Panel data: impossible insufficient observations

[stijn schouten]

Dear all,

I am doing research regarding on the effect of m&a league tables and deal characteristics in what I believe to be panel data.
I have deals, classified by unique DealPermID and the time being :Dateannounced.
Often, i have two or more advisors on one deal, they have the same dealpremID so the same ID. In this case, some DealpermID will have two or more rows, consisting of the full financial data and the data of the individual advisors.
I can make an unbalanced paneldata set, $t is daily based on dealannounced and $id is based on the dealpermID. when i try to use the command xtreg: i get the error: insufficient observations.

I do not understand why i do not have sufficient observations, i have over 1100 individual deals over 1997 to 2021.

On the other hand, every observation only exists once, however i have more id's as some id's add information regarding the 2nd or 3third advisor. (one row includes financials and one advisor, other rows only include advisor data)
so if i were to declare it as time series, i get: repeated time values in sample as there can be two deals on the same date.


I would like to know how why i get these errors, if it is panel or times series and how i can start with regressions on this dataset without getting the errors.

[Carlo Lazzaro]

Stijn:
if -DealPermID- are all different (something that cannot be doubke-checked if you post a screenshot instead of a excerpt/example of your dataset via -dataex-), -xtreg- is nit supposed to work, as you do not have a panel dataset.

[stijn schouten]

Carlo Lazzaro correct, DealPermID are all different numbers.
Thanks for your response! Therefore I should directly link the advisors to the deal number thus making it a cross sectional dataset instead of a panel data set right?
next stept is to figure out how I will do this effeciently

[Carlo Lazzaro]

Stijn:
if you actually have a cross-sectional dataset with a continuous regressand, you should go -regress-.