Probit Regression

[Theocharis Iosifidis]

Hi to everyone in the forum!

I am an MSc student in Financial Economics and I would like to raise a question. Any help would be much appreciated
I am running an ordered probit model in stata for my thesis in which my dependent variable takes three values denoted: 1,2,3 which represent the method of payment in mergers and acquisitions i.e. cash only payments, stock only payments, hybrid cash-stock payments.
While the main regression and the marginal fixed effects run perfectly fine, below the marginal effecs table i get an asterisk that says: dy/dx is for discrete change of dummy variable from 0 to 1. Does this indicate that there is a mistake in my calculations? Because I do not have a binary model but rather three values for my dependent variable and thus I cannot understand why it says 0 to 1. Could anyone please provide a clarification for that?
Thanks in advance,

[Felix Bittmann]

I am not an economist but are you sure the outcomes are ordered? As a robustness check I would consider a multinomial regression. Regarding the message, could you post the command you are using?

[Joro Kolev]

Felix is right, Theocharis. For your situation there is no natural ordering between the 3 methods of payment, therefore multinomial probit is the correct modelling strategy.

And the message that you are getting refers to the regressors/explanatory variables. Not to the regressand/outcome variable/explained variable.

Hi to everyone in the forum!

I am an MSc student in Financial Economics and I would like to raise a question. Any help would be much appreciated
I am running an ordered probit model in stata for my thesis in which my dependent variable takes three values denoted: 1,2,3 which represent the method of payment in mergers and acquisitions i.e. cash only payments, stock only payments, hybrid cash-stock payments.
While the main regression and the marginal fixed effects run perfectly fine, below the marginal effecs table i get an asterisk that says: dy/dx is for discrete change of dummy variable from 0 to 1. Does this indicate that there is a mistake in my calculations? Because I do not have a binary model but rather three values for my dependent variable and thus I cannot understand why it says 0 to 1. Could anyone please provide a clarification for that?
Thanks in advance,

[Theocharis Iosifidis]

Felix is right, Theocharis. For your situation there is no natural ordering between the 3 methods of payment, therefore multinomial probit is the correct modelling strategy.

And the message that you are getting refers to the regressors/explanatory variables. Not to the regressand/outcome variable/explained variable.

. oprobit PaymentMethod CashHolding Control COLLATERAL Unemployment GDPPERCAPITAGROWTH AssetGrowth lnM1 CrossIndustry CrossBorder lnDV, vce (robust). This is the command I am using. I alligned to the payment method the values 1,2,3 which correspond to cash only acquisitions, hybrid acquisitions, stock only acquisitions respectively. The methodology that I use is ordered probit which in contrast with a binary probit, gives me the ability to assign more than 2 values to the dependent variable. Why should I use a multinomial regression and not the ordered probit? Also, the problem with multinomial probit is that it produces a base category for which it provides me with no coefficients, something that I need to avoid since I need coefficients for all three cases.
Thanks for your advice!

[Theocharis Iosifidis]

I am not an economist but are you sure the outcomes are ordered? As a robustness check I would consider a multinomial regression. Regarding the message, could you post the command you are using?

. oprobit PaymentMethod CashHolding Control COLLATERAL Unemployment GDPPERCAPITAGROWTH AssetGrowth lnM1 CrossIndustry CrossBorder lnDV, vce (robust). This is the command I am using. I alligned to the payment method the values 1,2,3 which correspond to cash only acquisitions, hybrid acquisitions, stock only acquisitions respectively. The methodology that I use is ordered probit which in contrast with a binary probit, gives me the ability to assign more than 2 values to the dependent variable. Why should I use a multinomial regression and not the ordered probit? Also, the problem with multinomial probit is that it produces a base category for which it provides me with no coefficients, something that I need to avoid since I need coefficients for all three cases.
Thanks for your advice!

[Joro Kolev]

You should not use -oprobit- because there is no sense in which your 3 categories are ordered.

-oprobit- would have been appropriate if your three categories were "cash", "more cash", "even more cash".

There is not sense in which "stock acquisitions" is bigger/smaller, faster/slower, more pretty/less pretty, etc. that "cash acquisitions". Your three categories do not stand in an ordered relationship to one another.

[Theocharis Iosifidis]

You should not use -oprobit- because there is no sense in which your 3 categories are ordered.

-oprobit- would have been appropriate if your three categories were "cash", "more cash", "even more cash".

There is not sense in which "stock acquisitions" is bigger/smaller, faster/slower, more pretty/less pretty, etc. that "cash acquisitions". Your three categories do not stand in an ordered relationship to one another.

https://www.youtube.com/watch?v=c9kvqeLFF8U this video does the exact same thing as I do. Has 3 categories and assign values from 1 to 3 on them and uses ordered probit regression. Could you please help me understand how what I do is not the same with that?
Thanks in advance!

[William Lisowski]

Within the first minute of the video, the author tells us the dependent variable is health status, with three categories, fair, good, and excellent. Clearly good is better than fair, and excellent is better than good, so these are ordered.

Joro asks in post #6, in what sense are your outcomes ordered? You have not addressed that question, and the video is not the exact same thing you do.