RE: IBG YBG WBG

From: Tjemme
Affiliation:
Address: tjemmevdm@hotmail.com
Date: 15 Dec 2007
Time: 12:14:54

Comments

Allow me to give a few thoughts on buy-side risk management: (1) Traditional "accounting" does not provide sufficient "accountability" (www.bis.org/publ/work213.pdf). Ideally, every transaction is linked not just to a valuation model, but also to a risk model. This risk model is purely based on the past (and on exogenous credit ratings). It should one day be driven by standards, run by professionals, much like the accounting profession currently works. Ideally, however, we can tailor the amount of transparancy / accountability to an externally validated statement about the complexity of the underlying investments. This is because accountability has clear costs and benefits will differ depending on the process and stakeholders. We should also provide standard statements about the value of the portfolio for which no reliable risk models are available. I think this is where the benefit of standard risk models like Barra / Riskmetrics comes in, and where the benefit of outsourced risk measurement (prime broker, custodian, ..) comes in. The role of subjectivity is limited. This kind of risk measurement is driven by clients / regulators and the like. They should not expect perfection here. Accounting and risk standards always bound to be behind new developments. And, yes, we should continually improve this. (2) Then there is risk control. Buy-side risk control is vastly different from sell-side risk control. That is because the client (the fund board) is the owner of the parameters. Compliance ensures that the sell-side does what is sold and sells what is done. It ensures that regulatory requirements (providing transparancy about risks) is met, most likely in a rather simple manner. Limit monitoring is a compliance responsibility. (3) Then there is risk management. Yes, risk managers can support decisions. And, yes, internal models can be useful for this. But these internal models are ad-hoc, proprietary and subjective. They are tools, they need not be relied upon. The difference is like the difference between accounting and management information. Management information (ex-ante decision support and ex-post performance attribution) needs to be continually linked to the internal organisation and decision procedures and needs to change with it. It is an art to decide what elements of the complex world to spend time on modelling. Be glad that someone spends time on it. This is better than people taking ad-hoc decisions. We don't want internal controls around these things. Of course, when management information is routine, it may be worthwhile to document it and surround it with controls. But the business typically changes so rapidly that this is not worth it. (4) Counterveiling power. Within Banks, risk managers are often positioned as counterveiling power. Much like the court has two sides. As we know from the US, this is an extremely expensive resolution of possible conflicts, but provides an extremely valuable 'second opinion'. I think in many cases a standard 'risk - accountability report' is all that is needed. We only want American type legal situations for more complex / new transactions. Don't forget, we are not talking about financial risks, but we are talking about the myriad other kinds of risk as well. Should we have 'competition risk management', to monitor that the business has the right perspective on the competition? Running a business is taking risks. The business is the risk manager. A good business man knows when to ask for a second opinion ... but it seems too much to have too many procedures and controls around that.

Disclaimer