streetTRACKS Wilshire REIT (RWR)

All Comments on RWR

  • commenter
    Geoff
    Considine
    Jan 03 01:04 PM
    My Website
    Black Swans, Real Estate and Financial Stocks [view article]
    To VAMC:

    While testing models after the fact it is always a good idea to be careful. The model I used QPP was run with default inputs used in my articles for years. This is not a case of over-fitting. There is no reverse engineering issue here.

    The bottom line here is that you and everyone else are free to disregard available models because you feel they have the potential for 'black swan blindness.' You may as well argue that weather forecasts are pointless--and you are free to do so. Just carry your umbrella and slicker every day.     Reply
  • commenter
    Geoff
    Considine
    Jan 03 12:57 PM
    My Website
    Black Swans, Real Estate and Financial Stocks [view article]
    To dmnieren:

    I appreciate your enthusiasm, but you are not grasping the numbers. First of all, the probability of a cumulative 12-month decline is different than the various partial years, etc. Just to prove a point, I have gotten the 25-year history of C and analyzed the 2.5th percentile in returns for all 12-month periods. Guess what--its -36%! So, my analysis was predicting a substantially higher probability of the observed loss than history.     Reply
  • commenter
    Muzie
    Jan 03 09:36 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    vamc: Frankly I'd take Considine's flawed models over your vague wand waving any day of the week.

    I don't think Considine ever implied he tried to "predict all the things". Black swans will make any investment method fail whether they are based on fundamentals, technicals, psychology or whatnot.

    Yes, to one with a hammer everything looks like a nail, aka. the data fit problem.

    The flipside is to one with no (or inadequate) knowledge everything looks like randomness....     Reply
  • commenter
    vamc
    Jan 03 06:10 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    ceteris paribus - This might hold true when all the remaining things r constant. If it is not then what? Reply
  • commenter
    vamc
    Jan 03 06:01 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    The beauty lies in the eyes of the beholder

    The way you look at the market you can be insane or sane.

    Accept that one cant predict all the things and try to take advantage of the black swans.

    For that you need lot of patience finally that makes the difference between bleed or blowing up

    Reply
  • commenter
    SHIV DESAI
    Jan 03 05:24 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    I dont agree vamc. We are here (including Mr. Nasim Taleb) in the market to make money and not to prove which strategy is better and who is suffering for how many new biases. Whats the use of complex financial models if they are not able to yield the return and also of the black swan logic if I end up being paranoid and dont invest anywhere. ???? Reply
  • commenter
    vamc
    Jan 03 05:01 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    Mr. Geoff Considine :

    After reading your article i found the article suffers from
    1. cognitive dissonance
    2. Framework dependence
    3. Heuristic-driven bias
    4. Confirmation bias
    5. Round trip fallacy

    and the following things might also apply

    Empty-suit problem
    Future blindness
    Locke’s madman
    Reverse-engineering problem
    Black Swan blindness

    and

    Finally we are all fooled by randomness     Reply
  • commenter
    Muzie
    Jan 03 01:29 AM
    Black Swans, Real Estate and Financial Stocks [view article]
    dmnieren: I think Mr. Considine 1:40 odds are for any single one-year period... so of course over multiple years you will eventually hit a significant percentage. Besides you can't just compare a predicted one-year return with the cherry-picked arbitrary timeframes you've chosen, you're comparing apples and oranges.

    Mr. Considine: I'm sorry if my knowledge of statistics is relatively shallow - I'm assuming some kind of Monte Carlo simulation is at the heart of these predictions. Wouldn't it be possible to give a probability curve on the accuracy of the predictor itself, thus at least giving an estimate of how wide the black swan event window is?     Reply
  • commenter
    dmnieren
    Jan 03 12:42 AM
    My Website
    Black Swans, Real Estate and Financial Stocks [view article]
    I'll give you a point that Taleb sets up a bit of a difficult argument to prove or disprove, but that's not really his point. His point would be your model --Citigroup's decline is a low-probability event (as defined by you of 1:40 years)--is not real. It does not and cannot accurately quantify the likelihood of a 45% decline in C's stock. Your probability is based on past returns which do not account for the unknowable black swan even of the future (i.e. Enron-style fraud, Spitzer-like gov't interference, or worse). Therefore, your "prediction" that such a return is only a 1:40 event is useless.

    BUT the biggest nail in your silly probability coffin is that C had a >40% decline just 6 years ago in the August '01 to June '02 timeframe (less than a year; Black Swan=9/11). C had nearly a 50% decline from June -Sept '98 (Black Swan=Russia default). And another from June '90 to Oct. '90 (Black Swan=S&L?, Gulf War I?). And another from from Sept - Dec '87 (Black Swan=October '87 crash). And another from May '81 to March '82 (help me out on that one--interest rates?).

    FIVE such declines in the past 25 years. This is not a 1:40 event. Period. That is what Taleb was hammering on. These probabilistic models always underestimate the "tails" because stock returns do not follow a standard bell curve.

