Value at Risk: Enjoy the Cement Shoes

14 Jul

I have zero value to add here, besides to say that as a movement, can we all just fucking kill VaR, especially how it has/is currently used?  What a fantastically misappropriated “tool!”  Everyone, even financial novices know all about its limitations, but its still quoted with painful frequency.  This needs to stop.  VaR is Dead.  I am making it so.  Brothers, Sisters, and everyone in between, join with me to decry this nonsense once and for all!

Advertisements

16 Responses to “Value at Risk: Enjoy the Cement Shoes”

  1. Kid Dynamite July 14, 2010 at 9:13 am #

    but what’s better? VAR isn’t useless, AS LONG AS everyone knows it underestimates tail risks.

    • Anal_yst July 14, 2010 at 11:35 am #

      Why do firms report VaR, then? Everyone knows its incomplete (see my responses to other comments), but everyone still talks about it! Why?!?! Just because its convenient and easily quantifiable does not mean by any stretch of the imagination that it captures jack sh*t! Stop it! Bad banks! Bad SEC! Bad media! Stop pretending Var – in and of itself – is worth mentioning!

      • Byrne July 14, 2010 at 12:13 pm #

        Firms report lots of different incomplete measures. e.g. they’re willing to put a dollar value on inventory, or on an asset that’s in the fifth year of a twenty-year useful life, even though this is all really tough to estimate.

        Here’s something that would be interesting: how does reported VaR correlate with the size of future writedowns? You could get some really interesting data, though probably not enough to be statistically significant. I think what you’d find is that VaR measures risk for any situation that doesn’t lead to a bailout.

        So they should report VaR in a table that also includes how many of the company’s former employees work for the Fed, the Treasury department, etc…

      • Kid Dynamite July 14, 2010 at 1:43 pm #

        simply, it’s still useful as a RELATIVE measure of risk, vs an ABSOLUTE measure of risk.

  2. Carter the Examiner July 14, 2010 at 10:30 am #

    Can I disagree? VaR can be a useful too. The problem has become 1) management has forgotten how to use the tool, 2) naive observers treat the number as a maximum loss estimate rather than as the bound on a frequency measure (we will lose more the X Y% of the time), 3) many managers have failed to employ other complementary tools to examine risk.

    For a more thorough discussion of the tools of risk management, what worked, what didn’t and why, I’d recommend the Senior Supervisors Group 2008 report “Observations on Risk Management Practices
    during the Recent Market Turbulence” at http://www.ny.frb.org/newsevents/news/banking/2008/SSG_Risk_Mgt_doc_final.pdf

    Specifically, they write:

    “4. Informative and responsive risk measurement
    and management reporting and practices

    Firms that tended to avoid significant challenges through year-end 2007 typically had management information systems that assess risk positions using a number of tools that draw on differing underlying assumptions. Generally, management at the better performing firms had more adaptive (rather than static) risk measurement processes and systems that could rapidly alter underlying assumptions in risk measures to reflect current circumstances. They could quickly vary assumptions regarding characteristics such as asset correlations in risk
    measures and could customize forward-looking scenario
    analyses to incorporate management’s best sense of changing market conditions.

    Most importantly, managers at better performing firms
    relied on a wide range of measures of risk, sometimes
    including notional amounts of gross and net positions as well as profit and loss reporting, to gather more information and different perspectives on the same exposures. Many were able to integrate their measures of market risk and counterparty risk positions across businesses. Moreover, they effectively balanced the use of quantitative rigor with qualitative assessments. This blend of qualitative and quantitative analysis provided a high level of insight and consistent communications to management about evolving conditions,
    enabling the firm to pursue opportunities as they emerged and, more importantly, to reduce exposures when risks outweighed expected rewards.

    Firms that encountered more substantial challenges seemed more dependent on specific risk measures incorporating outdated (or inflexible) assumptions that management did not probe or challenge and that proved to be wrong. Some firms that lacked alternative perspectives on risk positions lost sight of how risk was evolving or could change in the future and what that might mean for the aggregate size of their gross versus net exposures. Some could not easily integrate market and counterparty risk positions across businesses, making it difficult to identify consolidated, firm-wide sensitivities and concentrations.”

    They go on to discuss VaR and stress testing in some depth.

