March 20, 2008
Operational Risks from External Events
Amidst all the carnage from the fall out of the sub prime crisis, operational risk has come back to occupy centre stage. For this, the credit goes to Jerome Kerviel, the Societe Generale trader whose allegedly unauthorised trades cost the bank US$7bn. Clearly, there was a breach in the internal process which allowed the rogue trader to build large positions in the futures markets.
Another ongoing (but not so well known) example of operational risk pertains to OTC currency contracts between banks and software exporters in India. The mid sized exporters whose earnings are primarily in US$, have faced the brunt of a depreciating dollar (and an appreciating rupee). This, coupled with higher operating costs, put pressure on the bottom line and naturally enough they were in search of some quick fixes. The villain in this piece viz. the Indian banks, apparently were only too eager to sell them currency derivatives. Corporates booked upfront the premium received, while in other cases boosted reported profitability through back to back structures. However, following a sudden and steep decline in the value of the US$ against the Japanese yen and the Swiss franc many corporates are staring at losses in their positions and are contemplating suing the banks on grounds of mis-selling complex structures. (Long term observers will note the similarity between this example and Procter & Gamble which successfully sued Bankers Trust in the mid nineties on similar grounds.)
The issue that that I intend to discuss here is not the reputational risks for Indian banks, but the more broader issue of definition of operational risk - defined as the risk of loss resulting from inadequate or failed internal processes, people and systems or from external events.
In a Basel II context, the goal of operational risk management is to focus on internal events (people process, technology) since they are only within the control of the firm. For example, a risk control self assessment process will help document operational processes, identify key operational risks and evaluate controls from an internal perspective.
What kind of external events pose operational risk to the firm? Political, tax, regulatory events can indeed lead to operational risk. To this list, one more factor can be added: risks of selling to uninformed or even illiterate clients. I deliberately choose to categorize this as an external event because the risks that clients are exposed to (arising from the advice given by the bank) is ultimately transformed into operational (legal) risk for the originating bank. Note that the asymmetry of risks in this case: if the client had made money on OTC contracts, the bank and the client would be mutually patting each other on the back and not engaged in a legal duel.
It is natural enough to evaluate the preparedness of banks in selling risky products. However one should not forget to evaluate whether clients are sophisticated or naive when it comes to buying these risky products. Do they have policies, tools, and methodologies to identify, measure and control risks? Do they have board approval? Are they hedging or speculating? In the Bankers Trust case, I have always wondered how P&G could claim that they were not fully informed of the risks in speculative derivative positions especially when their treasury was considered cutting edge. Surely, they would have also assessed downside risks? It seems a reasonable assumption, but no there is information on this aspect.
In the case of the Indian corporates clearly, they may not have cutting edge treasuries or the sophistication to understand risks in exotic instruments. (In fact a recent risk management survey conducted by a leading financial daily highlighted that most Indian companies are still quite far from having good risk management processes). Nevertheless, it remains a fact that they willingly entered into contracts on the hope of making money through speculation. It would be interesting to see whether the Indian corporates can follow the same path as P&G and recover not just their losses but end up making gains from suing banks! Interestingly, banks seem to have anticipated the possibilities of law suits. Reportedly, voice recordings with clients is likely to be a key input to demonstrate the banks had indeed clearly spelt risks to corporates before hand.
The above example only demonstrates the pervasive nature of operational risks for banks - they can be quite diverse and plentiful. Given this pervasive nature, is it possible to capture all key risk events? Finally, will the operational risk charge under Basel II fully capture all the risks faced by the bank?
Posted by aaaaaaa at 07:33 PM
| Comments (1)
March 09, 2008
Credit Risk / Investment Risk- two sides of the same coin?
Recently, I was talking to a senior risk manager in one of the offshore banks based in this region. This particular bank had private equity exposures, which are required to be classified in the banking book as per local regulations. Fair enough.
But what followed was confusing. Since private equity which has investment or price risk is classified alongside credit exposures, and the same advanced model (PD/LGD) is used to determine capital charges, the bank equated investment risk with credit risk and treated them as equivalent for practical purposes! While this was not what Basel II set out to achieve, nevertheless there are some areas with respect to equity exposures in the banking book which needs to be debated.
Risk weights for credit exposures vary from 0% to 150%. Capital is held is to protect against both migration risk and default risk. On the other hand, equity exposures in the banking book are weighted at 300% (publicly traded) or 400% (other equity holdings). The higher risk weights for equity exposures reflect the higher levels of unexpected losses as well as the greater variability of returns in equity exposures. On a portfolio basis, credit risks may be granular i.e. distributed across large number of clients, while equity exposures may be lumpy in nature. In addition, the additional risk charge (100%) for non traded equities reflects the liquidity risk in fire-sale situations. These three factors explains why price risk in equity investments is not the same as credit risk even if they belong to the same banking book.
The confusion is understandable if one looks at models based approach for estimating investment risk for equity exposures. Normally, the PD/LGD model is associated with the credit risk estimation; pray how is this relevant for estimating the investment risk of equity exposures?
In fact the initial consultative papers by Basel acknowledge that default risk is itself a difficult concept to define for equity. It goes on to clarify that the PD/LGD approach, incorporates elements of both general market and idiosyncratic (i.e. specific) risk associated with equity holdings. This unfortunately, is not readily apparent or explained.
If banks decide to use the PD/LGD model for measuring credit risk (oops! Sorry, investment risk!), there are further issues. Banks are required to estimate the PD of a corporate equity exposure in the same way that it would do for a debt exposure for that company.
What happens when the bank does not have a debt exposure for which it has an equity exposure? In such cases, it is difficult for banks to have access to information. Basel recognises that in such cases, “the equity definition of default is likely, on average, to deliver a later outcome than the corporate definition”. In recognition of this lagged/delayed effect, Basel has recommended a 1.5 scaling factor be applied to the PD/LGD weights in such circumstances.
Given these difficulties, most banks would prefer to stick to the simple risk weights method, rather than venture into the models based approach. While this means additional capital charges, it is not a problem for many Gulf based banks, as they have much higher capital adequacy levels than required by their regulators. That is, as long as credit risk and investment risk is not equated and are treated differently!
Posted by aaaaaaa at 11:13 AM
| Comments (0)
