interceptor

Novas mensagens, análises etc. irão se concentrar a partir de agora em interceptor.
O presente blog, Geografia Conservadora servirá mais como arquivo e registro de rascunhos.
a.h

Thursday, August 16, 2007

Gambling on tomorrow







Science & Technology

Statistics and climatology
Aug 16th 2007

From The Economist print edition
Modelling the Earth's climate mathematically is hard already. Now a new difficulty is emerging


Illustration by Dettmer Otto


“SCIENCE” is a recently coined word. When the Royal Society, the world's oldest academy of the discipline, was founded in London in 1660, the subject was referred to as natural philosophy. In the 19th century, though, nature and philosophy went their separate ways as the natural philosophers grew in number, power and influence.
Nevertheless, the link between the fields lingers on in the name of one of the Royal Society's journals, Philosophical Transactions. And appropriately, the latest edition of that publication, which is devoted to the science of climate modelling, is in part a discussion of the understanding and misunderstanding of the ideas of one particular 18th-century English philosopher, Thomas Bayes.


Bayes was one of two main influences on the early development of probability theory and statistics. The other was Blaise Pascal, a Frenchman. But, whereas Pascal's ideas are simple and widely understood, Bayes's have always been harder to grasp.

Pascal's way of looking at the world was that of the gambler: each throw of the dice is independent of the previous one. Bayes's allows for the accumulation of experience, and its incorporation into a statistical model in the form of prior assumptions that can vary with circumstances. A good prior assumption about tomorrow's weather, for example, is that it will be similar to today's. Assumptions about the weather the day after tomorrow, though, will be modified by what actually happens tomorrow.

Psychologically, people tend to be Bayesian—to the extent of often making false connections. And that risk of false connection is why scientists like Pascal's version of the world. It appears to be objective. But when models are built, it is almost impossible to avoid including Bayesian-style prior assumptions in them. By failing to acknowledge that, model builders risk making serious mistakes.
Assume nothing

In one sense it is obvious that assumptions will affect outcomes—another reason Bayes is not properly acknowledged. That obviousness, though, buries deeper subtleties. In one of the papers in Philosophical Transactions David Stainforth of Oxford University points out a pertinent example.
Climate models have lots of parameters that are represented by numbers—for example, how quickly snow crystals fall from clouds, or for how long they reside within those clouds. Actually, these are two different ways of measuring the same thing, so whether a model uses one or the other should make no difference to its predictions. And, on a single run, it does not. But models are not given single runs. Since the future is uncertain, they are run thousands of times, with different values for the parameters, to produce a range of possible outcomes. The outcomes are assumed to cluster around the most probable version of the future.
The particular range of values chosen for a parameter is an example of a Bayesian prior assumption, since it is derived from actual experience of how the climate behaves—and may thus be modified in the light of experience. But the way you pick the individual values to plug into the model can cause trouble.
They might, for example, be assumed to be evenly spaced, say 1,2,3,4. But in the example of snow retention, evenly spacing both rate-of-fall and rate-of-residence-in-the-clouds values will give different distributions of result. That is because the second parameter is actually the reciprocal of the first. To make the two match, value for value, you would need, in the second case, to count 1, ½, ⅓, ¼—which is not evenly spaced. If you use evenly spaced values instead, the two models' outcomes will cluster differently.
Climate models have hundreds of parameters that might somehow be related in this sort of way. To be sure you are seeing valid results rather than artefacts of the models, you need to take account of all the ways that can happen.
That logistical nightmare is only now being addressed, and its practical consequences have yet to be worked out. But because of their philosophical training in the rigours of Pascal's method, the Bayesian bolt-on does not come easily to scientists. As the old saw has it, garbage in, garbage out. The difficulty comes when you do not know what garbage looks like.

Monday, August 13, 2007

Além das ideologias

Esta história de separar política de economia, cultura de sociedade etc. pode até ter algum sentido na formulação de alguma teoria, academicismos etc., mas quando se pensa no óbvio, um problema que seja, ele não vêm separado no dia-a-dia: olha, esse aqui é um "problema social", este outro é "econômico" e por aí vai... Este "analiticismo" é herança positivista. Coisa mais engraçada é quando em alguma faculdade qualquer vemos um "Professor Doutor Coisa e Tal..." dizer, ou melhor, enrolar que chega a dar sono, dando voltas e voltas com seu academicismo na razão inversa ao que tem para dizer. Daí se recorre a todo um jargão, toda uma novilíngua "científica" (existe "ciência humana?”) para dizer o... óbvio!

