Pragmatic Asset Allocation (Quantpedia’s PAA)
Quantpedia's Pragmatic Asset Allocation, in a monthly form: the two strongest of the Nasdaq-100, global stocks, and emerging market stocks by 12-month return, each only while above its 12-month average. Failing slots, and the whole portfolio when the yield curve is inverted, go to the best of Treasuries, gold, and cash.
Designed by Radovan Vojtko and Juliána Javorská (Quantpedia), 2024. Implemented and tracked by Tactfolio.
| Mar 2009 – Sep 2026 | Strategy | SPY |
|---|---|---|
| Annual return (CAGR) | 9.6% | 15.7% |
| Worst drawdown | -22.0% | -33.7% |
| Sharpe ratio | 0.70 | 0.94 |
| Volatility | 14.6% | 17.2% |
| Annual return since publication (Feb 2024) | 25.5% | 20.2% |
Pragmatic Asset Allocation (PAA) is a model for semi-active investors from Radovan Vojtko and Juliána Javorská of Quantpedia, published in January 2024. It holds the two strongest of three stock markets, the Nasdaq-100, global stocks, and emerging markets, each only while in an uptrend. It moves to a hedging mix of Treasuries, gold, and cash when a market fails its trend or the yield curve inverts.
The idea#
The authors wanted rules an individual investor could follow with a few trades a year. Momentum picks the stock markets worth owning, and a trend filter keeps each out of a downtrend. The inverted yield curve, a classic recession warning, overrides everything and moves the portfolio to hedges.
In the original, the portfolio is split into four parts that are rebalanced in turn once a quarter and each held for a year, partly for long-term tax treatment. There is also a stop-loss for positions held more than a year.
How it works#
This version applies the same signals to the whole portfolio at the close of the last trading day of each month:
- If the 3-month Treasury yield is above the 10-year yield (FRED series T10Y3M below zero), hold only the hedging mix.
- Otherwise, rank QQQ, ACWI, and EEM by 12-month return and take the top two at 50% each. A chosen market below its 12-month average passes its half to the hedging mix.
- The hedging mix takes the two strongest of intermediate Treasuries (IEF), gold (GLD), and cash (BIL) by 12-month return. Treasuries or gold below their 12-month average are replaced by cash.
This is a proxy for the published model. Without the four staggered year-long parts and the stop-loss, it trades far more often and reacts to every monthly signal.
What the backtest shows#
The test starts in March 2009, once the global stock fund has a year of history. Since then it returned about 9.6% a year, against 15.7% for the S&P 500 and 10.4% for a 60/40 stock and bond mix. Its worst drawdown was 22%, similar to the 60/40's, and its risk-adjusted return trailed both. Its worst drawdown ran from April 2010 to June 2012, recovered by October 2013; it lost 12% in 2011 as markets whipsawed around their trends.
It did best in trending years, such as 2017 (34% against 22%) and 2020 (32% against 18%). In 2019 it made only 7% against 31%, partly because the inverted yield curve moved it to hedges for several months.
Since publication at the start of 2024 it has returned about 25% a year, well ahead of the S&P 500.
When it struggles#
- Whipsaws. Monthly trend checks on three volatile markets switch often in choppy years such as 2011 and 2015.
- False recession warnings. An inverted yield curve without a recession, as in 2023 and 2024, moves it to hedges during rising markets.
- Missing the S&P 500. Its stock choices are the Nasdaq-100, global, and emerging markets, so it can lag when US large caps lead.
Using it on Tactfolio#
The live strategy above runs the paper's funds and signals monthly on the whole portfolio. Copy it to change the markets or the hedges, or compare it with TrendYCMacro, another Quantpedia strategy that uses the yield curve.
Year by year
| Year | Strategy | SPY |
|---|---|---|
| 2026* | 25.7% | 14.0% |
| 2025 | 27.1% | 17.7% |
| 2024 | 14.2% | 24.9% |
| 2023 | 5.3% | 26.2% |
| 2022 | -16.7% | -18.2% |
| 2021 | 13.8% | 28.7% |
| 2020 | 31.9% | 18.3% |
| 2019 | 7.4% | 31.2% |
| 2018 | 1.9% | -4.6% |
| 2017 | 33.5% | 21.7% |
| 2016 | -0.6% | 12.0% |
| 2015 | -6.0% | 1.2% |
| 2014 | 8.7% | 13.5% |
| 2013 | 26.8% | 32.3% |
| 2012 | 1.5% | 16.0% |
| 2011 | -12.2% | 1.9% |
| 2010 | 3.0% | 15.1% |
| 2009* | 19.4% | 38.7% |
* Partial year.
The rules as implemented
This is the exact tree Tactfolio runs, rebalanced monthly with signals and trades at the close. Open it to inspect or copy it.
