Wednesday, April 25, 2018

"The World’s Longest Multi-Asset Momentum Investing Backtest!"

Two quick points:

1) When implementing momentum as a strategy always, always keep in mind the little "Trend, Friend, Bend, End" rhyme. If the momo mamas start moving toward the exits the resulting countermoves can be disastrous for folks still on the dance floor.

2) Those equity databases going back 200 years have a whole lot of problems starting with survivorship bias, small sample sizes, reliance on financials (banks), i.e. no wooden turnpike companies, failed canals... etc. etc.
In his monumental “COMMON-STOCK INDEXES 1871-1937″ Mr. Cowles is quite explicit as to the reasons he and the Commission didn’t go further back than 1871. (pg. 4) 506 page PDF hosted at Yale.
I've mentioned "Common Stock Indexes..." and the Cowles Commission a few times, usually when referring to one of the industrial companies, New York Guano but also because of the Cowles Foundation's connection to some Nobelists:

Robert Shiller
Tjalling Koopmans
Kenneth Arrow
Gerard Debreu
James Tobin
 Franco Modigliani
Herbert Simon
Lawrence Klein
Trygve Haavelmo
Harry Markowitz

Not as prolific as Cambridge's Cavendish Lab, 29 Laureates, mainly in physics, at last count but then the econ version isn't one of the original Nobels and hasn't been around as long.

From Alpha Architect:
As evidenced by the image below, interest in momentum research has taken off since the original 1993 Jegadeesh and Titman paper:
 data:image/png;base64,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
Source: “Two Centuries of Multi-Asset Momentum (Equities, Bonds, Currencies, Commodities, Sectors and Stocks)” https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2607730
Most of the research on momentum is repetitive and reaks of data torture, but Geczy and Samanov have been conducting some fascinating out of sample research on the topic.

We have discussed prior research by Gezcy and Samonov here on U.S. equity momentum over 200 years of data. The evidence is similar to the academic finding that, on average, intermediate-term winning stocks continue winning, while intermediate-term losing stocks continue losing–otherwise known as a continuation of momentum for stocks.(1)

But studying 200+ years of equity data was not enough for Gezcy and Samanov — they have a new paper titled, “Two Centuries of Multi-Asset Momentum (Equities, Bonds, Currencies, Commodities, Sectors, and Stocks).”

This paper builds on the original paper (mentioned here), but adds (1) other asset classes and (2) examines time-series momentum. The paper also looks at how momentum performs in other asset classes and international equities. (similar to what Asness, Moskowitz, and Pedersen show in, “Value and Momentum Everywhere:” the momentum premia exists in international equities as well as other asset classes.

To follow-up on the “Value and Momentum Everywhere” study, this paper digs into how momentum works both within and across asset classes, over a longer time-period. The image below gives a graphical depiction of the number of asset classes studied in the paper across time.
 Image result for "The World’s Longest Multi-Asset Momentum Investing Backtest!"
Source: “Two Centuries of Multi-Asset Momentum (Equities, Bonds, Currencies, Commodities, Sectors and Stocks)” https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2607730
Below we dig into the results from the paper.
The Results
First, the paper examines how momentum works within (and across) each asset class. This is done by sorting securities into a relative “winner” and a relative “loser” portfolio by using the classic 12_2 momentum screen — this is the twelve-month momentum of each security excluding the last two months (so 10 months of returns). Within each universe, the top third are deemed “winners” and the bottom third are deemed “losers.”

The table below examines the returns for each asset class to the Winner, Loser, and Winner minus Loser portfolios....MORE