One of the first measures of risk on Wall Street was a simple allocation of investment dollars: “How much do I have invested in Europe?” or “How much is in US Bonds?” Since the introduction of derivative instruments with inherent leverage, this concept has expanded to be a measurement of the full economic exposure of an investment. Instead of asking how much is invested, people are asking how much is controlled by the investment. While it may only take a few thousand dollars to open a margin account, that investment may control hundreds of thousands of dollars through the futures market.

Exposure is the measure of the full amount of money under control by an investment and is one of the inputs into a standard Value-at-Risk calculation.

Measures of volatility are nothing new to portfolio management. J.P. Morgan’s famous reply to the question about what the markets will do rings as true today as it did in his day.

Because it’s so fundamental to risk management, Investor Analytics measures the volatility of each security in a client’s portfolio every day using a variety of techniques. Typically, we use a rolling 100-day window and exponentially weight the historical returns to put more emphasis on recent information. But there are circumstances where it’s more appropriate to use a longer window or a flat weighting. Our Client Services team works closely with each client to implement the most appropriate measures for their portfolio.

Volatility measures how much a security’s price fluctuates around its own return, and is a second input into a standard Value-at-Risk calculation.

While volatility tells us how a single security fluctuates, the correlation tells us how two securities fluctuate compared to each other. Do they always move exactly together (correlation of 1), do they always move exactly opposite to one another (correlation of -1), or do they behave in between these two extremes?

Like volatility, correlations between instruments change constantly, so IA calculates the correlations between each of your securities every day. With this level of granularity, we are able to determine the correlations between any groups of securities within your portfolio: one strategy compared to another or one manager compared with another. We also calculate the correlation of any part of your portfolio to any number of market factors or benchmarks of your choice. IA provides an easy way to answer the question, “How correlated is my portfolio to the S&P?”

Correlations based on historical portfolio performance (say from the last 36 monthly returns) are influenced by trading decisions made long ago, and are good for demonstrating how the portfolio was correlated to market indices. But we take a different approach. Because we calculate portfolio-level correlations from the securities you currently have, we provide a more forward-looking correlation measurement. Of course, we can also provide a measurement based on historic returns for your comparison.

Correlations are the third input into the standard VaR calculation.


Value-at-Risk has become the industry’s standard approach to quantifying market risks. It provides a basis for measuring the expected volatility of a portfolio of securities under normal circumstances and is in wide use among banks, portfolio managers, pension funds, endowments and other money managers.

VaR is a statistical measure that attempts to answer the question “If my portfolio loses money, what amount of money is it likely to lose?” over a specific amount of time within a confidence interval. For instance, a portfolio with a 95% confidence, 1-day VaR of $10,000 is likely to lose more than $10,000 only 1 day in 20. With a 99% confidence 1-day VaR of $15,000, the portfolio is likely to lose at least that much 1 day in 100.

Investor Analytics uses several models to calculate a portfolio’s VaR. Our standard model uses the three concepts above: exposure, volatility and correlation. With each model, clients can specify the parameters and configure reports to display the VaR at each level within their portfolio. IA’s models are described further in the VaR Models section.

Although VaR has become an industry standard risk tool, it is only one of many measures that should be used. Some of our clients also look at a portfolio’s “worst case loss.”

As with any risk measure, there are multiple ways to calculate a worst case loss. For instance, determining the worst daily return in the portfolio’s actual history is one technique. Another is to determine the worst daily return of any security in the portfolio and apply that return to the entire portfolio. Although this is a very unlikely scenario (it would require all instruments to lose the same value simultaneously), it is a technique employed in the industry. A third method determines the worst loss suffered by each security in the portfolio over some time period (which can be considerably longer than the portfolio has existed) and adds those losses up. Again, this scenario is rather unlikely, but is used.

Comparisons of “worst case loss” and other risk measures (such as the more statistical VaR) provide a good way to monitor possible movements in the portfolio.


Determining the risk of a portfolio denominated in a foreign currency is not as simple as taking the final risk number and converting by the exchange rate. That's because each security's volatility is different when viewed from the foreign currency (since each security has a different correlation with that currency).

Because IA's system is completely currency neutral, and since all of our calculations are at a very granular level, we are easily able to present the risk of any portfolio, in any currency, and explicitly take into account the effects of foreign currency risk.

When people talk about "a stock's beta," it sounds like each stock has one and only one beta. Of course, each security has a beta for every index you want it compared to. IA uses the concept of a "Market Model," which is simply a list of indices or benchmarks that are meaningful to your portfolio. We calculate the betas of each security in your portfolio against each index in each of your Market Models.

We even calculate the betas of any part of your portfolio with those indices. Betas can be calculated to a single index or to several indices simultaneously, and the two methods often result in different beta values for the same security/index pair. While the formula for beta is similar to the formula for correlation, they are two different things: for example, correlation is always between +1 and -1, while beta can take on any value.


IA calculates all the Greeks for exchange traded and OTC options: delta, gamma, theta, vega and rho. Because we update these numbers daily, you're never left wondering about the sensitivity of your options positions to their underlying securities.

IA is working with select partner firms to add a suite of new capabilities to our system that include traditional fixed income analytics such as duration / convexity, DV01, and market value changes resulting from shocks to the yield curve. We’re also adding credit risk analytics for instruments like Credit Default Swaps. Please contact us to learn more.