Stress Tests
The easiest way to find stress tests in the application is to go to the Fields drop-down menu, and type “Stress Test”. We frequently add new stress tests. The application includes (at least) the following benchmark stress tests:
2020 US Election Stress Tests
Black Monday
COVID
Curve +1%
Emerging Markets Currency Crisis
Financial Crisis
First Gulf War
Flight to Quality
Flight to Euro
Jan 2009 Interest Rise
Market -10%
Quant Meltdown
Rates +1%
Sideways Market 30
Taper Tantrum
VIX(2x)
Volatility Surface Stress Tests
Worst Day
Stress tests are used to forecast how a portfolio will react in extreme market environments. Stress tests can be based either on historical events (Black Monday, Russian Debt Crisis, etc.), or hypothetical scenarios (e.g. equity markets down 10%, and gold up 20%).
Whether based on historical or hypothetical scenarios, stress tests are usually defined by at most a handful of factors. How the securities in your portfolio react depends on if the stress test is predictive or simple. In a simple stress test, only securities that are part of the stress test definition and securities that are derivatives of those securities are impacted. In a predictive stress test, securities react based on their correlation to the instruments that define the stress test. For example, suppose we have a portfolio that owns shares of GOOG and puts on the S&P 500, and we want to define a stress test where the S&P 500 moves down 10%. In the simple version of the stress test, the S&P 500 puts will make money, but the GOOG shares will not move. In the predictive version, the S&P 500 puts will make money, and (assuming GOOG is positively correlated with the S&P 500) the GOOG shares will lose money. Most of our default stress tests are predictive, but there are instances where using simple stress tests can be useful.
Because stress tests deal with extreme events it is important that options and any other derivative instruments are fully repriced. In our stress tests, derivatives are always fully repriced.
For more about stress tests, including return-based weighting (RW), see our white paper, Stress Tests and Correlation.
Predictive Stress Tests
2020 US Election Stress Tests
We have added two stress tests for the 2020 US Presidential Election, “Biden Surprise Win” and “Trump Surprise Win”. For more on how to interpret these stress tests, and how we designed them, read more here.
Black Monday
Black Monday occurred on October 19, 1987. This stress test is based on the 5-day period, 10/13/1987 – 10/19/1987. We assume a fall in the S&P 500 of −27.33% and the VIX going to 71.58. All returns for all other securities are based on their current estimated correlation to these benchmarks.
∆SPX = −27.33%
VIX = 71.58
COVID
This stress test focuses on the start of the COVID crisis, March 2020. The crisis impacted different markets at different times. Rather than choosing one 5-day period, then, we combined different 5-day periods, between 3/12/2020 and 3/19/2020.
∆SPX = −17.97%
∆US Bond Excess = −5.27%
∆Small Cap - Large Cap = −6.98%
∆Energy Excess = −15.05%
∆Technology Excess = +5.84%
Curve +1%
In this stress test, the USD Swap Curve shifts up by 1%. This stress test impacts all securities, regardless of asset class, based on their historical correlation to shifts in this interest rate curve.
Financial Crisis
Our financial crisis stress test is based on the 5-day period, 10/3/2008 – 10/9/2008. We assume a fall in the S&P 500 of −18.34%, a decline in financial stocks of −30.30% and a VIX of 54.46. All returns for all other securities are based on their current estimated correlation to these benchmarks*.
∆SPX = −18.34%
∆XLF = −30.30%
VIX = 54.46
*To avoid problems of multicollinearity between SPX and XLF, the calculations are actually based on our financials excess return factor, which is based on the spread between XLF and SPX. The return for this factor during the 5-day stress period was −4.96%.
First Gulf War
Iraq invaded Kuwait on August 2, 1990. This stress test is based on the 5-day period, 7/31/1990 - 8/6/1990. We assume a fall in the S&P 500 of −5.94%, an increase in oil prices of 38.79%, and a VIX of 35.91. All returns for all other securities are based on their current estimated correlation to these benchmarks.
∆SPX = −5.94%
∆CL1 = 38.79%
VIX = 35.91
Flight to Quality
∆SPX = −10%
∆GOLD = 5%
EURUSD = −5%
Flight to Euro
∆SPX = −10%
∆GOLD = 5%
EURUSD = 5%
Market -10%
In this stress test, the market index (default is S&P 500) undergoes a return of -10%. All returns for all other securities are based on their current estimated correlation to these benchmarks.
Quant Meltdown
Our quant meltdown is based on the 4-day period from 8/14/2007 – 8/17/2007. The scenario involves simultaneous shocks to three style factors:
∆Momentum Excess = -1.59%
∆Small Cap - Large Cap = 1.59%
∆Value - Growth = 2.60%
These represent approximately 5, 3, and 6 standard deviation moves for the momentum, small-large, and value–growth factors, respectively.
VIX(2x)
In this stress test, the VIX doubles from its current level. For example, if the VIX is currently at 15%, it goes to 30%. All returns for all other securities are based on their current estimated correlation to the VIX.
Simple Stress Tests
Emerging Markets Currency Crisis
This scenario is loosely based on the Asia Financial Crisis in 1997 and 1998. During the first week of 1998, an equally weighted basket of THB, MYR, IDR and PHP depreciated in value by 33% and an equally weight basket of the Thai SET, FTSE Bursa, NIFTY 50 and Hang Seng fell by 8.5%. In our scenario, all emerging market currencies fall in value by 25% and all emerging market equities fall by 10%.
Jan 2009 Interest Rise
Based on the 30 day period, 12/30/2008 - 1/29/2009. This is a simple stress test. The stress test assumes a Cox-Ingersoll-Ross adjustment, with interest rates increasing in proportion to the square root of their current level.
1-year: 0.11% to 0.47%
5-year: 1.50% to 2.39%
10-year: 2.75% to 3.70%
30-year: 2.93% to 3.39%
Rates +1%
In this stress test all interest rates increase by 1%, a parallel shift. This is a simple stress test. Fixed income instruments and interest rate derivatives are directly impacted. Futures and options in other asset classes are repriced using the new discount factors, but are otherwise unaffected.
Sideways Market 30
This scenario is designed for portfolios with significant option positions. In this scenario, we see what would happen to the value of a portfolio if the market moved sideways for 30 calendar days, that is assuming time to expiration decreased by 30 days, but the price of all underlying securities, and all other option parameters remained constant. For options that would expire within 30 days, we assume that price moves to the current intrinsic value. There is no P&L from non-option securities, and there is no P&L from interest or dividends.
Taper Tantrum
This scenario is based on 6/18/2013 to 6/25/2013. For fixed income we assume a Cox-Ingersoll-Ross adjustment, with interest rates increasing in proportion to the square root of their current level.
1-year: 0.12% to 0.14%
5-year: 1.06% to 1.49%
30-year: 3.34% to 3.63%
Volatility Surface Stress Tests
This family of stress tests alllows users to change the shape of the implied volatility smile, the volatility term structure or both. For more information see our white paper, Volatility Surface Stress Tests
Worst Day Stress Test
This stress test indicates the worst day that the current portfolio would have experienced over the past 500 days. For securities that age (options, bonds, futures, etc.), constant maturity series are used (maturity, moneyness, etc. are constant).
The standard version of this stress tests indicates the worst single day over the past 500 days. 2, 5, 10 and 20 day versions of this stress test are available, which show the worst performance of the current portfolio over a number of consecutive days.
OPERA Stress Tests
Northstar also provides several standard Open Protocol Enabling Risk Aggregation (OPERA) stress tests. See Stress Tests (OPERA).