US Risk Parity Trend-Following (The Trend is Our Friend)

May 2007 – Sep 2026	Strategy	SPY
Annual return (CAGR)	4.3%	10.7%
Worst drawdown	-10.1%	-55.2%
Sharpe ratio	0.74	0.62
Volatility	5.9%	19.7%
Annual return since publication (Feb 2016)	4.6%	15.7%

May 1969 – Sep 2026	Strategy	SPY
Annual return (CAGR)	8.7%	10.7%
Worst drawdown	-11.1%	-55.2%
Sharpe ratio	1.54	0.68
Volatility	5.5%	17.2%
Annual return since publication (Feb 2016)	4.6%	15.7%

US Risk Parity Trend-Following applies the method of The Trend is Our Friend, by Andrew Clare, James Seaton, Peter N. Smith, and Stephen Thomas, to five US asset classes. The paper first appeared as a working paper in 2012 and in the Journal of Behavioral and Experimental Finance in 2016. Each asset gets a share sized by its volatility, and any asset below its 10-month average has its share moved to Treasury bills.

The idea#

The authors combined two well-known tools. Risk parity gives each asset a weight in inverse proportion to its volatility, so calm assets such as Treasuries get large shares and volatile ones such as small caps get small shares, and no single asset dominates the portfolio's risk. Trend following, in the form Mebane Faber popularized, holds an asset only while its price is above its 10-month average.

In their tests, risk parity alone matched the return of an equal-weighted mix with much lower volatility, and trend following alone cut its drawdowns. Together they gave a higher risk-adjusted return than either, with much smaller drawdowns. Each asset keeps its slot even while in cash, so the portfolio never piles into whatever is left standing.

How it works#

At the close of the last trading day of each month:

Measure each of five funds' volatility over the past year: large-cap stocks (SPY), small-cap stocks (IWM), 7–10 year Treasuries (IEF), commodities (DBC), and real estate (VNQ).
Give each fund a share in proportion to one over its volatility, so the calmest funds get the most.
For each fund, hold it if its price is above its 10-month (210-session) average; otherwise hold its share in Treasury bills (BIL). The other shares do not grow to fill the gap.

As in the paper, each share comes from the fund's own volatility even while that share sits in bills, so the calmest fund, usually IEF, gets the largest slice.

What the backtest shows#

Over the ETF era, from mid-2007, the strategy returned about 4.3% a year, well below the S&P 500's 11% and a 60/40 mix's 8%. In exchange it was far steadier: its worst drawdown was about 10%, against 55% for the S&P 500 and a third for the 60/40. Its risk-adjusted return beat the S&P 500's and was level with the 60/40's.

That drawdown came quickly, in July and August 2011, and took until February 2013 to recover. In 2008 it gained 4% while the S&P 500 lost 37%. In 2022 it lost 5%, under a third of the S&P 500's loss, and in 2023 it lost 1% while stocks rebounded 26%. Its best year was 2021, up 13%.

Since the journal version appeared in 2016 it has returned about 5% a year with a better risk-adjusted return than over the full period and a worst fall under 10%.

With simulated history the test starts in 1969 and returns about 9% a year, close to the S&P 500's 11%, with a risk-adjusted return more than twice as high and a worst drawdown of about 11%, in the October 1987 crash. It gained in 1974 and in 2000–2002 while stocks fell. Higher bill and bond yields before 2008 helped; before 1980 gold stands in for the commodity fund.

When it struggles#

Low interest rates. Much of the portfolio sits in bills or bonds at any time, so returns shrink when yields are near zero.
Rising rates. Treasuries usually get the largest share, and they fell with stocks in 2022.
Sharp rebounds. Assets that dropped below trend come back only at a month end, after much of a recovery.

Using it on Tactfolio#

The live strategy above runs these rules with volatility from a year of daily returns in place of twelve monthly returns, a 210-session daily average in place of ten month-end prices, and dividend-adjusted prices for the trend test. Copy it to change the funds or the volatility window, or compare it with the equal-weighted GTAA 5, which uses the same trend filter.

