Tobit model explained: censored regression (Excel)

Подписаться 26 тыс.

Просмотров 9 тыс.

50% 1

How to estimate a regression model when the dependent variable is truncated or censored? This is a common problem when estimating, for example, consumption or dividend payouts that cannot be negative. Tobit model, developed by James Tobin, is a go-to technique in this case, that uses modified maximum likelihood to tackle this issue. Today, we are learning how to apply the Tobit model in Excel using a real-world example.
Don't forget to subscribe to NEDL and give this video a thumbs up for more videos in Econometrics!
Please consider supporting NEDL on Patreon: / nedleducation

Опубликовано:

9 ноя 2021

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 5

@NEDLeducation 2 года назад

You can find the spreadsheets for this video and some additional materials here: drive.google.com/drive/folders/1sP40IW0p0w5IETCgo464uhDFfdyR6rh7 Please consider supporting NEDL on Patreon: www.patreon.com/NEDLeducation

@janhoern6118 Год назад

Hi, thank you so much! Your video helped me a lot to prepare for a presentation at university. Especially your additional materials! Incredible quality content what you do here!

@addisugebremichael5071 Год назад

I am really appreciate your clear explanation and wonderful lecture. but I need on technical efficiency issues of Tobit model (half normal and exponential distribution using Stata . Thanks in advance. Addisu Gebremichael from Ethiopia

@fanaaraf8663 Год назад

Hi! Thank you for your video. Similar to dividend yields, can we also say interest rates are truncated as well? In the case of p2p lending or investment platforms, although there is an average interest rate, it would never be less than zero or zero. Does the lower limit have to be zero? I have a loan data with interest rate as low as 5.5% and as high as 21%. Can I say this variable is truncated as well? I believe the interest can go a bit higher than the 21% I have in my sample.

@NEDLeducation Год назад

Hi Fanaa, and thanks for the excellent question! Generally, yes, lending rates for consumers are bound from below at zero, so you can use a lower-bounded Tobit model to explain these. However, as your "empirical" lower bound is 5.5%, which is quite a bit higher than the "theoretical" lower bound of 0%, the Tobit results will be quite similar to OLS I believe. Nevertheless, it would constitute a good robustness check. As for truncated from above, there is no reason interest rates have an upper limit (only if there does not exist a straight-up ban on charging higher than some rate on a platform for regulatory reasons).