Models and Overconfidence

Whether a scientist, or a financial engineer, when we invest in an idea, a model, a way of thinking, by definition, we believe it to be an accurate representation of the world, and so interpret new information in terms of this model. New information that is consistent with this model makes us more confident in its use, while new conflicting information is often regarded as spurious, exceptional or irrelevant. The model provides us with an anchor, and biases us away from alternatives.  We protect our models from the world of evidence, in the same way we protect ourselves, no surprise here, as in some sense these models constitute ourselves. The models change only with crisis, some overwhelming surge of evidence that cannot be ignored. Ironically, financial models, (unlike models of physics) when broadly and consistently acted upon, will often build up arbitrage and economic pressures in the market to break the models. So no one can see the breakdown coming, everyone is shocked when it does happen, and yet everyone can rationalize after the fact why the crisis occured.
Perhaps the truth is that reality is more complex than we can ever know, and like a good buddhist, all we can do is acknowledge and respect the complexity that we face, realizing like the quantum physicist, that the act of understanding the world, also changes it into something else.

The Curse of Overconfidence and some revealing party tricks...

Irving Fisher, probably the greatest american economist of his generation (most known for his theory of interest rates), remarked famously just before the great crash of 1929 that "Stock prices have reached what looks like a permanently high plateau". To be fair, he put his money where he mouth was, and invested much of his assets in the stock market. Soon after he was broke, and had to sell his home in NewHaven Connecticut to pay his debts (fortunately his university bought his house and let him live there, or he would have been destitute!). Even after his experience, he continued to argue that he had been right.
The curse of overconfidence is not unique to economists, indeed one of the major trends in new economic thought is behavioral finance, applying all of social science theory to finance, not just economics. And one the cornerstones of this theory is that people are overconfident, inconsistent, lazy, and simply stupid at times. In short, not the "homo economicus" of classical theory. Consider one's own overconfidence - ask yourself or your friends to define a 90% [lower,upper] range for answers to the following questions (no cheating!), and find out if you indeed tend to get 9 out of 10 them right.

1. What is the population of Turkey?
2. What is the weight of the Empire State Building? (tons)
3. What is the GDP of Australia?
4. How many cells in the human brain?
5. How likely is it that you will be struck by lightening this year?
6. How many bibles are there in the world?
7. How many children will die of malnourishment in the world in the next day?
8. How likely is it that you will be in a plane crash this year?
9. What is the likelihood of a asteroid hitting the earth in the next year and destroying all life on the planet?
10. What is the likelihood of a AAA bond defaulting in the next year?

 Answers (no peeking)

1. 72,561,312 as of 1st January 2010
2. 365,000 tons
3. $1 trillion USD
4. 100 billion
5. 1/750,000
6.  6,000,000,000
7. 3000
8. 1/675,638
9. 0.00000002
10. 0.1

Most people (assuming they play fair) get significantly fewer than 9 out of 10 of these questions right. They believe the world is less complex and better understood than it actually is, and this influences their behaviour. Unfortunately sometimes with disasterous results - i am reminded of the quote from Larry Kersten, an american sociologist that "before you attempt to beat the odds, be sure you could survive the odds beating you.”

 

 

Moves to increase transparency in the OTC market

It’s clear that the juggernaut of transparency will be hard, if not impossible to resist. And this is generally a good thing for most derivatives users, even in sophisticated markets like the US. For far too long, OTC derivatives have hidden risk and complexity from users and regulators under the guise of “customization”, providing much value to product innovators but little real value to end users. Eventually, the essentially one-sided nature of much of this technology will be revealed. However, the real risk is that the pendulum will swing too far in the opposite direction, preventing even sophisticated end users laying off their risks with bilateral contracts. Part of the limitations of such a one-sided response is the assumption that transparent markets are a cure-all. They are not. Transparency is about having a information baseline against which all transactors in the market can make their decisions. However derivatives information does not necessarily equate to knowledge or expertise in derivatives product use. All the information in the world does not create an adequate valuation or hedging model, and even less does it create an understanding of the limitations of the valuation models used.
So Regulation should increase transparency of transactions, but it should also target the expertise of potential counterparties. Some institutions (much like non qualified investors) should not be transacting with counterparties if the expertise imbalance is simply too great. And frankly given the huge investments financial services have made in financial engineering human capital, this may often be the case.

