Market Risk Measurement
Market risks can and should be measured continuously to avoid unexpected losts.
When properly understood, risk controls can be integrated to the portfolio construction process which leads to superior Return / Risk ratios.
What's New:
More Educational Content coming soon
More educational content
to come. Sujects include double taxation of bonds and corporate governance.
April 19, 2012: Death of a great investment counselor, Carl H. Otto
Updated: April 20, 2012
Risk & Portfolio Construction History
In the 1950to70s, investment managers were using ad-hoc risk controls & portfolio construction processes. Some would insure that they had some industry diversification. Some looked at a series of financial ratios and aggregated them for the portfolio. The more sophisticated investors would compare aggregated industry exposure and financial ratio measures to same measures from some index benchmark. Managers had an idea of the portfolio characteristics that was preferable vis-à-vis the benchmark.
In 1975, BARRA introduced the first multi-factor risk model which allowed the estimation of ex-ante measure of risk for any portfolios. Managers could test the impact of portfolio changes on the expected risk before implementing those changes.
Once we have expected return forecast and a risk model, we can find the best portfolio that will maximize the expected return of the portfolio for a certain risk budget. This required fast large-scale optimization algorithm and fast computer. Such optimal portfolio construction tools were introduced in the 1990s and are today of common use by most investment managers who understand risk budgeting at the portfolio level.
In the 2000s, we saw the emergence of daily calibrated risk models. While daily risk models adapt faster to changes in market risk, they bring some serious issues when dealing with less liquid stocks. They also bring some serious issues of statistical robustness when estimating daily factor returns. While at lower frequencies, outlying observations were rarely an issue, the use of higher frequencies bring along some very serious issues of extreme outlying observations at the stock return level and more importantly at the factor level – something which is highly damaging to the risk model. In fact, improperly built daily risk models may have contributed to the huge sell pressure & liquidity crisis in 2007 and 2008.
AT the end of the 2000s, beginning of 2010s, we see increasing demand by sophisticated investment managers, and offer by risk model companies, of alpha-linked risk models, something that can only improve portfolio construction.
Dominic Clermont, ASA, MBA, CFA