Esses caras da Stratfor estão de parabéns, pois dizem o óbvio: "uma nação é uma nação" e não há como separar problemas de segurança nacional de problemas econômicos. Isto é Realismo puro. Só com esta visão é que podemos tentar compreender o que ocorre quando um país invade outro, pois qualquer ação desta monta é, antes de tudo, um investimento. O mais ideológico dos líderes fosse um Hitler, não poderia desconsiderar suas indústrias. A ação terrorista árabe conta com fanáticos dirão os missivistas, mas o Hamas e o Hezbollah são financiados; a Revolução Islâmica no Irã também teve um expediente econômico, a transferência de ativos de multinacionais aos bazaarin, "capitalistas nacionais". As burocracias encravadas nos pútridos estados terceiro-mundistas defendem com unhas, dentes e palavras que acusam conspirações, seus privilégios em detrimento da maioria que poderia ser contemplada pela liberalização dos mercados.

É sintomático que sites como o Mídia Sem Máscara, p.ex., neguem tacitamente as exigências da realidade. Certa vez quando escrevi um pequeno artigo que analisava como o continente africano era abordado em uma prova de vestibular, tive a primeira versão de meu texto cortada neste trecho:

Além das ideologias

Aliás, entender a geopolítica regional (e a global) segundo categorias ideológicas se torna falacioso:

“O apoio da Argélia à Frente Polisário [no Marrocos] não é devido ao respeito excepcional ao direito dos povos à autodeterminação. Como explicar o apoio de israelense, até 1977, ao regime etíope que reivindica o marxismo-leninismo a não ser pelo interesse de Israel em não ver o mar Vermelho se tornar um lago árabe? Os interesses de Estado determinam as políticas. Do mesmo modo a China de Mao Zedong apoiava a FNLA e a União Nacional para Independência Total de Angola (UNITA), também ajudados pelos Estados Unidos e pelo Zaire [atual República Democrática do Congo] porque os dois movimentos se opunham ao MPLA, apoiado pela URSS. Se a dimensão ideológica fosse determinante, a FPLE teria sido apoiada pelos países “‘socialistas’”.

Da mesma forma é interessante observar a posição européia, aparentemente ambígua do ponto de vista ideológico. Quando os movimentos revolucionários africanos tinham ingrediente nacionalista era de se esperar algum apoio europeu. Mas, sua oposição aos regimes nacionalistas devido ao seu controle de capitais, os empurrava para a URSS, mesmo quando contrários ao marxismo-leninismo. Já os EUA e sua posição visceralmente contrária à influência soviética os tornava mais pragmáticos.

A ideologia é totalmente inócua para se entender a África. A geopolítica e a razão de estado é que a explica.

O MSM é antimarxista e pró Israel, dizem. Então, como explicar que Israel apoiara, conjunturalmente, um regime comunista? Cortaram ao que protestei e a forma original do artigo acabou ficando.
----------------------------------------------------------------------------------
Subprime Geopolitics
August 13, 2007 19 17 GMT