    By not correctly estimating your downside risk, you greatly increase the odds you'll get blown up. Period. That's one of the big themes in Taleb's work.
     Reply
  • commenter
    d_teller
    Jan 02 07:53 PM
    Black Swans, Real Estate and Financial Stocks [view article]
    I really appreciate G.Considine's essays. Sometime late last spring, there was an article/posting on SA about the unreasonably low yield of REITS (their prices having risen, and their dividends remaining, not 200 bp above Treasuries, but 200 bp BELOW!) Fair value was estimated at 40 -60 % below their market pricing of that time. At that time I sold off a fewer of the weaker ones. I had no idea of SIVs, had not purchased any CDO's, etc. I think the clue to the turn in the market's fundaamentals (in the U.S.) lay in that peak or bubble ... after the physical real estate bubble had been identified in late 2006. Reply
  • commenter
    Geoff
    Considine
    Jan 02 07:40 PM
    My Website
    Black Swans, Real Estate and Financial Stocks [view article]
    I agree with Mr. Meisel above.

    To FH: your comments suggest that you are not really familiar with modern financial modeling. Modern risk modeling is very useful in helping investors estimate the risk levels of their portfolios. These models are well-tested and benchmarked (the good ones at least). This does not mean that users should be naiive with regards to the limits of the models. They are far from perfect representations of reality--but they are still very useful.


    Mr. Taleb is correct that we need to understand the limits of models. Professional energy traders rely on weather forecasts--knowing that these forecasts decrease in accuracy as you go out in time. If I know that there is a 5% chance of 100+ Deg high temperatures in Chicago on a given day, I would manage my positions accordingly. I would not simply ignore the forecast because I knew that the forecast model was imperfect.

    Now, I do not hope to convince you or anyone else in such a short article or comments below. I do hope to stimulate interest in the solid use of portfolio management tools such as those that Mr. Taleb dismisses as useless. There is an enormous well-established discipline of quantitative portfolio management and I believe that the basic lack of understanding of this discipline is a major source of information asymmetry between investors. Companies like RiskMetrics, SunGard, and (on a much smaller scale) my own little firm provide well-tested tools. No reasonable practitioner would suggest that people have too much belief in such tools, but they are an important tool in one's arsenal.

    When David Swensen of Yale says that he is targeting a specific average return on the Yale endowment with a specific target standard deviation in return, where do you think he gets these numbers?

    If you want to get a sense of what it being done by practitioners, simply do a web search on "Basel II VaR."

    Have a look at the CFA curriculum if you want more information on modern methods, too.

    Reply
  • commenter
    Jim Kerns
    Jan 02 07:29 PM
    Black Swans, Real Estate and Financial Stocks [view article]
    I would posit the critical factor in resolving this delightful blog is: repeatability. Can QPP repeat in providing small but real probabilities to high impact events. Dr. Considine has shown repeatedly in his series of articles that QPP has correctly identified real, but small, probabilities to other "crashes". See his archives. The reason for this capability is the long-term equity risk premium model that is the basic operating principle of the market and QPP. QPP will continue to provide such results as long as the equity risk premium does morph into a black swan, itself. Also, Dr. Taleb has been adept at using tools similiar to QPP and would probably use such data to buy an out of the money put and collect a high impact reward! Reply
  • commenter
    Paul Meisel
    Jan 02 05:30 PM
    Black Swans, Real Estate and Financial Stocks [view article]
    There are several good points on both sides of this argument, overstated a little perhaps. Modern portfolio theory and its academic supporters imply a precision that is not warranted --- but some data is better than none. The statistical assumptions and methods behind QPP have some validity, but are not perfect either -- nothing in forecasting ever is!
    Reply
  • commenter
    FH
    Jan 02 04:47 PM
    Black Swans, Real Estate and Financial Stocks [view article]
    The problem is that you don't know that you don't know! Yeah, you came up with your numbers after running it through some overnight job on multiple CPU processors and to me, all you are saying is that you were able to "guess" that Citibank "may" drop by some pct. And what does it tell me? And why should I take this number more seriously than a random guess by a taxi driver? Because it knows more based on history?

    I think the basic premise of Taleb's book is that these models helps us get into a false world were we think we know! and that's dangerous..     Reply
  • commenter
    Geoff
    Considine
    Jan 02 01:54 PM
    My Website
    Black Swans, Real Estate and Financial Stocks [view article]
    I am in agreement with ZB and User 127562: people don't like to look hard at numbers. Mr. Taleb's sweeping rejection of portfolio theory is the wrong solution--he has created a rhetorical position that cannot be disproven. If models did predict non-vanishing probability of a certain scale of event, it's not a black swan (by his definition). There is no way to reject Mr. Taleb's hypothesis from a quantitative standpoint.

    Now, I agree with Susan, too: Mr. Taleb's arguments are often worthy of thought and discussion. When he says, quite bluntly, that portfolio theory and quantitative methods as a whole are useless (as in the quote above) and even itellectually lazy, he is taking it too far.

    I am a big fan of logic and philosophy, and Mr. Taleb often uses strategies that are ill-founded. By using LTCM as an example, he is saying "hey, look at this case where smart people put too much faith in quantitative finance" and extrapolating to saying that the entire edifice of portfolio theory is wrong. There is a big unsupported leap in logic here. The insurance and banking industries have used these types of models very successfully for quite some time--and will continue to.

    Anyway, this is a useful topic and the 'black swan' metaphor motivates good discussions. Thanks.     Reply
Loading...
Symbols:
Authors:
Quick Links: