About Us | Search | FAQ | Contact Us
0-New demands on buy-side risk reporting
In future there will be an increased reliance on these 'forward looking' risk measures. This will be true both within asset management firms and, perhaps more interestingly, for their investors. The connectivity opportunities created by the Internet coupled with investors' growing demand for greater risk transparency both enables and provides the impetus for this form of external reporting to happen. Our focus here is primarily on the issues raised by an increased reporting of risk by asset managers to external parties.
At the same time institutional investors are demanding greater transparency and risk reporting from the external managers of their assets. An asset manager's ability to demonstrate superior risk process and reporting has become a differentiating factor in the management selection process. Investors are effectively demanding that their money managers make the investment in forward-looking risk systems.
There are important differences between measuring risk for internal management needs and reporting externally to investors. Internal measurement and reporting concerns the aggregation of risk across disparate business lines and systems. External reporting concerns the distribution of this information and ultimately the aggregation of risk across disparate firms. In the case of the investment management industry, the challenge is to aggregate risk from multiple asset managers and disparate investment strategies.
The (Near) Future
of Old and New
Observed volatility in the time series of Net Asset Value based reporting - the traditional approach to risk reporting within the investment management community, e.g., Sharpe Ratios (Fig. 1);
Potential loss of value due to the volatility and correlation of the constituent securities in the portfolio - the approach to risk reporting favoured by the sell-side, e.g., Value at Risk (Fig. 2).
There has been a significant amount of blood, sweat and tears invested in extracting information content out of a simple time series of NAV's. Examples include Sharpe Ratios, Information Ratios and alphas. Moreover, style analysis, performance attribution and historical risk/return analysis can all be derived from this type of information. These approaches are relatively simple to produce, are objective and support portfolio analysis. However, these measures suffer from being backward looking and most importantly they do not directly answer how much of the return is explained by:
n Risk taking.
There are countless examples of portfolios that earned spectacular returns with low volatility in NAV and then blew up in stressed conditions.
Value-at-Risk and scenario analysis are both examples of risk measurement approaches which can help differentiate between luck and inherent risk taking. Deflating returns to reflect the potential loss in the portfolio, as contrasted with the historical volatility in the fund's performance, has a number of advantages:
n The reported risk is insensitive to the managers' past performance
n A change in the risk profile is instantly observed (e.g., increase in the VaR)
A problem with these types of measures is that they require more information from the managers about the current composition of the portfolio and thereby how sensitive the portfolio is to changes in rates and prices. Moreover, there are many ways to describe these sensitivities, which in the absence of a standard makes it difficult to aggregate risk across portfolios. However, as we observed above, the ability to populate these models is improving as money managers invest in sell-side type risk analytics.
Internal Risk Reporting
n Reporting of risk will become more frequent
n Risk measurement methodologies will increase in sophistication
n Standalone reporting by each business entity will evolve into integrated portfolio risk measurement for a combination of business entities (notably investors' portfolios of investments). This aggregationacross funds will require a degree of standardisation in the reporting between the various asset managers.
These three elements are often in conflict with each other. Greater frequency in reporting may necessitate compromising the precision of the risk methodology employed. Similarly, portfolio reporting is constrained by the lowest denominator of the various reporting entities, in terms of both the frequency of reporting and the relative sophistication of the methodologies employed.
The challenge of arriving at a consistent and coherent portfolio risk measure is endemic to the majority of issues facing all risk managers, both sell-side and buy-side. The conceptual break through which enabled portfolio risk to be derived statistically (e.g., VaR) has proven to be a two-edge-sword. The 'devil in the detail' in implementing risk measurement across systems, business types and time zones is more difficult than what is implied by simply considering the math.
A typical response to the challenge of portfolio risk measurement within major financial institutions has been to commit huge technology budgets, often in the order of tens of millions of dollars. One popular approach has been to build a global warehouse and put a comprehensive risk engine on top, which embodies analytics as sophisticated as those found on most trading desks.
This approach while the most intellectually pure has often resulted in disappointment:
n Outright failures have been quietly brushed under the carpet,
n Unstable operating environments have resulted in frustration and challenges to the integrity of the metrics and
n Computational intensity has compromised timeliness, and resulted in numbers arriving too late to impact business decision making.
It was soon realised that throwing money at the problem (and the associated race to complexity) did not necessarily guarantee success or improve decision making. More enlightened risk managers opted for less ambitious solutions which gladly traded off complexity and accuracy for timeliness, breadth and cost savings.
n From a cost and operations perspective the win-win is to leverage the systems which have been developed for internal risk monitoring. Thankfully, the information requirements of external parties are similar and are generally a simplification of that required by internal decision-makers.
n Enabling external parties to consolidate their institution's risk information with other institutions' data will require data and methodology standards. Standards have the benefit of enabling comparison and aggregation but often at the expense of relevance and accuracy. Sell-side institutions have been wrestling with these issues for sometime, as part of developing enter prise wide risk systems and in dealings with regulators. Standards of the type being discussed here have yet to be promulgated. However, as people come to realise the bene-fits we can expect greater standardisation,either arising from a successful commercial venture or through the intervention of a third party (e.g., AIMR).
The implications of 'straight through connectivity' on governance arrangements which empowers third parties needs to be understood. Further clarification of roles and responsibilities will occur, as it did on the sell-side once enterprise-risk systems got up and running. Resulting structures and processes will need to reflect this greater transparency but protect the delegation of authority ñ we all want to avoid decision paralysis from having too many "back seat" managers.
been made to en
|Home | About Us | Search | FAQ | Contact Us|