- StrategyPragmatic Asset Allocation
- WeightEqual
- If10-year minus 3-month Treasury yield is below 0Then
- WeightEqual
- RankTop 2 · 252d cumulative return
- CandidateIEF
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEFThen
- WeightEqual
- TickerIEF
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEF
- WeightEqual
- CandidateGLD
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLDThen
- WeightEqual
- TickerGLD
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLD
- WeightEqual
- TickerBIL
- CandidateIEF
- RankTop 2 · 252d cumulative return
Otherwise- WeightEqual
- RankTop 2 · 252d cumulative return
- CandidateQQQ
- WeightEqual
- Ifcurrent price of QQQ is above 252d moving average of QQQThen
- WeightEqual
- TickerQQQ
Otherwise- WeightEqual
- RankTop 2 · 252d cumulative return
- CandidateIEF
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEFThen
- WeightEqual
- TickerIEF
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEF
- WeightEqual
- CandidateGLD
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLDThen
- WeightEqual
- TickerGLD
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLD
- WeightEqual
- TickerBIL
- CandidateIEF
- RankTop 2 · 252d cumulative return
- WeightEqual
- Ifcurrent price of QQQ is above 252d moving average of QQQ
- WeightEqual
- CandidateACWI
- WeightEqual
- Ifcurrent price of ACWI is above 252d moving average of ACWIThen
- WeightEqual
- TickerACWI
Otherwise- WeightEqual
- RankTop 2 · 252d cumulative return
- CandidateIEF
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEFThen
- WeightEqual
- TickerIEF
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEF
- WeightEqual
- CandidateGLD
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLDThen
- WeightEqual
- TickerGLD
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLD
- WeightEqual
- TickerBIL
- CandidateIEF
- RankTop 2 · 252d cumulative return
- WeightEqual
- Ifcurrent price of ACWI is above 252d moving average of ACWI
- WeightEqual
- CandidateEEM
- WeightEqual
- Ifcurrent price of EEM is above 252d moving average of EEMThen
- WeightEqual
- TickerEEM
Otherwise- WeightEqual
- RankTop 2 · 252d cumulative return
- CandidateIEF
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEFThen
- WeightEqual
- TickerIEF
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of IEF is above 252d moving average of IEF
- WeightEqual
- CandidateGLD
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLDThen
- WeightEqual
- TickerGLD
Otherwise- WeightEqual
- TickerBIL
- WeightEqual
- Ifcurrent price of GLD is above 252d moving average of GLD
- WeightEqual
- TickerBIL
- CandidateIEF
- RankTop 2 · 252d cumulative return
- WeightEqual
- Ifcurrent price of EEM is above 252d moving average of EEM
- WeightEqual
- CandidateQQQ
- RankTop 2 · 252d cumulative return
- WeightEqual
- If10-year minus 3-month Treasury yield is below 0
- WeightEqual
Sources and caveats
- Radovan Vojtko and Juliána Javorská, Pragmatic Asset Allocation Model for Semi-Active Investors (2024)
- Radovan Vojtko and Juliána Javorská, Pragmatic Asset Allocation Model for Semi-Active Investors, SSRN (2024)
- The source splits the portfolio into four tranches that trade one month after each quarter end and hold for 12 months, with a stop-loss for positions held over a year; the builder cannot hold tranches, so this version applies the same signals to the whole portfolio every month and trades far more.
- Stocks are QQQ, ACWI, and EEM and hedges IEF, GLD, and BIL, as in the source; momentum is the 12-month (252-session) return and trend the 252-session average.
- A stock among the two strongest by momentum that is below its average routes its slot to the hedges; the source picks the two strongest among those in an uptrend, which differs only when a weaker stock is trending and a stronger one is not.
- The yield curve test is the daily 10-year minus 3-month Treasury spread (FRED T10Y3M); the source uses monthly data.
- Each economic figure counts only from the day it was first published, and as first reported: at a month end the latest figure is the previous month's, the one-month lag the source describes.
- Signals and trades use the close of the last trading day of each month, as in the source.
Common questions#
What is Pragmatic Asset Allocation?#
It is Quantpedia's model that holds the two strongest of the Nasdaq-100, global stocks, and emerging markets while each is in an uptrend, and moves to Treasuries, gold, or cash when a market fails its trend or the yield curve inverts.
Why does the yield curve matter in PAA?#
When short-term Treasury yields rise above long-term yields, the curve is inverted, which has preceded most US recessions. PAA then holds only its hedging assets.
How is this version different from the paper?#
The paper splits the portfolio into four parts, each traded once a year with a stop-loss. This version applies the signals to the whole portfolio every month, so it trades more and reacts faster.
Does Pragmatic Asset Allocation still work?#
Since publication in 2024 it has returned about 25% a year. The live record on Tactfolio shows how it is doing now.