Volatility-Managed Portfolio (Moreira and Muir)

May 2007 – Sep 2026	Strategy	SPY
Annual return (CAGR)	7.6%	10.7%
Worst drawdown	-17.3%	-55.2%
Sharpe ratio	0.78	0.62
Volatility	10.1%	19.7%
Annual return since publication (May 2016)	10.3%	15.3%

Jan 1960 – Sep 2026	Strategy	SPY
Annual return (CAGR)	8.6%	10.6%
Worst drawdown	-34.0%	-55.2%
Sharpe ratio	0.88	0.70
Volatility	10.0%	16.4%
Annual return since publication (May 2016)	10.3%	15.3%

Volatility-managed portfolios come from Alan Moreira and Tyler Muir's paper "Volatility-Managed Portfolios", circulated as an NBER working paper in 2016 and published in the Journal of Finance in 2017. The rule scales stock exposure by the inverse of last month's variance: invest less after volatile months and more after calm ones. This page tests the paper's no-leverage version on the S&P 500, with the rest in Treasury bills.

The idea#

Volatility clusters: a volatile month tends to be followed by another. But high volatility has not been followed by higher returns to compensate. So in volatile times investors take more risk without being paid more for it. Moreira and Muir showed that cutting exposure when recent volatility is high, and restoring it when volatility is low, raised the return per unit of risk for the stock market and for many other factors.

The rule is not a trend signal. It does not ask whether prices are rising or falling, only how much they have been moving.

How it works#

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

Measure the S&P 500's (SPY) annualized volatility over the last 21 trading days.
Set the stock share from that volatility:
- below 12.5%: 100%
- 12.5–15%: 71%
- 15–18%: 49%
- 18–22%: 33%
- 22–27%: 22%
- 27–35%: 14%
- 35–45%: 8%
- above 45%: 4%
Hold that share in SPY and the rest in Treasury bills (BIL).

These steps approximate the paper's formula, a stock weight proportional to one over last month's variance, capped at 100%. The scale is set so that the uncapped version would be as volatile as the market; on S&P 500 data from 1928 to 2015, that puts the formula at 100% stocks at 11.5% volatility. Each step holds the formula's value at the middle of its band. The paper's main version also used leverage after calm months; this one never goes above 100% stocks.

What the backtest shows#

With real prices the test starts in mid-2007. Since then the strategy returned about 7.5% a year, against about 11% for the S&P 500 and 8% for a 60/40 portfolio. Its risk-adjusted return was a little better than both, and its worst drawdown, about 17%, was far smaller than the S&P 500's 55% and the 60/40 portfolio's 33%.

The rule worked as intended in 2008, when rising volatility cut its stock share early and it lost 11% against 37% for the market. Its worst drawdown ran from July 2007 to March 2009 and was recovered by early 2011. The cost is in rebounds. Volatility stays high for a while after a crash, so the strategy re-enters slowly. It gained 8% in 2009 against 26%, and lost 3% in 2020 while the market gained 18%. In calm years such as 2017 it matched the market exactly.

Since the paper appeared in 2016 it has returned about 10% a year, with a clearly better risk-adjusted return than over the whole period and a similar worst drawdown, about 17%.

With simulated history the test starts in 1960. Over that span it returned about 8.6% a year against the S&P 500's 10.5%, again with a better risk-adjusted return. It lost 9% in 1974 while the market lost 27%, and gained 19% in 1987 against 5%. Its worst simulated drawdown, about 34% from late 1968 to May 1970, shows its blind spot: that bear market was slow and calm, so volatility stayed low and the strategy stayed fully invested.

When it struggles#

Slow, calm declines. A market that falls steadily without large daily swings keeps the strategy fully invested, as in 1969–70.
Sharp rebounds. After a crash, volatility stays high while prices recover, so it holds little stock through the best part of the rebound.
Long bull markets. Any volatile stretch cuts exposure, and missed gains are rarely made up.

Using it on Tactfolio#

The live strategy runs the eight-step ladder each month on SPY and BIL. The steps are an approximation of the paper's continuous weight. Copy it to try a different volatility window, a lower cap point, or a leveraged fund for the calmest months.

