Risk Reporting – Here be Dragons

As an ex-physicist, I will always hold true to the, very subjective, opinion, that much of financial theory is just a case of physics wannabees. The fact that two psychologists, Kahnemnn and Tversky won (well at least one of them) a deservedly Nobel prize in economics, merely confirms my prejudice.
That being said, this note is not about such high-fulutin’ stuff but about the real information of risk management – risk reporting and the lingering influence of trader wannabees. In this sphere I think that we in the current risk world are still focused on information that is relevant to fund managers, or more realistically, the investment bank traders, from whom the fund managers acquired their risk management reporting style. What I would like to explore, in this note, is how independent risk manager risk reporting appears to have become a captive of fund manager reporting (which is so much easier to do, just replicate what already exists).
Fund managers need a vast amount of information to manage their portfolios successfully. I hope I have given you my opinion that the successful fund manager has a remarkably difficult job. They need a large amount of information to ensure that their current portfolio is consistent with their current views, with the impact of stresses on the portfolio and any other information that they think is necessary to do a professional job. The end result is a risk report that has a vast amount of information. It is not as if they look at it all every day, but that information has to be trivially easy to access as and when they need it. This leads to remarkably long risk reports for fund managers. This is not illogical, it is a requirement of the role.
However, such a report is not a requirement of the role of the risk manager or anyone else in the food chain of asset management. What is required, obviously, is that any given restriction is not being breached. This can easily be accomplished with any comprehensive risk management system. What is more important is the ability to monitor the risk profile of the fund. If the leverage of the fund has doubled in a month, still well within limits obviously, that is a risk management issue. If the liquidity of the fund has dropped from 90% in one day to 90% in 5 days, still no limits being breached, that is a risk management issue. If both happen at the same time, now you have a serious risk management issue. Might be nothing, but over and above any compliance issues, which are normally green, amber or red, this requires consideration. To address these issues, you need a comprehensive record of the actual risk stance of the fund over an extended period. Is this current risk profile normal, or is this unusual? Without historical fund data you are lost. The ability to monitor and demonstrate the fact that you are monitoring the total risk profile of the fund, over an extended period, is the key to properly risk managing portfolios.
The risk report that we have inherited from the long-gone traders is the standard issue. It is remarkably long and looks only at the current risk. It should be. Fund managers have multiple on-going investment and risk management decisions to make and they need to aggregate a large amount of information to make those decisions. This is almost the polar opposite of the risk managers requirement, where you need to define a minimum number of, hopefully orthogonal risk metrics, that allow you to monitor effectively the risk profile of the fund. As a simple point, a standard risk management report normally gives, down to individual asset level, losses under multiple stressed scenarios. Lots and lots of pages. Yes, we get it, the portfolio gets whacked when bad things happen, it always does. What the risk manager cares about is what you have done to increase or decrease the whack at the portfolio level. As a risk manager I am greatly unconcerned with the composition of your portfolio, unless you start changing your risk profile. Then I get interested, at which point I may need individual exposure data, but not as a general point.
As an example, I would offer liquidity risk measures. Whilst I like to believe that, at RiskSystem, we have an objective liquidity analysis process, I am not so confident to believe that our liquidity risk estimates are, on a single day’s analysis, a completely valid description of the actual liquidity of a fund. There is data, heuristics and assumptions that allow us to estimate a funds liquidity. None of them perfect. What I care about is the evolution of our analysis over time. Our assumptions do not change over time, thus any change in liquidity is due to changes in the portfolio, changes in the market, or a combination of both. But it requires that observation of change. Without a comprehensive database of the risk management profile of a fund, observable through a wide range and sometimes varying set of risk metrics, a risk manager cannot properly monitor and thus cannot challenge a manager on their management of the fund, and there be dragons.
To conclude, risk reports are never one size fits all. Every fund manager will have their own preferences. The risk manager will have a different set for each fund under their purview. The CRO will have another set and the board yet another. The data becomes more integrated, and more evolutionary as you move up the food chain, and a proper risk management system will have the capacity to provide for all of the requirements.