Volatility Contribution in Global Funds

View PDF here.

How do you identify volatility drivers in your fund?

Global funds pursue different investment strategies across multiple asset classes and countries around the world.  To carry out a portfolio construction process with a desired volatility range and periodic market risk evaluation, global funds need to look at how one or more position(s) contribute(s) to overall fund volatility and, more specifically, how node(s), a unique position, country or asset class, is/are contributing to fund volatility changes on a daily basis. This applies to absolute return strategy global macro hedge funds as well as relative return strategy mutual funds that have a benchmark index assigned. The end objective is to identify which security, node or strategy is driving fund volatility and the changes in fund volatility.

There are two methods to evaluate volatility contribution. The building block of each is to first calculate a simulated PnL distribution (historical or Monte - Carlo) for each position and all nodes. Once PnL distribution is calculated at position level, which is the lowest denominator, calculating the node PnL distribution is just a simple summation across time series for all securities that make up the particular node. Similarly, portfolio PnL distribution is a summation of either all non-overlapping nodes or all positions’ PnL distributions.

The first method calculates the node contribution by shocking the node exposures by a small amount, resulting in a small change in node PnL distribution (say by 1%). Change in portfolio nominal ($) volatility is calculated due to this small 1% change in node PnL distribution and then scaled to 100%. Dividing the scaled nominal volatility change in portfolio volatility by current portfolio NAV provides us contribution from this particular node. Please refer to the Appendix – Method 1 for more details on the mathematical formula. Summing volatility contribution numbers for all positions or all non-overlapping nodes leads exactly to portfolio volatility calculated directly from portfolio PnL distribution, which confirms that the calculations used are correct.

The second method relies on the fact that volatility contribution of a node is a product of a node’s isolated/absolute volatility and its correlation with the fund. Please refer to the Appendix - Method 2 for a detailed derivation of the concept. The additional benefit of this methodology is that it shows whether the node contribution is driven more by its own isolated volatility or its correlation with the fund. There are cases with nodes having high isolated volatility but contribute less towards overall portfolio volatility due to low correlation with fund. Similarly, nodes with low isolated volatility can contribute more towards overall portfolio volatility due to their high correlation with fund. 

This can be verified summing up all contributions from nodes, which equals portfolio volatility calculated directly from Portfolio PnL distribution. This proves that both methods are consistent with one another. It can be inferred that all security and node level contribution from both methods are equal to each other.  In implementations, we found that the choice to work with nominal {$) volatility and not return volatility for volatility contribution calculations works well with a multi asset portfolio with a lot of derivatives exposure. It avoids the complication that arises in defining return volatility when the exposures/market values approach zero in the simulation process.

Gravitas recommends that clients set up a periodic (daily, weekly and monthly) reviews of volatility contribution and the changes incurred from the last instance.  The report also highlights nodes contributing to more than a threshold level using a color-coded heat map. This threshold level in contribution calculation is a configuration choice that clients make depending on fund target volatility. The day-over-day/week-over-week change report highlights whether the change is due to an isolated/absolute volatility jump or a correlation jump. 

 

About The Gravitas Risk Solution.

The Gravitas Platform provides a unique integrated enterprise risk solution that positions funds to meet all internal, investor and regulator driven demands.  Funds are also prepared to meet current and potential future transparency and analytics demands including Opera and Form PF Risk.  The Gravitas Risk solution enables funds to augment portfolio management while providing insights to support investment decisions.  

The Gravitas Risk solution includes: Market Risk, Credit & Counterparty Risk, Investor & Regulatory Risk Reports, Custom Risk Dashboards, and PNL Attribution.  Benefits include: Institutional Coverage, Customization, Highly Integrated, Cost Savings, and Revenue Growth.

 

About Gravitas.

Gravitas is a leading collaborative outsourcing partner for the alternative investment space. Founded in 1996, Gravitas provides solutions for the front and middle office leveraging proprietary software, The Gravitas Platform, which is powered by The Gravitas Cloud.  The company is based in New York with offices in Chicago, Greenwich, Mumbai and Ahmedabad, India. 

Gravitas was most recently named "Best Middle Office Solution" in the HFM US Hedge Fund Services Awards, and "Best Outsourced Solution" in the 2015 HFM US Technology Awards. Past HFM US Hedge Fund Services Awards won by Gravitas include: “Most Innovative Technology Provider” (2014), “Best Outsourced Technology Infrastructure Provider” (2013) and “Best Overall Technology Firm” (2012).