eventstudy.Single.constant_mean

classmethod Single.constant_mean(security_ticker, event_date: numpy.datetime64, event_window: tuple = (-10, 10), estimation_size: int = 300, buffer_size: int = 30, keep_model: bool = False, **kwargs)

Modelise returns with the constant mean model.

Parameters

• security_ticker (str) – Ticker of the security (e.g. company stock) as given in the returns imported.

• event_date (np.datetime64) – Date of the event in numpy.datetime64 format.

• event_window (tuple, optional) – Event window specification (T2,T3), by default (-10, +10). A tuple of two integers, representing the start and the end of the event window. Classically, the event-window starts before the event and ends after the event. For example, event_window = (-2,+20) means that the event-period starts 2 periods before the event and ends 20 periods after.

• estimation_size (int, optional) – Size of the estimation for the modelisation of returns [T0,T1], by default 300

• buffer_size (int, optional) – Size of the buffer window [T1,T2], by default 30

• keep_model (bool, optional) – If true the model used to compute the event study will be stored in memory. It will be accessible through the class attributes eventstudy.Single.model, by default False

• **kwargs – Additional keywords have no effect but might be accepted to avoid freezing if there are not needed parameters specified. For example, if market_ticker is specified.

See also

market_model(), Single.FamaFrench_3factor()

Example

Run an event study for the Apple company for the announcement of the first iphone, based on the constant mean model.

>>> event = EventStudy.constant_mean(
...     security_ticker = 'AAPL',
...     event_date = np.datetime64('2007-01-09'),
...     event_window = (-5,+20)
... )