By George Friedman

The subprime crisis is worth analysis in its own right, though it also gives us the opportunity to discuss our own approach to economic issues. Stratfor views the world through the prism of geopolitics. In geopolitics, there is no such thing as separating a country's economy from its national security or its political interests. A nation is a nation. Academic departments divide themselves nicely into areas of study. In the real world, things are much too intertwined and sloppy for that. Geopolitics views the international system and nations as consisting of a single fabric of relationships, with economics being one of the elements.
Not all events have geopolitical significance. To rise to a level of significance, an event -- economic, political or military -- must result in a decisive change in the international system, or at least a fundamental change in the behavior of a nation. The Japanese banking crisis of the early 1990s was a geopolitically significant event. Japan, the second-largest economy in the world, changed its behavior in important ways, leaving room for another power -- China -- to move into the niche Japan had previously owned as the world's export dynamo. The dot-com meltdown was not geopolitically significant. The U.S. economy had been expanding for about nine years -- a remarkably long time -- and was due for a recession. Inefficiencies had become rampant in the system, nowhere more so than in the dot-com bubble. The sector was demolished and life went on. Lives might have been shattered, but geopolitics is unsentimental about such matters.
The Russian default of 1998 was a geopolitically significant event. It marked the end of the post-Cold War period and the beginning of the new geopolitical regime that is increasingly showing itself in Russia. The global depression of the 1920s and 1930s was enormously significant, transforming the internal political and social processes of countries such as the United States and Germany, and setting the stage for political and military processes that transformed the world. The savings and loan (S&L) crisis of the 1980s had no real geopolitical effect, and the collapse of Enron meant nothing. However, the consolidation of Russian natural gas exports under Gazprom's control is certainly a major change.
The measure of geopolitical significance is whether an event changes the global balance of power or the behavior of a major international power. Looking at the subprime crisis from a geopolitical perspective, this is the fundamental question. That a great many people are losing a great deal of money is obvious. Whether this matters in the long run -- which is what geopolitics is all about -- is another matter entirely.
The origins of the crisis seem fairly clear. Traditionally, when banks look at mortgages on homes, they carefully study the likelihood that the loan will be repaid, as well as the underlying collateral. Their revenue and profits come from the repayment of the loan or the ability to realize the value of the loan through the forced sale of the house.
Two things changed this simple model. The first started a long time ago. Encouraged by the federal government, banks that issued mortgage loans began selling those loans to other entities. This, then, created a large secondary market in bundled mortgages -- huge numbers of mortgages grouped together and sold and traded as if they were simply financial instruments, which, of course, they are.
As a result, banks began to view mortgages less as long-term investments than as transactions. They made their money on closing costs, rapidly selling the mortgages to aggregators, which in turn passed them on to others. The banks then loaned the money again. The more mortgages banks racked up, the more money they made. The risk was transferred to others.
In the past few years, two new groups of players entered the scene, one on either end of the spectrum. The first group comprised mortgage companies and brokers, nonbanking institutions whose business model was built primarily around the transaction. The brokers in particular had no skin in the game. Every time they executed a mortgage, they made money. If they didn't execute one, they didn't make money. The role of evaluating the borrower increasingly fell to these entities, neither of which was going to hold on to the debt instrument for more than a moment.
The second group was the final buyers of bundled mortgages -- increasingly, hedge funds. Hedge funds are monies gathered from various "qualified" investors -- otherwise known as rich people and institutions. They are private partnerships, so what they do with their money is between the managers and partners. No federal agency is responsible for protecting the private placement of money by the wealthy.
In a world of relatively low interest rates, wealth-seeking investors flocked to these hedge funds. Some of the older ones were superbly managed. The newer ones frequently were not. With a great deal of money in the system, there was a restless search for things to invest in -- and the secondary market in subprime mortgages appeared to be extremely attractive. Carrying relatively high rates of return, and theoretically collateralized by fairly liquid private homes, the risks of these deals appeared low and the returns on the mortgages -- particularly when you looked at the contracted increases -- seemed extremely attractive.
The fact is that no one really worried about defaults. The mortgage originators that prepared the documentation for these riskier loans certainly didn't care. They just wanted the mortgages to go through. The primary lenders didn't worry because they were going to resell them in hours or days anyway. The mortgage aggregators didn't care because they were going to resell them, too. And the final holders didn't worry because they assumed the system would permit easy refinancing of loans at sustainable interest rates, and that -- in a worst-case scenario -- they at least owned a portfolio of houses that they could bundle and sell to real estate companies, perhaps even at a profit.
The final owner of the mortgage, of course, is the loser. The assumption that subprimes could be refinanced if need be failed to take into account that higher interest rates priced these people out of the market. But the worst part is this: Many hedge funds leveraged their purchase of mortgages by using them as collateral to borrow money from the banks.
