My publisher and others frown upon my tendency to write dirty finance examples. Of course, we wouldn’t want to corrupt young minds under any circumstance. So to satisfy my urges I am just writing here in my little corner on the Net. Let’s keep it quiet! Shhh shh …

You might have heard of the Sharpe ratio, which is presumably one of the most common metrics used to pick stocks. Unfortunately, this metric assumes a normal distribution and therefore doesn’t take into account tail risk, Black Swans etc. Another ratio, called the Calmar ratio, is more robust in this respect. It is similar to the Sharpe ratio, but uses the maximum returns drawdown instead of the standard deviation. The maximum drawdown is the largest drop of returns, when considering sequential tops and bottoms. It is the largest peak to through decline. The following NumPy code calculates both ratios for a number of high dividend ETFs, that I selected from http://finviz.com/screener.ashx?v=111&f=fa_div_high,ind_exchangetradedfund,sh_avgvol_o50,sh_opt_option and writes the results to a file. So these are optionable (for hedging) ETFs with a bit more then minimal volume. I also computed the ratios for SPY as a benchmark. Also note that I am cutting corners here and not bothering with the risk free rate.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | import numpy as np import sys from scipy.signal import argrelmin, argrelmax import csv ... data download code omitted ... c = df['Close'] rets = np.diff(np.log(c)) #convert to simple returns rets = np.exp(rets) - 1 sharpe = rets.mean()/rets.std() mins = np.ravel(argrelmin(rets)) maxs = np.ravel(argrelmax(rets)) extrema = np.concatenate((mins, maxs)) extrema.sort() calmar = -rets.mean()/np.diff(rets[extrema]).min() with open('out/ratios.csv', 'a') as fp: writer = csv.writer(fp, delimiter=',') row = [sys.argv[1], round(sharpe, 4), round(calmar, 4)] writer.writerow(row) |

And I get the following table:

symbol,Sharpe,Calmar SPY,0.0284,0.0014 AMLP,0.0267,0.0024 CVY,0.0078,0.0006 DWX,-0.0047,-0.0005 EDIV,-0.0267,-0.0041 EMLC,-0.0129,-0.0018 HYD,0.0213,0.0017 HYG,-0.0012,-0.0001 HYLD,0.0142,0.0009 IFGL,-0.0061,-0.0005 JNK,-0.0008,-0.0 MDIV,0.0301,0.0048 PCEF,0.001,0.0001 PFF,-0.0003,-0.0 PGF,0.0008,0.0001 PGX,-0.0032,-0.0002 PSP,-0.0055,-0.0004 REM,-0.0282,-0.0027 SDIV,0.0018,0.0003 SJNK,0.0347,0.0046

Clearly just a handful of the selected ETFs have a performance that is comparable to the SPY benchmark.