Year by year

Year	Strategy	SPY
2026*	8.6%	14.0%
2025	11.4%	17.7%
2024	21.4%	24.9%
2023	15.3%	26.2%
2022	-6.1%	-18.2%
2021	17.4%	28.7%
2020	-3.3%	18.3%
2019	16.4%	31.2%
2018	1.0%	-4.6%
2017	21.7%	21.7%
2016	8.8%	12.0%
2015	-4.9%	1.2%
2014	9.0%	13.5%
2013	25.4%	32.3%
2012	10.8%	16.0%
2011	1.2%	1.9%
2010	8.4%	15.1%
2009	7.8%	26.4%
2008	-11.0%	-36.8%
2007*	-3.6%	-3.4%

Year	Strategy	SPY
2026*	8.6%	14.0%
2025	11.4%	17.7%
2024	21.4%	24.9%
2023	15.3%	26.2%
2022	-6.1%	-18.2%
2021	17.4%	28.7%
2020	-3.3%	18.3%
2019	16.4%	31.2%
2018	1.0%	-4.6%
2017	21.7%	21.7%
2016	8.8%	12.0%
2015	-4.9%	1.2%
2014	9.0%	13.5%
2013	25.4%	32.3%
2012	10.8%	16.0%
2011	1.2%	1.9%
2010	8.4%	15.1%
2009	7.8%	26.4%
2008	-11.0%	-36.8%
2007	3.1%	5.1%
2006	15.2%	15.8%
2005	1.0%	4.8%
2004	9.5%	10.7%
2003	14.9%	28.2%
2002	-5.5%	-21.6%
2001	-5.0%	-11.8%
2000	-5.0%	-9.7%
1999	11.4%	20.4%
1998	11.4%	28.7%
1997	14.4%	33.5%
1996	19.8%	22.5%
1995	38.0%	38.0%
1994	1.3%	0.4%
1993	9.7%	9.7%
1992	8.1%	7.6%
1991	18.0%	30.3%
1990	-1.1%	-3.2%
1989	27.3%	31.5%
1988	10.3%	16.4%
1987	19.2%	5.1%
1986	7.7%	18.6%
1985	31.6%	31.6%
1984	6.4%	6.2%
1983	17.9%	22.4%
1982	15.6%	21.6%
1981	-2.5%	-4.9%
1980	25.1%	32.4%
1979	14.5%	18.3%
1978	4.9%	6.5%
1977	-7.3%	-7.2%
1976	17.6%	23.8%
1975	14.8%	37.1%
1974	-8.9%	-26.5%
1973	-9.7%	-14.7%
1972	18.9%	18.9%
1971	9.8%	14.1%
1970	1.9%	3.9%
1969	-10.2%	-8.4%
1968	8.2%	10.9%
1967	23.8%	23.8%
1966	-10.6%	-10.1%
1965	11.6%	12.4%
1964	16.4%	16.4%
1963	20.6%	22.7%
1962	-12.9%	-8.8%
1961	26.8%	26.8%
1960*	7.7%	0.4%

* 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

Follows the paper's no-leverage variant (Table 5), with the stock weight min(1, c / realized variance) set from the previous month's daily returns. c makes the uncapped strategy as volatile as the market, which on S&P 500 data for 1928–2015 means full investment at 11.5% annualized volatility or less; the paper sets c on its own CRSP sample.
The continuous weight is approximated by eight steps: 100% below 12.5% volatility, then 71%, 49%, 33%, 22%, 14%, 8%, and 4% above 45%, each the formula's value at the middle of its band.
SPY stands in for the market portfolio and BIL for the risk-free asset; the past month is the last 21 sessions.
Signals and trades use the close of the last trading day of each month, as in the source.

Common questions#

What is a volatility-managed portfolio?#

It is a portfolio that scales its stock exposure by recent volatility: less exposure after volatile months, more after calm ones. Alan Moreira and Tyler Muir showed in 2017 that this improves the return per unit of risk of the stock market and many other factors.

How is volatility targeting different from trend following?#

Trend following reacts to the direction of prices, such as a fall below a moving average. Volatility management reacts only to the size of recent price moves, so it can cut exposure in a volatile rising market and stay invested in a calm falling one.

Does the volatility-managed portfolio beat the S&P 500?#

Not on return. Since 2007 it returned about 7.5% a year against about 11% for the S&P 500, but with a far smaller worst drawdown and a better risk-adjusted return.

Why does the strategy use Treasury bills?#

The paper scales exposure to stocks in excess of the risk-free rate, so the unused share earns the Treasury bill rate. BIL holds that share here.