Year by year

Year	Strategy	SPY
2026*	7.6%	14.0%
2025	4.9%	17.7%
2024	6.3%	24.9%
2023	-1.1%	26.2%
2022	-5.5%	-18.2%
2021	13.4%	28.7%
2020	4.6%	18.3%
2019	7.6%	31.2%
2018	0.3%	-4.6%
2017	6.4%	21.7%
2016	5.2%	12.0%
2015	-1.5%	1.2%
2014	5.5%	13.5%
2013	7.4%	32.3%
2012	3.6%	16.0%
2011	0.9%	1.9%
2010	7.5%	15.1%
2009	3.9%	26.4%
2008	4.2%	-36.8%
2007*	3.7%	-3.4%

Year	Strategy	SPY
2026*	7.6%	14.0%
2025	4.9%	17.7%
2024	6.3%	24.9%
2023	-1.1%	26.2%
2022	-5.5%	-18.2%
2021	13.4%	28.7%
2020	4.6%	18.3%
2019	7.6%	31.2%
2018	0.3%	-4.6%
2017	6.4%	21.7%
2016	5.2%	12.0%
2015	-1.5%	1.2%
2014	5.5%	13.5%
2013	7.4%	32.3%
2012	3.6%	16.0%
2011	0.9%	1.9%
2010	7.5%	15.1%
2009	3.9%	26.4%
2008	4.2%	-36.8%
2007	7.0%	5.1%
2006	10.7%	15.8%
2005	5.1%	4.8%
2004	11.5%	10.7%
2003	17.9%	28.2%
2002	7.2%	-21.6%
2001	4.5%	-11.8%
2000	17.9%	-9.7%
1999	6.4%	20.4%
1998	6.6%	28.7%
1997	13.3%	33.5%
1996	20.4%	22.5%
1995	22.9%	38.0%
1994	-1.4%	0.4%
1993	11.5%	9.7%
1992	7.8%	7.6%
1991	23.4%	30.3%
1990	0.8%	-3.2%
1989	21.7%	31.5%
1988	8.4%	16.4%
1987	3.8%	5.1%
1986	13.2%	18.6%
1985	23.5%	31.6%
1984	8.9%	6.2%
1983	14.9%	22.4%
1982	23.4%	21.6%
1981	7.1%	-4.9%
1980	15.6%	32.4%
1979	18.3%	18.3%
1978	7.1%	6.5%
1977	5.2%	-7.2%
1976	22.9%	23.8%
1975	5.3%	37.1%
1974	10.7%	-26.5%
1973	-1.6%	-14.7%
1972	14.0%	18.9%
1971	8.9%	14.1%
1970	10.4%	3.9%
1969*	-2.4%	-10.4%

* 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.

Sources and caveats

The paper uses five global indexes; this US adaptation, tested by Allocate Smartly, holds SPY, IWM, IEF, DBC, and VNQ.
Each asset is weighted by the inverse of its own trailing volatility, and an asset below its 10-month average keeps that weight in Treasury bills, as in the paper; the portfolio is not renormalized.
Volatility uses 252 daily returns instead of 12 monthly returns, and the trend test compares the close with its 210-session average instead of ten month-end closes, on dividend-adjusted prices; the paper uses price indexes for stocks and real estate.
Signals and trades use the close of the last trading day of each month, as in the source.

Common questions#

What is risk parity trend following?#

It is a portfolio that sizes each asset by the inverse of its volatility and holds each one only while it is above its 10-month moving average, moving that asset's share to Treasury bills otherwise. Clare, Seaton, Smith, and Thomas tested it in The Trend is Our Friend.

What ETFs does the US version use?#

SPY, IWM, IEF, DBC, and VNQ as the five asset classes, and BIL for cash. The paper itself used five global indexes; see the global version.

How is it different from GTAA 5?#

GTAA 5 gives each of its five assets an equal fifth. This strategy gives calmer assets more weight, which tends to lower both volatility and return.

Why does it hold so much cash?#

Each asset's share goes to Treasury bills whenever that asset is below its 10-month average, and some asset usually is. Its share stays in bills until the asset recovers.