The problem with financial modelling – comment by Joe Cambria

By skepticlawyer

Joe Cambria is a trader who has been watching blood spread all over the stockmarket floor during the last week. In the UK, short-selling of financial stocks has now been banned, although this move doesn’t address the fundamental problem of building wealth on debt. Joe has also spotted another problem – the increased reliance on financial modelling. Graphic is by DeusExMacintosh.

I think we all know what has been going on in the finance markets of late without having to go into it. There is more than enough evidence to suggest bankers were too heavily reliant on financial modelling in helping them price fairly complex asset instruments across various asset classes. According to sensible geeks banks continued to follow these models even when it was obvious there were red lights going off everywhere in terms of pricing and valuation.

I’ve worked as a trader for a while and saw my fair share of models used either for trading or valuation purposes. Trading models were basically useless as they were essentially trend following in various degrees. They made money when the trend was in full swing, but they gave all the money away when there was no trend. Valuation models to assess risk across markets arriving at a firm wide risk envelope were not only silly, they were actually quite dangerous.

Why then are we relying on models to predict climate change and adjust our way of life as a result? Are they more accurate than financial models in figuring the impact of GHGs in climate for a period of 100 years? The IPCC has handed out confidence levels of 90% as a result of models suggesting global temps will rise around  2 degrees over the next 100 years.

I highly recommend reading the link; it shows just how human minds can close down as a result of groupthink. I’m seeing speculative evidence this is also happening to climate scientists that mostly rely on models to make climate predictions.

We live in interesting times.

Update

LE notes that ASIC is also clamping down on short-selling of shares.

26 Comments

  1. conrad
    Posted September 19, 2008 at 5:32 am | Permalink

    JC learns the first trick of magic, to cause false associations with independent events.

  2. Posted September 19, 2008 at 6:27 am | Permalink

    I’m thinking Warren Buffet is in a position to get on his high horse and tell us all ‘I told you so’. He was the one who always reckoned you should never invest in something where you couldn’t figure out how it made its money. I took that to heart and have always gone for resources stocks… am I ever glad of that. I think MacBank might even be toast in the long run, unless they own enough ‘bricks and mortar’ assets to offset all the other smoke and mirrors.

  3. Posted September 19, 2008 at 8:08 am | Permalink

    And of course libertarians never suffer from groupthink. The financial crisis has arisen for a whole host of reasons but libertarians just keep saying it is the State that has caused this problem. Blinded by the Group.

  4. jc
    Posted September 19, 2008 at 9:42 am | Permalink

    LE

    margin lending is pretty plain vanilla stuff. It’s just lending against collateral which in this case is stocks etc.

    However, we’re now relying on modeling to predict future climate and make some really big changes to our lives using models that haven’t been able to predict the current short term cooling.

    This is madness.

    I’m not suggesting GHG’s aren’t a problem as they could very well be, but reliance on models is going to get us into trouble again: this time with taxes and other things that will cause major economic disruptions.

    It’s no good someone telling me financial models are different to climate models. They’re not really as they trying to reach the some end- predict the future. With climate models we’re trying to predict within a couple of degrees.

    This is nonsense.

  5. Posted September 19, 2008 at 10:12 am | Permalink

    I’m not suggesting GHG’s aren’t a problem as they could very well be, but reliance on models is going to get us into trouble again: this time with taxes and other things that will cause major economic disruptions.

    Much of biomedical science is based on modeling of biological processes. Should we abandon modern medicine? Modern aircraft are designed with computer models and fly straight off the plan. These types of modeling work much better than economic modeling because the models are built by relatively disinterested and unbiased creators. Economics will continue to be a guessing game because so much modeling is predicated on preconceived notions of what constitutes a good economy.

    Models can be useful even when they are wrong. The self-nonself model of immunology reigned for decades and is clearly wrong yet it does provide a good deal of explanatory power. The modularity model of the CNS is wrong but still provides explanatory power. Modern pharmaceutical drugs are built on the basis of models that are clearly incomplete but modern drugs are wonders to behold. Many physicists argue that the Standard Model of QM is incomplete but it has tremendous explanatory power.

    We have to create models. To make a blanket statement that because models are incomplete we should not rely on them is conceptual defeatism.

  6. jc
    Posted September 19, 2008 at 11:26 am | Permalink

    No suggestion we abandon modeling on anything , John. But it seems to me that climate modeling is being passed off as THE science.

    Boeing built a wing before and knows what a wing looks like before it started to model. They also do a lot of lab testing.

    I doesn’t seem to be happening in climate science as much as it should be .

  7. Jacques Chester
    Posted September 19, 2008 at 1:39 pm | Permalink

    These types of modeling work much better than economic modeling because the models are built by relatively disinterested and unbiased creators.

    It’s got nothing to do with bias. It’s about complexity. The laws of physics with regards to aerodynamics, structural integrity, etc are very well understood and completely deterministic. This makes them very easy to reliably model by computer.

