SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : The *NEW* Frank Coluccio Technology Forum -- Ignore unavailable to you. Want to Upgrade?


To: Frank A. Coluccio who wrote (31332)9/13/2009 11:31:13 PM
From: axial  Respond to of 46821
 
If, then, else... three words for contingency in the programmer's lexicon.

We can't even get the weather down: too many variables. We express outcomes as probabilities. The mere existence of the phrase "butterfly effect" demonstrates that we also understand a minor perturbation can induce chaotic effects that exceed the input by orders of magnitude. That, in a system where no constituent exhibits volition. So now let's model a system where each part of the system is independently driven: not just driven, buts also motivated. We make an illogical leap:, where we know and concede that we can't model an inanimate system like weather, to betting trillions on a vastly more complex - and animate - system like economics.

Would you bet a billion on 80% probability? How about a leveraged bet? How about a leveraged bet, using someone else's money? Would you create incentives for such betting, and institutionalize it?

Wait a minute; that bet is based on expected human behavior. It's not based on unanticipated events, or even worse, concatenated behavior arising from those events: the so-called domino effect. What's more we have concatenated behavior by the actors in the box, and the actors outside the box, who profess to be modeling everything "inside". Nope, hold it. That doesn't work, because the box doesn't exist; they're both part of the same continuum.

"Another harmful abstraction used in macroeconomic models to facilitate growth forecasts is to leave out the banking system and the buildup of related financial assets such as mortgage and other consumer debt. Rather than attempt to understand and model the complexities of the financial sector, economists simply assume it away."

What about economic and financial mindsets? After all, there's considerable variance among economists; how much is that variance represented in policy and practice among the biggest national players? The prevailing view at the Fed has been enforced rigidly for decades; dissenters have been ejected. The Fed, and thus Wall Street, have been ruled by a system of belief in which there are only captive orbiters - no one else need apply. In the century of energy abundance these people have established an outsize effect on the world's economic thinking and financial practice.

But are they right? Or is the model, the mindset derived from living in a century of plenty now fading from relevance? And is it "groupthink"?

It's all models: insufficient abstractions in words and practice. Adam Smith, Bagehot, von Mises and Marx have all sought to render vast complexity into coherent expression. The subset of economic orthodoxy is current financial practice.

The new orthodoxy is that banks are players. Are they? A bank's customers invest, speculate, and gamble. That's their business. The bank's business: to lend and protect. A banker is responsible for the welfare of others; prudence is obligatory.

Ordinary people could instruct these high flyers: don't bet more than you can afford to lose, and always set aside something for a rainy day. Because you never know. But heck, that's just the view of ordinary people, a distillation of centuries of experience. It's not quantitative esoterica, just common sense.

Derivatives are a two-edged sword, with no handle. Those who use them must have a demonstrable interest, and those who offer them must pledge hard assets. Yet in a crashing market, even the value of hard assets declines, and therein lies the paradox: derivatives will fail just when they're needed. Thus, they increase specific risk - and systemic risk - in catastrophic events.

How many times, in how many ways, must we see derivatives fail before we understand?

'"Economics is just entertainment," says Naseem Taleb, a former options and derivatives trader and author of a best-seller on finance The Black Swan. Taleb has been highly critical of the risk management practices employed by financial economists at investment banks leading up to and during the recent financial crisis. Inside the banks, experts in finance were making dangerous assumptions of their own involving statistics or probability theory. These experts, with their PhDs in finance, were getting paid millions of dollars to manage risks they either did not understand or chose to ignore. They turned a blind eye to the frequency and magnitude of what Taleb calls rare events -- perfect storms in finance -- because they are difficult to model."

Prudence is a distant dream. We're dancing in minefields.

Jim

---

[All quotes from: Economics: Dismal Science or Science at All?]

financialpost.com




To: Frank A. Coluccio who wrote (31332)12/29/2009 2:15:12 AM
From: axial  Respond to of 46821
 
Frank, on wrong-headed application of science and technology, worth noting:

Financial instruments could be spiked with unfindable risks

Message 26207149

Note the nested link: Credit Derivatives Are Not 'Insurance'

Message 26192161

From the parent post: "The risk models proved myopic, they say, because they were too simple-minded." There's an assumption of innocence in that statement.

---

From Rong Ge's derivative FAQ - cs.princeton.edu

5. Does your paper imply that CDOs and related products should be banned?

For now, that would be a hasty conclusion. One can imagine some free market solutions to the issues raised in the paper such as (a) turn over design of CDOs to a neutral 3rd party to avoid cherry-picking. (But, keep in mind the experience with rating agencies in recent years.) (b) use CDOs with parameters where the planted densest subgraph problem is easy. (But, keep in mind that the market currently lacks full transparency, so it is unclear how to even obtain the graph on which to solve densest-subgraph!)

Of course, one must then investigate if these new fixes allow other kinds of manipulation. Our paper is only a first cut at establishing a role for computational complexity in thinking about derivatives. Hitherto this aspect has been ignored.


---

Jim