The Rise of Collateral and the Urgent Need for Collateral Risk Management

The downturn has clearly pushed Collateralization to the fore of OTC derivatives, repos and securities financing. What was often viewed as a luxury, is now seen as a necessity as counterparty risk is not confined to low rated entities (sic. Lehmans!). But the rise of collateralization does not destroy the problem faced by counterparties. Collateral may increase some business opportunities but it also converts counterparty credit risk into operational and legal risks and into residual market and residual credit risk. This new bundle of risks is certainly different from what was held before and arguably is more complex. Consider Operational Risk is implied in everything from reconciliation to posting collateral with a custodian, from collateral valuation to settlement. Legal Risk is endemic to the OTC derivatives space and is of course the rationale for the various master ageements (ISDA, GMRA etc) and associated credit annexes, many of which remain to be tested in emerging markets legal jurisdictions. Residual Risk always remains, whether it be from market risk - the changing value of the collateral, or the changing risk sensitive margins, or even reinvestment risk; or from credit risk - wrong way risk - the possibility of correlations between counterparty default and collateral values. In short, collateralization is a good thing, one to encourage, provided one realizes that risk is not removed, merely converted into another form. And these new risks must be managed if collateral is not to give a false sense of security in these turbulent times.

The coming Pension Fund Crisis and how it will change the world

The next big thing, after the sovereign debt crisis that is slowly percolating through the financial system, is the pension fund crisis. Basically the problem is that public and private pension funds across the world have long managed their balance sheet as two independent elements - assets and liabilities and the concept of ALM has never been accepted or acknowledged within the industry. Liabilities - the pensions themselves, very much managed as an actuarial portfolio, usually based on the demographics of a particular fund and its covered employees is best thought of as a set of zero coupon bonds with limited optionality, the duration of which reflects the average life expectancy of the average insured individual. This of course varies by pension plan, and no surprise therefore that the asset side, the funding of these liabilities must depend on a detailed analysis of the cash flows associated with these liabilities. In short we need to build a replicating portfolio for the liabilities (just like in Bank ALM) and use that to define the performance expectations of the asset portfolio. This is rarely done - instead pension funds use fund of funds to "divide and conquer" the asset portfolio and evaluate that performance on existing standard market benchmarks for bonds, equities and other assets. But of course, performance for a pension fund means the ability to fund the liabilities, it does not mean the ability to outperform a bond index like the Barclays Capital Aggregate Bond Index - which is typically industry practice. Analysing risk for a pension fund means analysing the risks to achieving performance, in other words, how likely is it that our asset portfolio will underperform the liability portfolio. With a liability replicating portfolio, we can build var and shortfall models that allow us to monitor and manage the net asset position over time. The lack of this capability is not merely a problem for a few pension funds. The mismatch of mark to market values of asset and liabilities in corporate pension funds is of the order of trillions (yes - trillions) of dollars in the US alone (UK, Russia, Italy also big problems). If there is continued fall in the value of equities this is likely to get worse, and be a drain on corporate performance for years to come.

What does it mean to hold a diversified portfolio?

Imagine we have a portfolio of equal positions in different securities? Is this diversified? Probably not. What if these securities were all bonds with small volatilities, and one equity position with a much higher volatility - that clearly would not count as properly diversified. So individual security variances need to be considered when thinking of diversification. What about correlation? This too should be included. The best measure of diversification probably considers the contribution of the pieces to the overall variance of the portfolio. But a complication here is that variance is not additive - so instead we often talk of factor variances, calculated based on some form of Principal Component Analysis (PCA) - basically computing the eigenvectors of the original covariance matrix in order to produce orthogonal risk vectors (whose variances are then additive). Having PCA risk factor variances then allows us to estimate the cumulative contribution of different securities to aggregate variance starting with the greatest factor contribution. Much like a Pareto Analysis, we ask how different is the actual contribution to perfectly diversified case - a straight line, and this gives us a robust measure of diversification.

Socialism and Risk Management

So what does socialism have to do with risk management? Everything! Socialism, whether it is in the form of a kibuttz, or Robert Owen's New harmony, or Karl Marx's Communism, is ultimately about the provision of a safety net for all workers. High but unequal and uncertain levels of income are sacrificed for equality. Or to put more poetically in the words of Marx, from each according to his ability, to each according to his need. It is a form of risk reduction based on pooling of risks and of resources. Much like social security or insurance, it works because individual losses (and gains) are shared across a wider population. And of course losses are usually feared more than gains are valued, so such insurance is deemed to have value. The problem with such socialism (and for that matter, communism and insurance) is moral hazard. The old Soviet joke "we pretend to work and they pretend to pay us" captures it all. Moral Hazard is endemic to risk sharing particularly when the potential outcomes are partly under the control of the individuals themselves. The fire insurance that encourages arson, the seatbelts that encourage fast driving, the federal deposit insurance that discourages due dilligence in banks. Most of political philosophy is actually an extension of this idea, the production and the distribution of resources in an uncertain world. Unfortunately our positions on such political ideas are usually coloured by our current states, and the inherent moral hazard associated with that state as we try to game the system for our personal benefit. Philosophers like John Stuart Mill, Rawlings have said much the same from the liberal tradition.Howver even some conservative thinkers, particularly communitarians, have realized that we are part of a society and our survival as individuals depends on our building networks that share opportunities and risks within that society.