February Hedge Fund Performance Forecast Review
Two weeks ago we wrote a post attempting to forecast February performance of three of BarclayHedge’s monthly hedge fund indices - the Barclay CTA Index, the Barclay Hedge Fund Index and the Barclay Equity Long/Short Index - using a set of daily factor indices. The technique was very simple - find the 10 highest correlating factors for each index for the past year and use their February returns as a proxy for the monthly index where the return for the month is not yet available. Here is the chart of our custom proxy indices vs their monthly targets:
The proxies have some high monthly correlations to their targets, too:
BarclayHedge has since published their official monthly numbers and this is our predicted vs current actual number (the current number keeps fluctuating for a while as more funds report their returns to BarclayHedge):
Hedge Fund Index Forecast -0.36% Actual -2.67%
CTA Index Forecast 0.60% Actual -0.17%
Equity Long/Short Index Forecast -0.77% Actual -2.06%
Of course, we know that using 10 daily vehicles to predict returns of an index consisting of thousands of funds has its limitations and while we did not expect a perfect match, the point of the experiment was to see if we can get anywhere close, at least directionally. And I think it worked reasonably well from that point of view: the Hedge Funds and Long/Short Equity indices got in the right direction, and although CTAs did not get into a positive territory they still show approximately 2% better returns than the other two indices and that is similar to our estimate - we expected CTAs to do better than the main index and the equity long/short index. The method was off on volatility/scale, and that is something we will keep working on to improve.
In a world with so much data it is easy to build models with such levels of complexity that they become impossible to understand or interpret. What we are trying to do here is the complete opposite - we want to find a simple and intuitive model that, while not perfectly accurate, is able to provide some meaningful information. Simpler models can shed light on the relationships behind complex systems with many moving parts in a way that is intuitive and, therefore, predictive, and that is the best property a financial or investment related model can provide.
Dmitri Alexeev is Founder and CEO of AlphaBot, a collaborative platform for alternative investment research.
© The Sortino Group Ltd
All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency or other Reprographic Rights Organisation, without the written permission of the publisher. For more information about reprints from AlphaWeek, click here.