    • Anal_yst July 14, 2010 at 11:29 am #

      Appreciate the response, especially since you very succinctly stated explicitely my qualms with VaR that I was too impatient to write earlier this morning:

      “1) management has forgotten how to use the tool, 2) naive observers treat the number as a maximum loss estimate rather than as the bound on a frequency measure (we will lose more the X Y% of the time), 3) many managers have failed to employ other complementary tools to examine risk. ”

      Perfectly sums-up what I’m talking about, but to elaborate why I’m suggesting firms – and more importantly, the media and the SEC – abandon VaR is just like the parent who lets his/her kid get a bb gun. Sure, you’ve trained your kid, and repeated ad nauseum all of the safety precautions for using the gun, and made it very clear there will be consequences if the gun is used inappropriately or worse, done so intentionally, but it doesn’t matter; eventually, either your kid or one of his friends is going to screw up, and you, as a parent are going to regret it (hopefully not to any extent that cannot be ameliorated!).

      I’ve always and loudly advocated that we should not blame the tools, but rather the improper usage of them, but at a certain point, as a prudent parent, we have to keep certain tools (instruments) out of our childrens’ hands.

  3. JM July 14, 2010 at 11:00 am #

    Hey. VaR isn’t the problem, it’s just an imperfect way to quantify downside. It is the synchronization of selling is the problem because it makes self-fulfilling prophecies happen. That is what markets do.

    It’s not about VaR. When stop-loss tools (which is at the core of all risk management) are homogeneous, everyone liquidates at the roughly the same time and crash risk rises. If everyone used a thumb in the wind indicator it would be the same.

    CVar, expected shortfall, variations on Frechet, whatever. They all do the same thing with varying tolerances. It is better than using a horoscope.

    • Anal_yst July 14, 2010 at 11:34 am #

      Excellent points! My point though (see my reply to Carter the Examiner) is that internal and external reliance on VaR has gotten out-of-hand. Why do firms have to report it? Why do they use it (independently or along-with other risk-management tools/methods)? Why is the media obsessed with it (don’t get me started)?

      I’m almost certain it’ll be replaced in common practice and parlance with some other incomplete tool/method, but I’d hope – perhaps naively – that by eliminating it from the above uses, it’ll make the point that risk management cannot and should not be boiled-down to one overly-simple, convenient, and horribly incomplete number.

      • JM July 15, 2010 at 1:49 pm #

        Affirm.

  4. professorpinch July 14, 2010 at 9:56 pm #

    Sorry I’m late to this party, but I had to weigh in on this.

    I think one of your points was brought up in Larry Mcdonald’s book when he described the Delta trade their desk had. They bought them for $0.18, but the firm was antsy. Why? Because VaR is, in its essence, like a lot of other tools out there: it tells where you’ve *been* and perhaps where you *are* – but not where you’re *going*.

    I have issues with using the normal distribution, because frankly, losses are not normally distributed. They are highly correlated and the distribution is non-normal. So why use the normal distribution other than for elegance and convenience? I have no idea.

    We need to have a way to quantify our risk and VaR, CVaR, etc. provide ways to do that. But we need to take a much closer look at the ways we describe losses and analyze how losses act/behave.

    Because the assumptions we use currently aren’t the right ones.

    • Anal_yst July 15, 2010 at 1:40 pm #

      Its not just about using the “correct” distribution (one of my issues with the paper on portfolio management/loan loss distributions we were talking about the other night).

      In order to quantify risk (which should also be very well qualified to boot) a variety of measures and tools are needed to get a holistic view.

  5. tiredAnd_Full July 14, 2010 at 11:04 pm #

    VaR in and of itself is not harmful when used in context with other measures (Stress Tests, etc). The problem is management at many banks have no comprehension of even the most basic statistical concepts. They look at a VaR of X, get themselves comfortable with that #, (“It’s only 1% of revenue”, etc) then move on confident that the risk is capped.

    I’m not sure what the solution is. VaR is useful in the everyday risk management of a portfolio, at the least to get some relative sense of what your loss’s could look like. However, when used as part of a banks limit structure, i’m afraid the simplicity of the measure exposes the naivete of management and acts as an enabler of out-sized and irresponsible risk taking.

  6. Anal_yst July 15, 2010 at 2:39 pm #

    Agreed, and agreed re: paragraph 1.

    All the criticism of VaR has already been made by people far more knowledgeable than I, so I’m not going to rehash it here, but because it encourages so many bad habits, I suggest we should do-away with it entirely.

  7. capitalistic July 21, 2010 at 7:27 pm #

    I think VaR is essential, in terms of taking a snapshot. However, I believe that using historical figures to predict the future is pointless, and good risk managers know this.
    It’s a like a retail store owner who decides to sell his inventory at a 30% mark down at the end of the month. His decision has nothing to do with historical data. It has to do with his interpretation of real time sentiment.

    • Anal_yst July 22, 2010 at 4:01 pm #

      I don’t think its absolutely essential, especially considering how often/easily its (ab)used as a shortcut.

      The point is that if you absolutely insist on using VaR, then it should ONLY be used as one part of a holistic risk management approach.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: