Early assignment for SPX?

In the context of options trading, early assignment refers to the situation where the option holder exercises their option before its expiration date. This typically happens when the option is in-the-money, meaning the underlying stock's price is beyond the strike price of the option. In such cases, the QuantConnect backtesting engine simulates an assignment event to reflect this real-world scenario (source).

However, for index options like SPX, early assignment doesn't quite apply in the same way. This is because index options are cash-settled, meaning there are no actual shares to be assigned. Instead, the holder of an in-the-money index option will receive a cash payment. The error message you're seeing suggests that the backtesting engine is trying to simulate an assignment event for an index option, which isn't realistic (source).

To avoid this issue, you can disable early assignment in your backtests by setting the option assignment model to the NullOptionAssignmentModel. Here's how you can do this in Python:

security.set_option_assignment_model(NullOptionAssignmentModel())

(source)

Please note that while this will prevent the simulated early assignment from occurring, it may not fully reflect the real-world behavior of index options. It's always a good idea to thoroughly debug your algorithm and understand the implications of the models you're using.

I think the main issue is that the other leg of the spread (long put) doesn't seem to be exercised automatically when early assignment happens. Is this a bug or expected behavior?

If early assignment is disabled, both options are exercised correctly, and there is no unexpected dip in the account.

In QuantConnect, when an option is in-the-money, the backtesting engine will trigger an assignment event to simulate the real-world scenario where the option holder may choose to exercise the option. However, this automatic exercise/assignment only applies to in-the-money options. If the other leg of your spread (the long put) is not in-the-money, it won't be automatically exercised (source).

If you want to exercise the other leg of the spread when early assignment happens, you can do so manually in the OnOrderEvent method. Here's an example of how you can do this in Python:

def OnOrderEvent(self, orderEvent):
    order = self.Transactions.GetOrderById(orderEvent.OrderId)
    if order.Type == OrderType.OptionExercise:
        # short options get assigned
        if orderEvent.IsAssignment:
            # do stuff
            pass
        # long options were exercised
        else:
            # do stuff
            pass

(source)

In this code, orderEvent.IsAssignment will be True if the order event is an assignment (i.e., one of your short options got assigned). You can then add your logic to exercise the other leg of the spread in the # do stuff section.

Please note that exercising an option manually will require sufficient buying power in your account. If you're seeing insufficient fund errors, it could be because your account doesn't have enough buying power to exercise the option. You may need to adjust your algorithm or add more funds to your account to avoid these errors.

In this case both options are in the money. But the other option is not getting exercised even on expiration (if early assignment simulation is enabled).

Another confusing thing is that the time of the early assignment is 12:00 on expiration day. How is that even possible? These options should be settled after hours, so what exactly does this time represent?

confusing time of early assignment

After disabling early assignment simulation, both options are exercised as expected:

both options are exercised 

However, it's also not clear why does it say 01:00 as the exercise time? Should these options be settled after hours on expiration day, so should this time be “2022-01-24 16:00:00” or something like that? 

The time of the early assignment you're seeing (12:00) is likely due to the way QuantConnect's backtesting engine simulates option assignments. The engine doesn't necessarily reflect the exact timing of real-world option assignments, which can occur at any time before expiration for American style options. The simulated assignment time is not meant to represent the actual time an assignment would occur in live trading (source).

As for the other leg of the spread not getting exercised, this could be due to the way your algorithm is set up. If you want to ensure that both legs of the spread get exercised, you may need to manually exercise the other leg in the OnOrderEvent method, as I mentioned in the previous response.

Regarding the exercise time of 01:00, it's important to note that the S&P 500 Index Option contracts market trades in the America/Chicago time zone (source). This could explain the discrepancy you're seeing in the exercise time. However, without more specific information about your algorithm and its settings, it's hard to provide a more detailed explanation.

If you continue to experience issues, I recommend posting a more detailed question on the QuantConnect forum where the community and the QuantConnect team can provide more targeted assistance.

I have a backtest attached in my original message. To summarize, here are my questions that I hope someone can answer after reviewing the code:

1. With early assignment simulation enabled (by default), why is the long put option not exercised at expiration, even though it's in the money?
2. What does the early assignment time “2022-01-24 12:00:00” mean for the option that expires on 01-24. I don't think options can be settled in the middle of the day like this?
3. After disabling early assignment simulation, both options are exercised, however the exercise time shown is “2022-01-25 01:00:00”. Should it instead show the time at which SPX closing price was calculated like “2022-01-24 16:00:00”?

1. The issue with the long put option not being exercised at expiration, even though it's in the money, could be due to the way QuantConnect's backtesting engine simulates option assignments. The engine doesn't necessarily reflect the exact timing of real-world option assignments. If you want to ensure that both legs of the spread get exercised, you may need to manually exercise the other leg in the OnOrderEvent method (source).

2. The early assignment time “2022-01-24 12:00:00” is likely due to the way QuantConnect's backtesting engine simulates option assignments. The simulated assignment time is not meant to represent the actual time an assignment would occur in live trading (source).

3. The exercise time shown as “2022-01-25 01:00:00” could be related to the time zone of the S&P 500 Index Option contracts market, which trades in the America/Chicago time zone (source). However, without more specific information about your algorithm and its settings, it's hard to provide a more detailed explanation.

If you continue to experience issues, I recommend posting a more detailed question on the QuantConnect forum where the community and the QuantConnect team can provide more targeted assistance.