    Things like weather, markets and the like are chaotic systems. We can generate general principles and make predictions, but unless one is careful, precision will be mistaken for accuracy.

    We get this in computer science too. Model building is, in a sense, what we’re all about. Yet it’s a chaotic system built from deterministic parts.

  8. jc
    Posted September 19, 2008 at 2:49 pm | Permalink

    Jacques, are the models used in climate science so much easier/reliable than those used to value a collateralized obligation?

    We’re being asked to steak 1% of GDP for the next 100 years or so on models that weren’t able to predict the most recent cooling patterns.

    Meanwhile banks went broke relying on math models telling them what was their inherent risk.

    This is nuts.

    ————————
    They weren’t economic models by the way…. whoever said that. They are different. These were valuation models.

  9. Posted September 19, 2008 at 3:59 pm | Permalink

    He was the one who always reckoned you should never invest in something where you couldn’t figure out how it made its money.

    Funny how this is ‘heretical insight’ not ‘well d’uh’.

  10. Posted September 19, 2008 at 5:52 pm | Permalink

    “Sure, we can try to make predictions, but they should not and cannot be treated as gospel.”

    That’s why climate scientists talk about balance of probabilities, error margins, uncertainty bands, confidence levels, and so on. The claim that climate models are presented as gospel (or whatever one’s favourite metaphor is) is simply wrong. Their shortcomings are constantly pointed out by the modellers, usually in the vain hope that the media will understand the inherent complexity and uncertainty, and thus convey that to the public. We all know how well the media does that.

    As for the models’ capabilities, they perform very well considering climate is a chaotic system. They do especially well with temperatures. If anyone really cares to know, the IPCC has it all laid out here (PDF). Those who don’t want to wade through that much text (though the graph on page 12 is instructive), you can always glance at Wikipedia.

    It’s also worthwhile noting that the models differ in degree, not substance. That is, they all point to global warming.

  11. Posted September 19, 2008 at 6:26 pm | Permalink

    fatfingers – I just had to let you out of the spam can – sorry! I have no idea what is going on with the spaminator. It chucked Adrien and John in there yesterday.

  12. DeusExMacintosh
    Posted September 19, 2008 at 7:48 pm | Permalink

    I think as FatFingers suggests, a lot of scientific ‘qualifications’ get left behind in the reporting and policy discussions about climate change. The people doing the actual research are up front about the impossibility of accurately modelling chaotic systems like financial markets or the environment, and as long as you accept that we’re talking about taking action JUST IN CASE we’re facing the worst case scenario it’ll put things a bit more in perspective. You are right that it there’s a certain amount of it being ‘sold’ as inevitable but that I think has primarily been a response to those whose vested interest in the status quo has discouraged investment in change. Now that even Bushie has admitted climate change is probably happening, perhaps the commentators and pressure groups will feel freer about qualifying their statements.

  13. jc
    Posted September 19, 2008 at 9:52 pm | Permalink

    fatfingers:

    That’s why climate scientists talk about balance of probabilities, error margins, uncertainty bands, confidence levels, and so on.

    Yeah, like who fats. Jim Hansen wants anyone bucking the view to be tried and sent to prison. His side kick, Algore calls it the moral imperative of our time. The IPCC ascribes 90% confidence weightings to their views that are based on about 1/2 dozen models.

    Who are these people that demonstrate humility, fats?

    The claim that climate models are presented as gospel (or whatever one’s favourite metaphor is) is simply wrong.

    Are you delusional? Anyone that doesn’t sign up to the cause is treated like a Holocaust denier.

    Their shortcomings are constantly pointed out by the modellers, usually in the vain hope that the media will understand the inherent complexity and uncertainty, and thus convey that to the public.

    What are you smoking, fats? See the above comment.

    As for the models’ capabilities, they perform very well considering climate is a chaotic system.

    Yea, point to one model that actually predicted the recent cooling. One, fats.

    They do especially well with temperatures.

    See above.

    It’s also worthwhile noting that the models differ in degree, not substance. That is, they all point to global warming.

    Gee, I wouldn’t have guessed. But please show us one model that predicted the recent cooling.

  14. conrad
    Posted September 20, 2008 at 5:20 am | Permalink

    JC,

    you are confusing poor interpretation of models with models. If Al Gore (or his equivalent on the other side) has some crazy interpretation of something, it doesn’t discredit the models, nor the real scientists (who you don’t see no TV) that do the work to generate and test them.

    The other thing you should realize is that the reason you should care about them more than financial or other types of models is that the risk profile is far different, and it’s completely asymmetrical for climate models. If we go too far down the left tail in both models (i.e., we are too conservative), we don’t lose too much (a few missed profits or money spent pointlessly reducing emissions). If we lose on the right side of economic models, then we have a recession and everything goes back to normal in a year or two. Alternatively, if we lose on the right side of climate change models, it’s the end of the world as we know it. So worries about this arn’t necessarily groupthink — it’s a reasonable reaction to potentially highly negative outcomes. It’s the same reason you get checked for cancer, despite the probability of you having it is small.

