Taleb, i modelli e il letto di Procuste

Oggi vi propongo un video di Nassim Taleb sui limiti dell'approccio probabilistico in economia e finanza e più in generale sul problema della decisione in condizione di incertezza. Come sempre è un polemista divertente e in questa conversazione con l'Economist se la prende con gli eccessi del data-mining ("the more statisticians you put on a job the more false results you get...the spurious relationshisps rise to the surface") e in generale con la presunzione di poter prevedere il mercato a proprio vantaggio ("It's very hard just by skills to become a trillionaire or a billionaire...you need skills plus a lot of luck...luck is more valuable, for very large deviations"). Secondo Taleb le società (e gli investitori) devono organizzarsi in modo da essere "immune to forecast errors". Proprio per questo è importante non dimenticare che ci sono anche sistemi (e strategie di investimento) che invece di essere fragili sono anti-fragili, "systems that benefit from shocks". Ma il vero colpo di genio polemico è verso la fine, quando paragona un certo modo di usare i modelli in economia e finanza al letto di Procuste. Un uso che, combinato con l'eccesso di leva finanziaria, ha condotto il sistema alla crisi del 2008.

Ancora sulla Grecia e gli scricchiolii dell'euro. Due video sugli alchimisti di Wall Street.

John Mauldin dedica gran parte della sua newsletter settimanale ad un'analisi della crisi greca. Ecco la sua difesa degli hedge funds dall'accusa di cospirare contro l'euro: 

The Euro and a Conspiracy of Hedge Funds

The lead story in this morning's Wall Street Journal is that hedge funds are holding "idea meetings" and deciding that shorting the euro is a good bet. Der Spiegel called them "secret meetings," as if somehow a cabal of hedge funds is conspiring to push the euro down. A few points for the writers of Der Spiegel:

• There is no secret about the problems with the euro. Let's see, when the head of Germany's leading debt-management agency warned this week that the euro would collapse if any member defaulted on its debt, was he part of a secret conspiracy? If he is right, do you want to bet that Greece will behave, and go long the euro?
• The currency market is a $2 trillion dollar a DAY market. That's over $50 trillion a month. Even with 20:1 leverage, $50 billion in hedge funds shorting the euro is a drop in the bucket, and I seriously doubt anywhere close to that much is at risk. George Soros won his bet against the pound sterling because the pound was fundamentally flawed and overvalued, and he put his money where his mouth was.
• If a hedge fund is betting against the euro, someone has to be on the other side of that trade. Are those guys (on the other side) conspiring in secret to drive the euro up and the dollar down? Are they in "secret" meetings to take advantage of the poor, dumb, misinformed hedge funds? Who are they? The world needs to know who is conspiring against the dollar and other currencies! Whatever. One side will be wrong. Fundamentals will out.
• I get invited to "idea dinners" from time to time. They are indeed private, but they don't rise to the level of "secret." I do very little trading, but these meetings help to hone my ideas, and I hope that helps make this letter a better source for you.

The Journal wrote that these hedge-fund managers expect the euro to go to parity with the dollar, as if that is some novel idea. I made that prediction in 2002 when the euro was at $.88, suggesting that it would rise to $1.50 and then fall back to parity by the middle of the next decade. Maybe it will get there a little faster than I thought. Stay tuned, and I do NOT suggest making 20:1 bets on currency moves. A lot of those hedge funds will lose a lot of money if the market moves against them.

Da consumato professionista mescola argomenti ottimi con argomenti meno buoni (ad esempio le considerazioni sulla leva sono risibili e trascurano le nonlinearità). Sono però totalmente d'accordo con lui con le considerazioni su Soros e la sterlina e sul fatto che la sopravvalutazione dell'euro (almeno quando viene scambiato a 1.5$) non è un mistero per nessuno.

Sempre Mauldin propone una semplice (forse troppo semplice) ma interessante analisi della struttura dei tassi a 10 anni delle obbligazioni governative dei paesi dell'area euro a cura di GaveKal dalla quale si è portati a concludere che:


Thus, for all of the complaints about market manipulation, it seems that the hierarchy of spreads over German Bunds has followed, since the beginning of this year, a pretty rational walk. Actual debt levels and short-term pressure on government accounts have systematically explained more than 85% of yield spreads. When a liquidity risk premium is applied, the explaining power of our model rises to above 95%. A very interesting predictability which leads us to the following conclusions:

1. The hierarchy of sovereign spreads on Euro financial markets closely follows the logic of the Stability and Growth Pact (SGP). The European bond markets are thus remarkably consistent and sending a clear and powerful message to the governments of the euro zone to stick to the agreed rules or suffer the consequences.
2. As of today, there seems to be no additional risk premium related to the possible dislocation of the Eurozone. Clearly, this possibility would have such devastating effect on world financial markets that investors cannot even think of it (even if many talk about it).
3. Based on the fitted curve we can identify a few markets that potentially deserve either a lower or a higher spread than the one currently prevailing. Of course, one should not build too much on small deviations to the fitted curve, but one sovereign bond market looks particularly expensive, namely Belgium. The OLO market should normally trade with yield significantly above that of Austria, or even Italy (see The Italian Job). The credentials for the management of the Belgian debt crisis of 1993 may play a role in the overvaluation of the OLO market. But with Belgian public debt now rising again (it should pass the 100% mark next year) and with the Belgian banking system in deep crisis, we see little reason to hold any OLO in bond portfolios. Rather, we would advise to sell OLOs against a basket of Austrian and Italian bonds. Aside from Belgium, France and Ireland are the two markets that currently look vulnerable.

4- The Return of Country Risk

As mentioned above, it is hard to foresee how the current situation in Europe will play out, but the one thing we can be sure of is that the crisis is an important turning point for European investors in that it marks the return of country risk. Indeed, regardless of what the European Union may do to help Greece through its current crisis, the new reality is that Greece's funding costs (along with those of other European nations) will, from now on, increasingly be a reflection of Greek fundamentals rather than German fundamentals. This must mean that, like a phoenix rising from the ashes, country risk in Europe is all of a sudden back from the dead. It also means that discerning which countries are set to experience a rise in financing costs, and which countries enjoy a pull-back will once again be a driver of relative stock market performance. In that regards, we hope that our reader will find the equation we offered above both interesting and useful.


L'equazione alla quale viene fatto riferimento nel testo è il risultato di una semplicissima regressione
lineare su tre fattori:


since the beginning of the year, we find that, on average, 10-year bond yields in the Euro area have been organized according to the following pattern:


Yield = 4.31+ 0.0201*(excess debt)+0.4723*(EDP adjustment)-0.564*(liquidity)


Attenzione che la serie impiegata è davvero molto corta, ecco perchè ho enfatizzato in rosso il caveat iniziale.


Vi segnalo infine due video sul ruolo avuto dai Quants nella crisi finanziaria: il primo è di Andrew Lo: ResearchChannel - Are Mathematical Models the Cause for Financial Crisis in the Global Economy e potete scaricarlo dal sito della National Science Foundation a questo link

Il secondo video è un documentario (teoricamente in olandese ma tutte le interviste, cioè oltre il 90% del contenuto, sono in inglese) Quants - the Alchemists of Wall Street  con interviste a Paul Wilmott, Emanuel Derman, Matthew Goldstein e Mike Osinski. Tra gli spunti interessanti di questo documentario mi ha colpito la descrizione che Willmott fa al 19esimo minuto di strategie note come "raccogliere monetine davanti al rullo compressore"...in cui piano piano fai profitti
esponendoti a un rischio estremo (inciampare mentre raccogli la monetina di fronte al rullo compressore che avanza piano piano ma inesorabilmente...).  Sarei davvero contento di scoprire che tra i miei 25 lettori c'è qualcuno attivo nel mondo Hedge e  conoscere la sua opinione sui commenti di  WIllmott e in particolare a proposito della sua affermazione seguente:

"I think it is fantastic that people that people who take risk should be compensated for taking risk. But only if they are
taking risk themselves...taking risk with other people's money you should not be compensated for... I am sorry...I do not know if that fits with economic theory but taking risk with other people's money does not get rewarded"

Financial Armageddon? Aggiornamento al 12 febbraio 2010

Se non avete ancora preso le vostre decisioni di investimento per il 2010 e siete interessati all'opinione di un pool di gurus di tutto rispetto potete ascoltare il lungo dibattito Investments: how to earn in 2010? What are the risks? tenutosi a Mosca una settimana fa: ecco in breve i loro punti di vista

