Saturday, February 27, 2010

Economics 27/02/2010: How to reform our broken risk pricing system

This is an unedited version of my article in March 2010 edition of Business & Finance magazine:

There are several deeply rooted problems with the current analysis of the ongoing financial crisis. These relate to the sources of the crisis itself and to the solutions proposed for ensuring that a new financial bubble will not emerge out of the ashes of systemic risk under pricing that characterized the period of 2003-2007 around the globe.

So far, the public aspects of the regulatory responses to the crisis have been focused on ‘political’ topics, such as executive compensation. Fine: the incentive for banks executives to structure their own compensation to reflect short term gains is well established.

Political issues are non trivial as well. We all are aware of the fact that politicians – from Bill Clinton to Gordon Brown to Bertie Ahearn and on – have strong incentives to placate voters through fattened Exchequer revenue, expanded public spending and broadened access to credit irrespective of risks. Active encouragement of loose lending standards (especially in the case of the US SGEs: Fannie Mae, Freddie Mac and Ginnie Mae) were enshrined in regulatory and legislative mandates. And look no further than Greece, Portugal, Spain and Ireland as to the troubles this can cause – politicised spending breeding scores of vested interest groups that cannot be disentangled from the feeding trough.

All of these forces, underlying the crisis emergence, are well known. What is less frequently discussed in the media is that wrong incentives alone are not a sufficient condition for markets malfunctioning, since in efficient markets, a contrarian view should be able to price out those players aligned with wrong incentives.

It is a much deeper question as to whether this has happened in the case of the current financial crisis. Anecdotal evidence suggests that this was indeed so. Early in 2007-2008, a number of short positions, including those taken well in advance of the crisis, were generating the payouts consistent exactly with the rapid pricing-out of the malfunctioning lending strategies. Ironically, banning short sales has resulted in the restoration of the mis-aligned incentives in the market. An act by the regulators aimed at restoring order in the financial markets turning out to be nothing more than reinforcing the very causes of the crisis.

This means that we must look back beyond the immediate crisis to find any evidence to either support or dispute the proposition that mispricing of risks by the financial system was systemic (in the sense that existent models of risk pricing could not have allowed for contrarian pricing strategies).

It remains a puzzle that the main villains of the game, sub-prime mortgage packages (the famed ‘Collateralised Debt Obligations’), seem to have been so badly mispriced. This apparent mispricing lay not so much in the slicing of the mortgages but in the failure to price into the packages as a whole the apparent systematic risk due to the general response of property prices to the business cycle.

Suppose that the current view that greed blinded markets participants to the fact that CDOs packages were not properly pricing risks is correct. This explanation requires that not a single market participant was willing to take a contrarian strategy betting against the consensus view. Alas, this is patently untrue.

So ‘collective madness’ explanation does not hold and the crisis roots lie somewhere else – more likely, in the balance of incentives. My suggestion is that on the margin, regulatory and market incentives led to favouring of underpricing risks inherent in CDOs and MBSs. Thus, on the margin, excess returns to unpriced risk for going long on mortgage-backed products were made greater than the expected returns to shorting mortgage-backed products once the price of insuring / shorting these products was taken into account.

In other words, it was a combination of:
  • Artificially low perceived cost of long positions;
  • Artificially high cost of shorting; and
  • Recklessly elevated correlations between product risk (mortgages risk) and insurer risk (AIG)
that drove the bubble formation. The reason why this understanding is important is the following:

If contrarian strategy could have been formulated based on existent risk pricing systems, then short sellers were ‘fundamentally’ justified in their positions and their gains were not ‘speculative’. Furthermore, this would imply that the current crisis is not systemic from the financial point of view, but is driven by incentives and regulatory failures.

If, however, existent risk pricing systems were not sufficient to support the contrarian investment strategy and those short sellers who were betting against the consensus obtained speculative profits, then the markets are not efficient and the crisis is systemic in nature.

So, if we take information about the markets available pre-August 2007, could the crisis been pre-priced? Put differently, were Irish or for that matter UK or US property and credit bubbles predictable on the back of fundamentals or were they random events?

The systemic crisis argument supporters show that since Irish property prices have indeed collapsed on the back of weaker-than-expected ‘fundamentals’ market price risk discounting has failed.

Those opposing the argument point out that the fundamental in the housing market is ultimately driven by income dynamics, which in turn are driven by productivity. In the case of Ireland, productivity growth (income growth) should follow a random walk because innovation is largely unforecastable and in the case of a small open economy it is also subject to global trends. In other words, Irish productivity growth should be following a random walk that is even ‘more random’ than the productivity growth processes in the rest of the world.

Both are wrong. It turns out that close to the crisis – at least from 2003 on – Irish property prices appeared to become a non-stationary process having been largely stationary in the previous decade. Ditto for the US and UK, and even Spanish, Russian and Dutch property prices.

This means that the conditional forecast of the property prices in Ireland was best modelled by a reference to the current prices. More importantly, from the point of view of risk pricing, expected conditional volatility of house prices was a scale factor of the observed current volatility. In other words, the degree of risk 6 quarters from, say February 2003 was simply 6 times greater than very low volatility observed back in February 2003.

The scaling relationship, alas, failed to hold in the real world. As the property boom became, using Bertie Ahern’s terminology, ‘boomier’ through 2003-2006, property prices stopped following non-stationary process and their volatility became largely trend driven. The trend presence means that at least in part, future risk could have been priced into lending decisions by the banks and regulators. Alas, it was not. All evidence on lending suggests that the banks lending margins were heading down during 2003-2008 period, not up. In other words, Irish mortgage lenders, with tacit consent of the regulators, were pricing in decreasing risks into the future during the bubble inflation time. Ditto for all other countries that have experienced the collapse of property markets.

