QQQQ With Collars

An interesting paper pointed out in a post by Mark Wolfinger on results of trading QQQQ with collars during different market conditions over the last ten years.

They modeled two methods, one a passive collar strategy and the other an active one.

Passive Strategies

On the Friday before expiration (on Saturday) sell calls on QQQQ that expire in a month and buy puts to collar the ETF. They tested using at the money options, 1%, 2%, 3%, 4%, and 5% out of the money options as well as 1 month, 3 month, and 6 month puts. So they tested 3 * 6 = 18 strategies here.

Active Strategies

The calls sold are all one month out and the puts bought are all six months out. But the number of calls is adjusted and the amount out of the money is adjusted as well. To do this, they look at three market signals, momentum, volatility and macroeconomic data, at three time horizons, short term, medium term, and long term.

Momentum Signal

To see if the market is trending up or down, they look at a simple moving average cross over. If the fast line is above the slow line, they take that as an up trend and if it is below, a down trend.

• Their short term signal is a 1 day simple moving average compared to a 50 day moving average.
• Their medium term signal is a 5 day simple moving average compared to a 150 day moving average.
• Their long term signal is a 1 day simple moving average compared to a 200 day moving average.

If the momentum signal is bullish, they widen the collar and move the call and the put one percent more out of the money. If bearish, they move them 1% closer to at the money.

Volatility Signal

Puts are always bought to match the number of shares of the ETF bought. Not so with the calls. If the market is in a state of “high anxiety” that is a bullish signal and so they want more exposure to the market. To do that, they sell only 0.75 calls per 100 shares. If the market is in a “low anxiety” state, that complacency is bearish, so they sell 1.25 calls per 100 shares.

They measure the anxiety level across the three time horizons.

• If the spot VIX is one standard deviation, or more, above the 50 day moving average of VIX we are in short term high anxiety state.
• If the spot VIX is one standard deviation or more below the 50 day moving average of VIX we are in short term low anxiety state.
• If the spot VIX is one standard deviation, or more, above the 150 day moving average of VIX we are in medium term high anxiety state.
• If the spot VIX is one standard deviation or more below the 150 day moving average of VIX we are in medium term low anxiety state.
• If the spot VIX is one standard deviation, or more, above the 250 day moving average of VIX we are in long term high anxiety state.
• If the spot VIX is one standard deviation or more below the 250 day moving average of VIX we are in long term low anxiety state.
• There is also a neutral zone, within one standard deviation of the moving average, where they sell 1 call per 100 shares.

Macroeconomic Signal

This is a two part process. If the NBER defines the time period as expansionary, new unemployment claims above the moving average is bullish is their claim. If we are in a recession, the opposite is true, new unemployment claims above the moving average is bearish.

• The short term signal is to compare the week’s unemployment claims to the 10 week moving average of unemployment claims.
• The medium term signal is to compare the week’s unemployment claims to the 30 week moving average of unemployment claims.
• The long term signal is to compare the week’s unemployment claims to the 40 week moving average of unemployment claims.

So if bullish, we want more upside exposure so shift the call and put to higher strikes. This shifts the collar upward. If we have a bearish signal, shift the call and put to lower strikes.

Trading Rules

Call % OTM = 2 + Momentum Signal + Macroeconomic Signal

Put % OTM = 3 + Momentum Signal – Macroeconomic Signal

Where the signals are +1 or -1 depending on the values described above. Also, puts are more expensive than calls so to get close to a zero cost collar, they start the put at 3% out of the money and the call 2% out of the money.

Number of Calls per Put = 1 + 0.25 * Volatility Signal

Tomorrow, I’ll discuss the results.

Here is a copy of the paper:

qqq_collar_study