We might be thinking holding a security overnight, especially a highly volatile security like XIV, bears substantial risk. In a world full of uncertainties we are wondering is it worth holding XIV overnight. I have conducted a simple test since the inception of XIV to compute the distribution of the intraday and overnight percentage returns for XIV. The intraday return is simply the (daily close/daily open) -1 while the overnight return is (daily open/previous close) -1.

The intraday return is highly skewed with heavy left tails. And the mean is 0.07684657 which means if you buy at open and sell at close you are expected to make 7.68 USD per 10K investment. XIV has been in existence for 710 trading days and you could make 5452.8 USD if you gamble with 10000 dollars every day without assuming any cost and with perfect execution.

With a typical trading cost ~$4/trade if you only have 10K to start with basically you are only making your brokerage firm happy.

The overnight return is more symmetric. And the mean is 0.2469942 which means if you buy at open and sell at close you are expected to make 24.7 USD per 10K investment. XIV has been in existence for 710 trading days and you could make 17537 USD if you gamble with 10000 dollars every day without assuming any cost and with perfect execution. You are more than doubled.

With a typical trading cost ~$4/trade if you only have 10K to start with you can almost double your fortune and beat the SP500 index handily.

The overnight gamble has a better return profile. And the reason is simple: the overnight roll yield. There are ups and downs associated with the changes of VIX or market fears but the law of large number and long term mean-reversion of VIX makes the distribution resemble a Gaussian distribution. Eventually the roll yields give you a nice positive mean return. Still, we need to remember that holding XIV overnight is not free money, you are facing total destruction nevertheless.

I will explore the overnight/intraday histogram under different market conditions, say VIX level and backwardation/contago, in future studies.

We might be thinking holding a security overnight, especially a highly volatile security like XIV, bears substantial risk. In a world full of uncertainties we are wondering is it worth holding XIV overnight. I have conducted a simple test since the inception of XIV to compute the distribution of the intraday and overnight percentage returns for XIV. The intraday return is simply the (daily close/daily open) -1 while the overnight return is (daily open/previous close) -1.

The intraday return is highly skewed with heavy left tails. And the mean is 0.07684657 which means if you buy at open and sell at close you are expected to make 7.68 USD per 10K investment. XIV has been in existence for 710 trading days and you could make 5452.8 USD if you gamble with 10000 dollars every day without assuming any cost and with perfect execution.

With a typical trading cost ~$4/trade if you only have 10K to start with basically you are only making your brokerage firm happy.

The overnight return is more symmetric. And the mean is 0.2469942 which means if you buy at open and sell at close you are expected to make 24.7 USD per 10K investment. XIV has been in existence for 710 trading days and you could make 17537 USD if you gamble with 10000 dollars every day without assuming any cost and with perfect execution. You are more than doubled.

With a typical trading cost ~$4/trade if you only have 10K to start with you can almost double your fortune and beat the SP500 index handily.

The overnight gamble has a better return profile. And the reason is simple: the overnight roll yield. There are ups and downs associated with the changes of VIX or market fears but the law of large number and long term mean-reversion of VIX makes the distribution resemble a Gaussian distribution. Eventually the roll yields give you a nice positive mean return. Still, we need to remember that holding XIV overnight is not free money, you are facing total destruction nevertheless.

I will explore the overnight/intraday histogram under different market conditions, say VIX level and backwardation/contago, in future studies.