So whilst the financial markets were correct in pricing risks, subject to significant regulatory incentives constraints that skewed their willingness (and later ability) to actually adjust their risk pricing positions, mortgage lenders and regulators were grossly mispricing risks. This realisation leads to two major conclusions.

It shows that globally, financial crisis of 2007-2010 has been driven by the risk mis-pricing that originates in the very institutions whose business is preventing this from happening – the banks and the regulatory bodies.

And it shows where the future reforms attempting to address the issue of financial bubbles formation must lie. And it certainly has nothing to do with bankers compensation packages. The main solution to the problem suggested today is heavy re-regulation of bank risk; moving Basel I and II up to Basel III to include a pro-cyclical risk capital provision.

While a useful idea, greater buffer reserves of risk capital built over the years of credit expansion cycle, are not a panacea to the problem outlined above. The reserves are only sufficient in so far as they reflect actual risk expectations. Missing risk forecast will, in the end, still imply sub-optimally low levels of capital.

Instead, the answer to the problem of how can we prevent future bubble formation similar to the one that has been deflating since August 2007 lies in a more holistic approach to risk pricing reforms. This approach must involve several policy changes along the following directions.

Firstly, a more transparent early warning system must be deployed across the financial markets that would make short trading positions a part of open market pricing mechanisms. Put more simply, short trading must be allowed to operate on unrestricted basis, but all short trading positions must be disclosed and reported in the market in the same way in which we current disclose long positions.

Secondly, property must be treated as investment instrument, with full price and hedonic information disclosure rules mirroring those required for liquid financial instruments under MiFID.

Thirdly, there must be a clear set of strict ‘no pain, no bail out’ rules that will impose severe penalties on the management, bondholders and shareholders in financial institutions seeking public assistance. If countries can change governments within weeks after elections, banks can be weeded of their failed management as a matter of months. Instead of restricting their pay, in the future, we must make bankers accountable for their failures.

Third, regulatory authorities must be beefed up with independent, fully protected risk analysis boards drawn across the broader economy. These boards must be politically unconstrained, free of interest groups influence and must be operated behind a strict Chinese Wall relative to the entire regulatory process. A formal requirement must be imposed that at least 1/3 of the board members should be drawn from outside financial services sector, with the same proportion of members being required to hold a publicly verifiable policy positions that are contrarian to the consensus.

What the current crisis has taught us is that in the environment where politicians and industry drive risk pricing-related policies, failures of the market to cope with distorted incentives and incomplete regulatory oversight will be spectacular. Crises are a natural way for the markets to reassert proper order on inept regulatory and institutional systems.


artied said...

Great analysis, Constantin. But I think fundamentally your prescription implies a change to fundamentals of human nature.

The slow steady growth of the bubble meant that all sectors of society and the administration fell victim to 'group-think' and things that were rash before became the norm (viz 35yr int only mortgages on 7 times income).

There is no prophylactic for 'group-think'.

TrueEconomics said...

That is true - group think is a powerful force driving everything to a consensus/median forecast.

This is why it is very important to allow contrarian risks payoffs to take place (short selling is one example) and also to impose direct pricing on decisions of those who are happily rolling with the crowd (no bailouts clause and automatic destruction of management, equity, debt holders in this specific priority).

It is equally important to make sure that the voices of those who disagree with the group-think are being heard. In Ireland this requires a statutory limit on cronyism.

What the latest experience has shown is that cronyism is still alive and well. Since the beginning of the crisis, 90% of all appointments made to various new regulatory structures and supervisory / investigative / policy bodies made by the current Government - including the Greens - are made from the same old-boys networks of people who have neither intellectual, nor professional experience capacity and certainly no track record to think and act independently.

We need a statutory requirement for inclusion of 'uncomfortable' voices into the policy making process here. Otherwise, there is no hope.

I know of several people who have done more work on Nama than the entire Nama board combined. None of them are allowed by the State to contribute to the working of Nama. Ditto for consumer protection, financial regulation, financial risk assessment, tax policy, and so on.

Group-think in this country is alive and well, and it must be broken, completely demolished, if we are to have any hope of getting out of this mess.

artied said...

Completely in agreement with all of that comment.

James said...

"Secondly, property must be treated as investment instrument, with full price and hedonic information disclosure rules mirroring those required for liquid financial instruments under MiFID."

With all due respect, this proposition is problematic because of how property is traded and the heterogeneous nature of property assets. While adjustments can be made through the hedonic, those observations are always backward looking and generally lagged for a significant period....and don't even start me on the valuation process....

Unless the fundamental process of trading property is altered, it is almost impossible to have information provided in a timely fashion that is also useful. The best alternative is to have most countries disclose accurate and timely mortgage data, including application data for comparison. The HMDA data in the US, if properly disclosed and analysed, could have provided a better indication of the impending disaster than property price data.

TrueEconomics said...

James, I am not suggesting that we have daily quoting on property assets that are of the market. What I am saying is that property transactions must be disclosed - sale price, hedonic information at the time of sale. And that this information must be made public. That real estate agents must treat their claims relating to investment quality of a property or to the general property market as investment advice.

Give you one example. In an investment prospectus on a 5-star hotel (that is heading for Nama now), Hooke & MacDonald - the sole agent for the sales of hotel units - makes a statement that they see no reason as to why no market will exist for this type of investment in the future. The statement is made in 2006. It is unbacked by any research and contains no estimates or data.

There is another claim that in the view of investment management firm, the yield on investment can be expected at x%.

The rest of prospectus contains estimates of financials, etc, but not when it comes to the actual property valuation or to the yield.

In other words, the part that makes property an investment in this case is the same part that contains no real information.

This has to end. If property is an investment - agents must provide information to the public. If it is not, they should stop making statements relating to returns, investment and savings.