  15. Posted September 20, 2008 at 12:11 pm | Permalink

    LE:

    Perhaps the relevant question is: who is qualified to question the hypothesis?

    For example, I just put up a post on cannabinoids and Multiple Sclerosis. A number of countries now permit the use of medicinal cannabis. About time, the research makes it obvious that cannabinoids have therapeutic potential for conditions ranging from MS to brain tumours to atherosclerosis. Moreover this can be done without having to smoke cannabis or even get stoned because the non-psychoactive cannabinoid, cannabidiol, shows great promise. In relation to brain tumours the resistance of politicians and the some in the medical community is unadulterated bias and very distressing because over the last 20 years there has been virtually no progress in treating brain tumours. The incredible irony is that contrary to widespread opinion cannabinoids are first rank neuroprotective agents. Now ask yourself this: when was the last time you heard a positive report about cannabis in the Australian media?

    A few years ago a study group in Aus recommended the introduction of medicinal cannabis. Christopher Pyne offered no comment. Why? Because John Howard hates cannabis and because supporting the study conclusions could be perceived as supporting cannabis consumption. That is, his lack of response had nothing to do with the objective and relevant data and everything to do with politics.

    That is the problem with AGW, it has very significant economic implications which means it will become highly politicised. Economics often is, the Right Left dichotomy is very much about economic philosophy. A great weakness in economic reporting is the inability to remain aloof from political considerations. Bias is endemic and virtually impossible to eradicate because economics remains a science in becoming. Even reporting on solid science can be severely distorted so economics has little chance.

    With AGW there are so many wannabes throwing out their opinions that the average Joe and Jane quickly becomes discouraged in being able to know who to trust about the issue.

    There is one possible escape. There are first rate science journalists(eg. Phillip Ball who regularly writes for Nature, George Johnson is another one) and if push came to shove I would lean on their opinions rather than anything in the Australian media, Al Gore, Andrew Bolt or any other number of self proclaimed experts.

  16. jc
    Posted September 20, 2008 at 12:18 pm | Permalink

    Conrad:

    The entire IPPC framework is based on modeling. Can anyone show me any model that actually predicted the most recent cooling from 2002?

  17. Posted September 20, 2008 at 1:10 pm | Permalink

    JC,

    You might be erecting a straw man here. The IPCC was modeling climate not weather. Weather is short term, climate is long term.

    In any event I find the whole debate silly. As I have previously argued hot or cold eventually the Earth’s climate is going to radically change and the great technological challenge for humanity is Climate Control. Hot or Cold the consequences of the change will be catastrophic. But of course we’ll just keep arguing about a model instead of addressing the undeniable fact: Catastrophic climate change is inevitable and it may well happen in this century.

    Thus I return to my tired old point: we should not dismiss climate modeling we should pour resources into it because then and only then do we have any hope of preventing runaway greenhouse or a brutal and pulverising ice age.

  18. Posted September 20, 2008 at 1:13 pm | Permalink

    And while I’m at it! ….

    In paleoanthropology it is generally accepted that our species arose because with the commencement of the ice ages climate change was so severe that it played a major role in our evolution through the radical transformations in Africa’s climate and landscape. That’s what I mean by catastrophic.

  19. jc
    Posted September 20, 2008 at 2:07 pm | Permalink

    Not putting up anything, John. these models are supposed to be gods gift to science. Weather is a season or two. 2001 to 2008 is climate.

  20. conrad
    Posted September 20, 2008 at 5:38 pm | Permalink

    I suppose this is the type of cooling you are talking about JC:

    http://environment.newscientist.com/channel/earth/climate-change/dn14527-climate-myths-global-warming-stopped-in-1998.html

    Personally, I don’t think taking 2 points and looking at the gradient is helpful in a time series. In addition, if this year or then next turn out to be really hot, I’ll just assume you’ll stay consistent and won’t take it as a fluctuation on the upside.

  21. Sinclair Davidson
    Posted September 21, 2008 at 11:17 am | Permalink

    Imagine four quadrants with “easy” “complex” on the y-axis and “understand” and “don’t understand” on the x-axis. A lot of modelling falls into the “complex” and “don’t understand” quadrant, whereas modellers like to tell us that it falls into the complex – understand quadrant. The difference between financial modellers and climate modellers is that the financial modellers get caught out every so often.

  22. conrad
    Posted September 21, 2008 at 11:42 am | Permalink

    Here’s and alternative modeling view:

    Let’s say I have a black box that predicts some outcome 3% better than most people could year after year. I know little about how it works, apart from that fact it selects “good” over “bad”. I’m going to call that Warren Buffet’s brain. Should I ignore it?

One Trackback

  1. By forexbroker on November 8, 2011 at 8:04 pm

    forexbroker…

    […]Skepticlawyer » The problem with financial modelling – comment by Joe Cambria[…]…

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*