Is there a better way to gauge potential losses in a portfolio than the statistical model known as Value at Risk?
Four alumni of online brokerage TD Ameritrade think there is. And are starting to market risk management software to hedge funds, proprietary trading firms and active traders that uses a combination of backward- and forward-looking models to do so.
The Portfolio Defense program developed by Omaha-based PrairieSmarts assigns a higher than typical probability that unexpected big events, commonly called Black Swans, will occur when it calculates risks facing positions in stocks, options, mutual funds and exchange-traded funds.
The goal: Help traders figure out now just the range of risk – how far up or down the security could go in the immediate future –but the right amount – or size of position – to hold.
“The reality is, if you’ve got really good risk management software, you should be able to gauge what your position sizing is,’’ said Chris Nagy, one of the four principals, as well as “what you’re going to be long in and what you’re going to be short in. “
Nagy was managing director of order routing, sales and strategy at TD Ameritrade. Last year he founded consultancy KOR Trading. Other founders of PrairieSmarts include Casey Rockwell, a former TD Ameritrade engineer, Mike Chochon, former head of risk and trading at TD and Ron Piccinini, who was its director of portfolio modeling.
Managing partner Piccinini is the quant of the crew and contends classic models of risk underestimate Black Swan events, such as a credit crisis or market crash.
Into Portfolio Defense, Piccinini has incorporated a “look-back” model that tracks price movements of roughly 35,000 securities and mutual funds back to January 2006 and a “look-forward” model that projects future returns for about 5,000 names that have options traded on them, using a study of every option traded in the United States since 2002.
For any given combination of stocks, options and funds, the program takes about three-tenths of a second the value of positions, the amount at risk and how often in the past seven years the portfolio would have experienced a “ruin day.’’ That’s a day when the entire equity in the account is lost.
In a representative risk sample, a portfolio might have a value of $61,760, with 65.77% or $42,669.54 or 65.77% at risk in the next day of trading and a projected annualized return of -33.32%. And face one “ruin day.”
“A big part of it is open your eyes to what could happen,” Chochon said.
The system can adjust the time horizon for calculating the amount at risk for any number of days, weeks or months.
The system subsequently alters these estimates whenever new prices or positions are available, on a look-back basis always and on a forward-looking basis, where applicable.
“In practice, you want to take risks. Risk management is exactly that,” Piccinini said. Risk “is controlled through position sizing.”