Michael Power: Put your money in Singapore dollars until mid-year. Then invest in emerging markets, coal and consumer stocks in emerging markets with current account surpluses. Avoid the West. Avoid banks in the West. Put some money in frontier markets. Chinese growth will continue to be driven by the rise of the middle class.
David North: Buy protection on credit, as markets are overcooked (e.g. CDS on banks). Buy dividend futures, especially in the UK, as cash flows are high. Take the opportunity to bet against rate increases, as they will not happen.
Michael Gomez: Invest in places that will benefit from rebalancing. Buy Asia and Asian currencies, which should appreciate. Avoid the euro and emerging European currencies. Europe is prone to defaults, given weak balance sheets. Invest in strong balance sheets in countries with good monetary policy and inflation targeting, such as Brazil. You can buy 5y bonds there at 13%. Among G7 countries, Germany is in favor. There is a need to differentiate in emerging markets and buy Brazil over Turkey, for example.
Nassim Taleb: Go no-risk with 80% percent and high-risk with the rest. For the speculative component, short the S&P and be long in precious metals; bet on hyperinflation with OTM calls on gold and puts on bonds; short USTs as long as Bernanke and Summers are in office; trade on the breakup of Europe. Russia is stable.
Hugh Hendry: Think of Keynes’ “Economic Consequences of the Peace”. The euro is the new gold, and governments cannot repay. The amount of debt taken on is enormous and still highly risky, as none of the debt has gone. Mr Hendry would not take any risk now. Debt will squeeze the vitality out of economies. Buy the idea that rates will not rise. The policymakers will not be able to create inflation, as monetary policy does not work. People are all long risky assets as a hedge against inflation and short treasuries, which is dangerous. He likes land and agricultural equities in the long term, but they are risky in the short term. The problem with Asia is excess capital and low returns on capital. It makes no sense for the Chinese yuan to be so cheap. China could fail and go from first to last, as its economic model is unsustainable.
Ashot Khachaturyants: Buy Russia, especially oil stocks with exposure to East Siberia. Nanotechnology and project financing are promising.
Marc Faber: The consensus is to buy Asia. Japan is a good idea as a huge economy where the market has been down for over 20 years. When inflation comes, avoid currency and bonds and buy equity and foreign assets. One should buy land as civilization may collapse; farms in Argentina are promising. Water is a real issue, especially in India. Chinese economic growth is likely to slow this year.

Dopo che avrete deciso come investire, e su quali collassi puntare (se ascoltate Taleb) sarete pronti
a prendere in attenta considerazione le profezie di TraderMark su SeekingAlpha:
HFT, Algorithmic Trading, l'ombra di Jim Simons (al quale va la mia ammirazione: leggete qui) che oscura la vita sul pianeta, altro che inverno nucleare...sentite qui e immaginatevi lo S&P500 che perde il 22% in pochi minuti...

For those who have been around a while you know I constantly refer to the "supercomputers at the hedge funds" controlling things or at least being the marginal decider of prices. As a participant in markets for a while now I have to say some of the things we're seeing the past year or so are beyond compare.
My thesis has been the quantitative hedge funds really have changed the nature of the markets and trading. The most successful and famous is Renaissance Technologies, led by Jim Simons. The track record of success there has been fantastic, and it's all computer driven. Success begets copy cat behavior - and a flood of funds trying to replicate the grand chief have been born. Hence when I refer to "algorithms" dominating trading, I am speaking to this bevy of pooled capital, all doing (or trying to do) almost the exact same thing and taking stocks farther (both higher and lower) than makes any logical sense. And this, in my opinion, is simply crowding out people who use fundamentals or logic. The machinations of August 2007 really was the first time this hit me in the face as I was seeing action that were in no way explainable by any reasonable data point. Much of it was "liquidations," i.e. hedge funds over levered, and much like an individual investors gets a margin call - so did they. So they had to sell what they could, not what they wanted to, as many hold esoteric positions that had no market (hedge funds don't just own simply stocks and bonds). So they sold off the liquid parts of their portfolios. And once a selling begins, a waterfall effect hits as technical conditions are triggered in computer after computer throughout New York City (and in fact the world) and you get this cascading effect.

I am of the belief that eventually something will go very wrong in a system like this, on a much larger scale than August 2007 or some of the cascading selloffs in 2008. "They" assure us they provide liquidity. But with some of these computers doing 1000s of trades a second, and so many quants doing the exact same thing - all set to hair trigger off each other - it is very easy to imagine a scenario where October 19, 1987 happens not in hours but in minutes. Hopefully I am very wrong on this, but I certainly don't take any solace in the assurances from either the quants themselves or their regulators (who are either asleep at the wheel or captured), that it can't happen. These are the same groups of people who assured us they had everything under control and all risk was arbitraged away pre-2007. There is no way to assure anything with a disparate electronic ecosystem spread over countless desktops storage facilities.

Sempre sul ruolo dei quants e sul prop trading ecco il sig. Patterson, giornalista del WSJ e autore di
The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It:

 
Certo di avervi rovinato il sonno per qualche settimana mi dichiaro soddisfatto e vi lascio con l'immancabile aggiornamento al 12 febbraio.