Paired Switching (quarterly SPY and TLT rotation)
Paired switching from Akhilesh Maewal and Joel Bock: at the end of each quarter, hold whichever of the S&P 500 and long-term Treasuries returned more over that quarter. The portfolio is always fully in one of the two, and it relies on the pair usually moving in opposite directions.
Designed by Akhilesh Maewal and Joel R. Bock, 2011. Implemented and tracked by Tactfolio.
| Oct 2002 – Sep 2026 | Strategy | SPY |
|---|---|---|
| Annual return (CAGR) | 9.1% | 11.4% |
| Worst drawdown | -35.6% | -55.2% |
| Sharpe ratio | 0.63 | 0.68 |
| Volatility | 15.7% | 18.6% |
| Annual return since publication (Sep 2011) | 6.5% | 15.1% |
| Apr 1962 – Sep 2026 | Strategy | SPY |
|---|---|---|
| Annual return (CAGR) | 9.3% | 10.6% |
| Worst drawdown | -35.6% | -55.2% |
| Sharpe ratio | 0.71 | 0.69 |
| Volatility | 14.0% | 16.6% |
| Annual return since publication (Sep 2011) | 6.5% | 15.1% |
Paired Switching is the stock and bond rotation that Akhilesh Maewal and Joel R. Bock described in their 2011 SSRN paper Paired-Switching for Tactical Portfolio Allocation. At the end of each quarter it moves the whole portfolio into whichever of the S&P 500 and long-term Treasuries returned more over that quarter.
The idea#
Maewal and Bock started from a simple observation about negatively correlated assets. If stocks and long Treasuries tend to move in opposite directions, their good stretches rarely overlap. By holding whichever did better last quarter, an investor rides each asset's stretch of above-average returns and switches when the lead passes to the other.
The authors used Vanguard's S&P 500 and long-term Treasury funds. Lewis Glenn later popularized a monthly version with a 3-month lookback, which Allocate Smartly tracks; the signal is the same, checked three times as often.
How it works#
At the close of the last trading day of each quarter (March, June, September, and December):
- Measure the total return of the S&P 500 (SPY) and long-term Treasuries (TLT) over the past quarter (63 sessions).
- Hold 100% of whichever returned more until the next quarter end.
There is no cash option. The portfolio is always fully in stocks or long bonds.
What the backtest shows#
The ETF-era test starts in late 2002, once TLT has a full quarter of history. Since then Paired Switching returned about 9% a year, against 11% for the S&P 500 and 8.4% for a 60/40 stock and bond mix. Its worst drawdown, about 36%, was smaller than the S&P 500's but deeper than the 60/40's, and its risk-adjusted return trailed both.
Its strength showed in 2008, when it gained 34% while the S&P 500 lost 37%, and in 2011, up 39%. Its weakness showed when stocks and bonds fell together: it lost 23% in 2022, more than the S&P 500. It also lost 14% in 2015 and 4% in 2024, a year stocks gained 25%. The worst drawdown ran from August 2019 to the March 2020 crash and was recovered in April 2021.
Since the paper appeared in 2011 it has returned about 6.5% a year with a lower risk-adjusted return than over the full period.
With simulated history the test starts in 1962 and returns about 9% a year against 11% for the S&P 500, with a similar risk-adjusted return and a worst drawdown of about 36%, still the 2019–2020 fall. It softened the 1974 bear market, losing 5% while stocks lost 27%, but lost more than stocks in 1973 and 2001.
When it struggles#
- Stocks and bonds falling together. The strategy depends on the two moving in opposite directions; in 2022 both fell, and it had nowhere to hide.
- Quick reversals. A quarterly check can hold the previous quarter's winner for months after the market turns.
- Long bond risk. TLT is volatile, so even the defensive side can lose heavily when rates rise.
Using it on Tactfolio#
The live strategy above follows Maewal and Bock's quarterly rule with SPY and TLT standing in for their Vanguard funds. Copy it and switch the rebalance to monthly to get Lewis Glenn's version, or compare it with Glenn's five-asset Quint Switching Filtered.
