually impossible inductively whereas repudiation is easy. it only takes ONE black swan to disprove the statement "all swans are white" whereas you can see 1 trillion white swans and be unable to conclusively prove that "all swans are white."

similarly, it only takes a few observations of past data to prove that "financial markets are not gaussian (or anywhere close)"

examination of that data complies far better with scalable distributions.

this approach looks at the extreme events (whose cumulative effects are massive) and then treats normal events as secondary. gaussian approaches do the exact opposite. the latter claim that the cumulative effects of the 20 sigma events wash out over time. instead they are hugely impactful.

[ QUOTE ]

Tying back to my original post, I squirm when people simultaneously say:
[1] LTCM was modelling risk in some way (e.g., using some type of VaR metric) when in fact such future risk is simply unknowable.

[2] They underestimated the risk of some event (e.g, a particular liquidity crisis).

I think it's a little weird to simultaneously claim that an event is unmeasurable AND that you know the direction that they misestimated. In my eyes, it suggests an omniscient speaker.

[/ QUOTE ]

no. let me rephrase. when you construct an alpha portfolio, the goal is (in my mind) to first and formost, NOT to blow up. lets take one thing first. towards the end, LTCM was taking positions that, should they move against them, would constitute a massive exposure and force margin calls. one of the nobel laureate principles intimiated to the fund manager that this was a problem but the fund manager replied by saying it isn't likely to happen based on the models they were using (i.e. "YOUR models").

further, they were taking positions in areas in which they had no serious expertise. they were using a hammer to make a milkshake. the tool didn't fit the task. this is deficient management by all accounts. it doesn't matter what you know or don't know. when the counterparties got a hold of LTCMs (until then extremely well guarded) books, they were absolutely appalled at the state of affairs. just the amount of leverage they were using.

so the 2nd point you say is that i stated they "underestimated some specific event." what i meant is that they didn't precisely think through what would happen if something that wasn't taken into account int he model happened, i.e. what happened if correlations moved? are correlations static? should they be treated that way? what happens if they aren't? it isn't impossible to ask these questions at that time. my former employer was asking (and answering) them very well and did very well during 1998, and 2000.

what was deficient wasn't the misestimation of ONE SPECIFIC event, it was the LACK OF THOUGHT that went into "risk management" vs. "alpha creation." the way the risk management worked at LTCM is laid out very well in the book i recommend below. it was seriously deficient and not reminiscient of a top flight group (like the one i used to work for).

i hope this answers the above, if not, please ask again.

[ QUOTE ]

As a side note, where did you get your information about LTCM? You make several very specific statements (they didn't consider the possibility of a 20% loss, they didn't use monte carlo simulations to stress test their models, they made predictions with 100% precision in their shareholder letters) that sound unlikely to me.

[/ QUOTE ]

this is not unlikely. it is fact. read up on "When Genius Fails." they reprint shareholder letters there and convey exactly what i stated. they actually do claim that the probability of an %X loss in Y time periods is precisely Z%. the X stopped at 20%. i.e. in the letter they gave to shareholders they didn't consider the possiblity of losing 20% in a year since the probability of it, based on the normal distribution they were using, was negligibly small.

[ QUOTE ]
I’ve read a bit about LTCM and would appreciate any references you have that would support your claims. [I don’t mean this as a dick-waving challenge – I would genuinely like to read any sources that you have.]

[/ QUOTE ]

for one, like i said, "when genius fails" is great. it takes you step by step down the road of LTCM's blow up.

hope this helps, feel free to ask anything else and i'm sorry about "closing" the other thread. i honestly didn't see that last post there.

Barron

[Apokrupto]

If you are a complete 24 y/o n00b at anything investing and have $1500 that you want to begin with on E*Trade, what would you do?

[DcifrThs]

[ QUOTE ]
so the 2nd point you say is that i stated they "underestimated some specific event." what i meant is that they didn't precisely think through what would happen if something that wasn't taken into account int he model happened, i.e. what happened if correlations moved? are correlations static? should they be treated that way? what happens if they aren't? it isn't impossible to ask these questions at that time. my former employer was asking (and answering) them very well and did very well during 1998, and 2000.

