#Max drawdown geekforgeek series

But investors should still ask the tough questions: These fund managers dutifully pressed onward using their investment mandates. Many active funds and index funds struggled during the 2008 downturn. This fund has the rather unappetizing combination of low return and a large Maximum Drawdown. UBS Global Allocation Fund ( BPGLX) has a Maximum Drawdown of 48.7% and annual return of 2.6%.Vanguard LifeStrategy Growth Fund ( VASGX) has a Maximum Drawdown of 47.6% and annual return of 4.7%.Sound Mind Investing Fund ( SMIFX) has a Maximum Drawdown of 50.7% and annual return of 5.3%.Rowe Equity Income Fund ( PRFDX) has a Maximum Drawdown of 53.2% and annual return of 5.5%. Examples for the 10-year period ending January 31, 2017: By looking at the names or annual total returns for the funds below, we might expect them to be less risky.īut we can see a more complete picture by looking at the Maximum Drawdown: these funds each suffered a dramatic drop in value during the last market downturn. As you get older, you don't have the luxury of waiting and watching.' A quote from a subscriber sums up the difficulty of owning a portfolio with a large maximum drawdown: 'When you see 25, 30, 40% drawdown… you feel so helpless. Caveat emptor!Īnother reason this is a feared metric: it could frighten investors.and rightly so. It took the fund about 3 years to recover from that drawdown. This is 52.2% drawdown from a fund with 'Diversified' in its name. Hypothetical growth of $10,000 invested in Vanguard Diversified Equity Fund ( VDEQX) over the past 10 yearsīy looking at the graph's underlying data (available by hovering over the graph on the Vanguard website) we can see a peak value of $11,127.66 in 2007 followed by a low value of $5,318.39 in early 2009. Exhibit A (below) shows a graph that Vanguard publishes on .Įxhibit A.

To be fair, the fund providers are not obscuring or hiding the data, but it's not always easy to find.Ī savvy investor needs to be aware of the Maximum Drawdown for any investment under consideration. Fund providers don't want to point out the Achilles' heel of their funds. This metric would put many mutual funds and actively managed portfolios in an unfavorable light. The 'Risk Potential' ranges from 1 (low risk) to 5 (high risk). provides a sparse set of risk metrics for each mutual fund: R-Squared, Beta, and a 'Risk Potential' score.'s mutual fund screener includes Sharpe ratio, standard deviation, beta, and R-squared.Additionally, standard deviation penalizes portfolios by treating downside deviation (which is bad) and upside deviation (which is good) equally. Other risk metrics like standard deviation, while popular, can average out and thereby disguise periods of poor performance. Because Maximum Drawdown is by definition 'extreme,' it allows investors to see the worst-case that a portfolio has suffered. In other words, it's a portfolio's greatest loss in value over the time period you're considering.įor our analysis, the Maximum Drawdown is based on the total monthly return, including distributions. 'Maximum drawdown is defined as the peak-to-trough decline of an investment during a specific period and is usually quoted as a percentage of the peak value,' according to the asset management firm Robeco. You can't find this metric on the websites of Morningstar, Yahoo, Vanguard, or Fidelity. This metric has the power to make a decent-looking fund or portfolio look quite dreadful.

#Max drawdown geekforgeek series

While passing over the series we keep track of two numbers - the peak of the series and the maxDrawdown in price between peak and i.What is the metric that mutual fund providers fear? Iterating over each index i of the series in order, the maximum decline ending at point i will start at the highest point of the series so far. Instead of comparing every value with every other value, we can exploit the sequential requirement and make only n - 1 comparisons: var peak = 0 įor ( var i = 1 i dif ? maxDrawdown : dif For series covering a long period of time or with a great deal of granularity, this algorithm might be too slow even for a fast computer. The number of comparisons increases faster than n - if there are 100 values, 4950 comparisons are needed, while 200 values requires 19900 comparisons. If the series has n prices, n*(n - 1)/2 comparisons are needed (the number of comparisons done during the ith pass through the inner loop is n - i - 1 the sum of all those comparisons is 1 + 2 + 3 +.