That was the tipping point. When the subprime defaults started to hit, the banks that had loaned money against the mortgage portfolios re-evaluated the loans. They called some, they stopped rollovers of others and they raised interest rates. Basically, the banks started reducing the valuation of the underlying assets -- subprime mortgages -- and the internal financial positions of some hedge funds started to unravel. In some cases, the hedge funds could not repay the loans because they were unable to resell their subprime mortgages. This started causing a liquidity crisis in the global banking system, and the U.S. Federal Reserve and the European Central Bank began pumping money into the system.
Told this way, this is a story of how excess emerges in a business cycle. But it is not really a very interesting story because the business cycle always ends in excess. As economic conditions improve, more people with more money chase fewer investment opportunities. They crowd into investments that seem to guarantee vast or sure returns -- and they get hammered. The economy contracts into a recession, as it tends to do twice every decade, and then life goes on.
There currently are three possibilities. One is that the subprime crisis is an overblown event that will not even represent the culmination of a business cycle. The second is that we are about to enter a normal cyclical recession. The third, and the one that interests us, is that this crisis could result in a fundamental shift in how the U.S. or the international system works.
We need to benchmark the subprime crisis against other economic crises, and the one that most readily comes to mind is the savings and loan crisis of the 1980s. The two are not identical, but each involved careless lending practices that affected the economy while devastating individuals. But looking at it in a geopolitical sense, the S&L crisis was a nonevent. It affected nothing. Bearing in mind the difficulty of quantifying such things because of definitions, let's look for an order of magnitude comparison to see whether the subprime crisis is smaller or larger than the S&L crisis before it.
Not knowing the size of the ultimate loss after workout, we try to measure the magnitude of the problem from the size of the asset class at risk. But we work from the assumption that proved true in the S&L crisis: Financial instruments collateralized against real estate, in the long run, limit losses dramatically, although the impact on individual investors and homeowners can be devastating. We have no idea of the final workout numbers on subprime. That will depend on the final total of defaults, the ability to refinance, the ability to sell the houses and the price received. The final rectification of the subprime will be a small fraction of the total size of the pool.
Therefore, we look at the size of the at-risk pool, compared to the size of the economy as a whole, to get a sense of the order of magnitude we are dealing with. In looking at the assets involved and comparing them to the gross domestic product (GDP), the overall size of the economy, the Federal Deposit Insurance Corp. estimates that the total amount of assets involved in that crisis was $519 billion. Note that these are assets in the at-risk class, not failed loans. The size of the economy from 1986 to 1989 (the period of greatest turmoil) was between $4.5 trillion and $5.5 trillion. So the S&L crisis involved assets of between 8 percent and 10 percent of GDP. The final losses incurred amounted to about 3 percent of GDP, incurred over time.
The size of the total subprime market is estimated by Reuters to be about $500 billion. Again, this is the total asset pool, not nonperforming loans. The GDP of the United States today is about $14 trillion. That means this crisis represents about 3.5 percent of GDP, compared to between 9 percent and 10 percent of GDP in the S&L crisis. If history repeats itself -- which it won't precisely -- for the subprime crisis to equal the S&L crisis, the entire asset base would have to be written off, and that is unlikely. That would require a collapse in the private home market substantially greater than the collapse in the commercial real estate market in the 1980s -- and that was quite a terrific collapse.
Now, many arguments could be made that the estimates here are faulty or that different concepts should be used. We will concede that there are several ways of looking at this crisis. But in trying to get a handle on it strictly from a geopolitical perspective, this gives us a benchmark with which to analyze the mess.
Can it balloon into something greater? The big risk is that the weak hands in the game, the hedge funds, are suddenly coming into possession of a great number of houses that they will have to put on the market simultaneously in fire sales. That could force home prices down. At the same time, most homes are not at risk, and their owners are not hedge funds. Moreover, it is not clear whether most of the hedge funds that own subprime mortgages will be forced to try to monetize the underlying assets. It is far from clear whether the crisis will affect home prices decisively. If home prices were to collapse at the rate that commercial real estate collapsed in the 1980s, we would revisit the issue. But, unlike commercial real estate, in which price declines force more properties on the market, home real estate has the opposite tendency when prices decline -- inventory contracts. So, unless this crisis can pyramid to forced sales in excess of the subprime market, we do not see this rising to geopolitical significance.
From this, two conclusions emerge: First, this is far from being a geopolitically significant event. Second, it is not clear whether this is large enough to represent the culminating event in this business cycle. It could advance to that, but it is not there yet. We cannot preclude the possibility, though it seems more likely to be a stress point in an ongoing business cycle.
Apart from discussing the subprime issue, this crisis offers us an opportunity to explain how we view economic activity. First, we try to understand, at a fairly high level, what exactly happened, much as we would approach a war or a coup. Then we try to compare this event to other events whose outcomes we know. And, finally, we try to place it on a continuum ranging from fundamental geopolitical change to normal background noise. This is more than normal background noise, but it has not yet risen even to the level of a routine, cyclical shift in the business cycle.