Amber Quant Lab research
Overview / Strategies / Risk parity
Baseline · family risk_parity

Risk parity

Baseline  A well-known reference strategy. Not a candidate: it sets the bar new ideas must clear.

Validation Sharpe0.77train 0.47
Validation CAGR9.1%train 5.1%
Max drawdown-26.6%train -26.4%
Correlation to 60/400.65validation
Deflated Sharpe96%only 1 trial counted, so not yet discounted
PerformanceHow it works

Performance

Backtest after costs. Train is where the idea was developed; validation is where it was judged.

Growth of $1Risk parity60/40SPY
trainvalidation$0.5$1$1.5$2.5$42008201020122014201620182020Risk parityRisk parity $2.6560/4060/40 $3.49SPYSPY $4.51
Drawdown from peakRisk parity60/40
-30%-20%-10%0%Risk parity-25% (Oct 2008)60/402008201020122014201620182020
Calendar-year returnsRisk parity60/40
-20%0%20%2007 Risk parity: +23.5%2007 60/40: +7.5%’072008 Risk parity: -9.9%2008 60/40: -17.9%’082009 Risk parity: +8.5%2009 60/40: +12.8%’092010 Risk parity: +20.6%2010 60/40: +13.5%’102011 Risk parity: +14.0%2011 60/40: +7.7%’112012 Risk parity: +12.1%2012 60/40: +11.2%’122013 Risk parity: -9.3%2013 60/40: +15.6%’132014 Risk parity: +1.3%2014 60/40: +11.8%’142015 Risk parity: -8.2%2015 60/40: +1.7%’152016 Risk parity: +9.1%2016 60/40: +7.7%’162017 Risk parity: +19.7%2017 60/40: +13.7%’172018 Risk parity: -8.4%2018 60/40: -2.0%’182019 Risk parity: +24.7%2019 60/40: +22.0%’192020 Risk parity: +4.1%2020 60/40: +16.4%’202021 Risk parity: +8.7%2021 60/40: +15.0%’21
MetricTrain 2007–15Validation 2016–21
CAGR5.1%9.1%
Volatility10.4%11.0%
Sharpe0.470.77
Sortino0.650.87
Max drawdown-26.4%-26.6%
Calmar0.190.34
Worst month-13.9%-8.6%
Months up57%65%
Turnover / yr1.3×1.4×
Avg gross exposure1.61×1.84×
Correlation to SPY0.480.53
Correlation to 60/400.600.65
Beta to SPY0.230.32

Average holdings

TIPUS inflation-protected Treasuries45.3%
IEFUS Treasuries 7–10y38.8%
TLTUS Treasuries 20y+17.6%
SPYUS large-cap equities15.8%
GLDGold14.7%
EFADeveloped ex-US equities13.7%
DBCBroad commodities13.7%
EEMEmerging-market equities10.3%

How it works

Inverse-volatility weights across stocks, Treasuries, inflation-linked bonds, gold and commodities, levered to a volatility target (capped). Monthly.

Why it should work. Balance risk, not dollars, across assets that win in different economic regimes (growth up/down, inflation up/down) — inspired by Bridgewater's All Weather. Relies on bonds and stocks not falling together for long (2022 was the stress test).