What is the PAA Strategy — Rules and Method

The idea of protective momentum

Keller and Keuning published PAA in 2016. Their aim was to keep losses small enough to compete with a one-year term deposit, and the paper calls the mechanism 'protective momentum'.

Ordinary dual momentum looks at each asset's trend on its own. PAA looks at the breadth of the whole market: it counts how many of twelve risky assets still have a live trend, and moves more into bonds as that count falls. Because assets tend to move together, the trend of the others carries information about risk too.

The universes

The risky universe holds twelve: US large caps, the Nasdaq 100 and small caps, Europe, Japan, emerging markets, US REITs, commodities, gold, high yield corporates, investment grade corporates and US long treasuries.

The paper used EEM for emerging markets and IYR for REITs. This site uses VWO and VNQ, which the same authors' DAA paper names for those slots — so the twelve risky assets are exactly DAA's.

Following the paper's main results, the safe asset is a single one: US intermediate treasuries (IEF).

The score — against a 13-month average

The score is this month-end's price divided by the average of the last 13 month-end prices, minus one. It is 13 prices including this month because the paper defines a 12-month lookback as 12 intervals, which means 13 prices.

A score above zero counts as good; zero or below counts as bad. The same score is used to rank the assets that get held.

The score is not a return. It says how many percent the current price sits above or below that average, which is why this site labels it 'vs. 13-month average'.

Same number of holdings, a moving bond share

Each month-end we count the good assets. Six or fewer means 100% bonds. Seven means 83% bonds, eight 67%, nine 50%, ten 33%, eleven 17%, and all twelve 0%. The bond share has only these seven steps.

Whatever is not in bonds is split equally across the six highest scores. With 50% in bonds, each of the six gets about 8.3%. Keeping six positions and shrinking their weights is what separates PAA from DAA.

The paper describes three levels of protection. This site calculates only the strongest, PAA2, which the paper adopts. With the weakest, PAA0, the bond share falls in proportion to the number of good assets; the medium level, PAA1, sits in between.

Seen alongside the others

GEM moves between stocks and bonds on a single 12-month return. VAA evacuates entirely when one of four readings turns negative. DAA keeps two detector assets apart and steps back in two stages.

PAA is the only one of the four that uses a moving average, and its bond share moves in seven steps. That is why PAA often reads differently even in months when the other three point the same way.

Reading it on this site

The signal page opens with the allocation, and the line below says how many of the twelve sit above their 13-month average along with the bond share. A defensive label on its own cannot tell 17% bonds from 100%.

The bar table shows the score this strategy uses by default. The switcher can show 13612W or period returns instead, but the value behind the decision is the score against the 13-month average.

Worth knowing

This page explains a rule. It is not advice to buy or sell anything; the decision and its consequences are yours.

The paper reports results over past periods. We do not reproduce those figures — we have not verified them.

This strategy's archive builds up from the month it was added to this site; earlier months are not filled in.