Year by year
| Year | Strategy | SPY |
|---|---|---|
| 2026* | -0.3% | 14.0% |
| 2025 | 4.0% | 17.7% |
| 2024 | -4.3% | 24.9% |
| 2023 | 26.1% | 26.2% |
| 2022 | -22.9% | -18.2% |
| 2021 | 19.4% | 28.7% |
| 2020 | -1.9% | 18.3% |
| 2019 | 12.2% | 31.2% |
| 2018 | -4.7% | -4.6% |
| 2017 | 16.7% | 21.7% |
| 2016 | 12.0% | 12.0% |
| 2015 | -13.5% | 1.2% |
| 2014 | 17.5% | 13.5% |
| 2013 | 32.3% | 32.3% |
| 2012 | 8.7% | 16.0% |
| 2011 | 39.0% | 1.9% |
| 2010 | 8.1% | 15.1% |
| 2009 | 26.7% | 26.4% |
| 2008 | 34.0% | -36.8% |
| 2007 | 6.4% | 5.1% |
| 2006 | 8.9% | 15.8% |
| 2005 | 4.9% | 4.8% |
| 2004 | -4.1% | 10.7% |
| 2003 | 16.7% | 28.2% |
| 2002* | -0.1% | -1.1% |
| Year | Strategy | SPY |
|---|---|---|
| 2026* | -0.3% | 14.0% |
| 2025 | 4.0% | 17.7% |
| 2024 | -4.3% | 24.9% |
| 2023 | 26.1% | 26.2% |
| 2022 | -22.9% | -18.2% |
| 2021 | 19.4% | 28.7% |
| 2020 | -1.9% | 18.3% |
| 2019 | 12.2% | 31.2% |
| 2018 | -4.7% | -4.6% |
| 2017 | 16.7% | 21.7% |
| 2016 | 12.0% | 12.0% |
| 2015 | -13.5% | 1.2% |
| 2014 | 17.5% | 13.5% |
| 2013 | 32.3% | 32.3% |
| 2012 | 8.7% | 16.0% |
| 2011 | 39.0% | 1.9% |
| 2010 | 8.1% | 15.1% |
| 2009 | 26.7% | 26.4% |
| 2008 | 34.0% | -36.8% |
| 2007 | 6.4% | 5.1% |
| 2006 | 8.9% | 15.8% |
| 2005 | 4.9% | 4.8% |
| 2004 | -4.1% | 10.7% |
| 2003 | 16.7% | 28.2% |
| 2002 | -1.5% | -21.6% |
| 2001 | -16.9% | -11.8% |
| 2000 | 13.3% | -9.7% |
| 1999 | 1.3% | 20.4% |
| 1998 | 13.7% | 28.7% |
| 1997 | 33.3% | 33.5% |
| 1996 | 6.9% | 22.5% |
| 1995 | 26.8% | 38.0% |
| 1994 | 0.3% | 0.4% |
| 1993 | 16.8% | 9.7% |
| 1992 | 7.3% | 7.6% |
| 1991 | 14.0% | 30.3% |
| 1990 | -4.5% | -3.2% |
| 1989 | 18.4% | 31.5% |
| 1988 | 13.9% | 16.4% |
| 1987 | 5.0% | 5.1% |
| 1986 | 13.5% | 18.6% |
| 1985 | 26.9% | 31.6% |
| 1984 | 10.0% | 6.2% |
| 1983 | 22.4% | 22.4% |
| 1982 | 46.2% | 21.6% |
| 1981 | 2.8% | -4.9% |
| 1980 | 5.3% | 32.4% |
| 1979 | 1.7% | 18.3% |
| 1978 | -3.3% | 6.5% |
| 1977 | -1.3% | -7.2% |
| 1976 | 30.5% | 23.8% |
| 1975 | 36.6% | 37.1% |
| 1974 | -5.1% | -26.5% |
| 1973 | -16.5% | -14.7% |
| 1972 | 15.4% | 18.9% |
| 1971 | 4.3% | 14.1% |
| 1970 | 10.4% | 3.9% |
| 1969 | -6.0% | -8.4% |
| 1968 | -5.1% | 10.9% |
| 1967 | 23.8% | 23.8% |
| 1966 | 1.3% | -10.1% |
| 1965 | 3.1% | 12.4% |
| 1964 | 16.4% | 16.4% |
| 1963 | 22.7% | 22.7% |
| 1962* | 15.1% | -6.7% |
* Partial year.
The rules as implemented
This is the exact tree Tactfolio runs, rebalanced quarterly with signals and trades at the close. Open it to inspect or copy it.
- StrategyMaewal-Bock Paired Switching
- WeightEqual
- RankTop 1 · 63d cumulative return
- TickerSPY
- TickerTLT
- RankTop 1 · 63d cumulative return
- WeightEqual
Sources and caveats
- Akhilesh Maewal and Joel R. Bock, Paired-Switching for Tactical Portfolio Allocation (2011)
- Lewis A. Glenn, Simple and Effective Market Timing with Tactical Asset Allocation (2014)
- SPY and TLT stand in for the paper's Vanguard 500 Index (VFINX) and Long-Term Treasury (VUSTX) funds; the prior quarter is 63 sessions.
- Follows the original quarterly rule. Lewis Glenn's later version, which Allocate Smartly tracks, checks the same 3-month return monthly.
- Signals and trades use the close of the last trading day of each quarter.
Common questions#
What is paired switching?#
It is a rotation between two negatively correlated assets, usually stocks and long Treasuries: at each period end, hold whichever did better over the period just ended. Maewal and Bock described it in 2011.
What ETFs does paired switching use?#
SPY for the S&P 500 and TLT for long-term Treasuries. The original paper used Vanguard's 500 Index and Long-Term Treasury mutual funds.
Is paired switching monthly or quarterly?#
Maewal and Bock's original switches quarterly, on the prior quarter's return. Lewis Glenn's 2014 version checks monthly using the past three months' return, which is the version Allocate Smartly tracks.
How is paired switching different from dual momentum?#
Dual momentum chooses between two stock markets and falls back to bonds only when stocks trail Treasury bills. Paired switching compares stocks directly with long bonds and never holds cash.