[/ QUOTE ]

i want to expand on this a bit, if ya'll don't mind.

i don't mean to say here that they should have been looking specifically at the correlation issue (b/c that ended up being the problem and thus i can be accused of telling history backwards).

i mean to say that the model was/is extremely sensitive to small changes in the underlying assumptions. therefore, given what they had riding on the model (their entire portfolio), wouldn't it make sense to break the assumptions apart and test each one thoroughly? that is what i mean.

ANY testing of correlations reveals they are not static. ANY testing of volatilities reveals they too are unstable. at my old job, to sell clients on our portfolio construction process (For which we used a monte carlo simulation around the underlying assumptions vs. the traditional static estimation) we showed those above two in pictures. we showed the rolling correlation of the S&P500 with the US10yr and the rolling volatility of the S&P500 over 1,3,5,10, and 30 year time frames. seeing those charts immediately dispells any notion that either correlations or volatilities are static.

within the B-S-M world stability of those isn't required, but for simplicity/tractability often are required (we now have volatilities that are treated as random quantities and correlations that are functions of volatilities or some such specification etc. but overall, to be made mathematically tractable - i.e. to derive a closed form solution- they must all adhere to some gaussian distribution)

and thats all i had to say about that (for this line, please supply your own soft southern accent)

thanks,
Barron

[CrushinFelt]

I'm absolutely positive they knew that correlations were not static. They also likely knew that their positions were sensitives to small changes.

It wasn't a small change that did them in. It was the BIG change that screwed it all up.

According to what you're saying, people who lost a lot on leveraged positions in the Asian markets after the Kobe Earthquake would also be foolish for not having accounted for the proper probability of a catastrophic earthquake.

[DcifrThs]

[ QUOTE ]
I'm absolutely positive they knew that correlations were not static. They also likely knew that their positions were sensitives to small changes.

It wasn't a small change that did them in. It was the BIG change that screwed it all up.

According to what you're saying, people who lost a lot on leveraged positions in the Asian markets after the Kobe Earthquake would also be foolish for not having accounted for the proper probability of a catastrophic earthquake.

[/ QUOTE ]

they would be if they built houses & buildings upon which the lives of millions depended where a 1.0 richter (sp?) scale earthquake could destroy everything (vs. building structures for minimal more cost that would withstand stronger tremmers).

analagously, the MAIN problem i see with LTCM was simply that they took far too much exposure to a) areas in which they had no expertise, and b) in products where ANY decent sized move, regardless of hte correlation to the rest of the portfolio, would put their entire portfolio at risk (specifically, there were some non-trivially sized merger arb bets that were levered almost 100:1. i don't care what model or methodology you are using, that is unacceptable from a risk-reward standpoint).

thats it.

Barron

[niffe9]

I'm interning at an asset management company that sub-advises a large foreign blend fund. This fund lost a good amount using opportunistic hedging via equity index futures and I was asked to look into it. So far I've calculated the net gain/loss from the futures with the corresponding hedged positions and the correlation in ROI between the two. I'm also trying to look at the hedge ratios and holding periods for the futures. I'm thinking about what I can do to compare this strategy to the hedge everything (Tweedy Browne and others) and hedge nothing strategy. I was wondering if you had any insights on the issue. Thanks.

[DcifrThs]

[ QUOTE ]
If you are a complete 24 y/o n00b at anything investing and have $1500 that you want to begin with on E*Trade, what would you do?

[/ QUOTE ]

well what is your goal? $1500 isn't that much but it is something if you want to start testing the waters.

EDIT: well you asked what I would do so i'll tell you. i would start creating a passive portfolio that would earn the best risk adjusted return i could. i don't know how to most efficiently do this with $1500 but thats where i'd start (i.e. i'd worry more about the long term allocation of my money vs. the trading aspect).

but thats me.

Barron