HAC standard errors explained: Newey-West procedure (Excel)
HTML-код
- Опубликовано: 30 июл 2024
- Heteroskedasticity and autocorrelation consistent standard errors (HAC) have become a staple in time series econometrics since their development by Newey and West (1987). Today we are investigating the implementation of the HAC covariance matrix, the mathematics behind it, and its optimal specification in Excel based on a simple 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
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
THANK YOU, this is the clearest explanation I have seen so far!!!!
Very useful. Thank you!
Hey man great!!!! This is a great topic in econometrics but i could not find any great excel resources on it. This is a great addition to my learning!!
Hi NEDL, briliant content - just have a question, if i were to model a GARCH (1,1) model of the residual of a regression and assess HAC for the params, does the alpha and beta term from the GARCH model fall under the design matrix and can i use the same process as in the video to evaluation GARCH params? Thanks
Very helpful, thank you.
you saved my life
Great video!! But I have a little confusion. In the W formula, there is a term [T/(T-k)] multiplied to the error matrix and max function, but I did not see you multiply that [T/(T-k)] term. Thank you for this video btw
He talks about it @14:38. He did not put it to make the result more intuitive, guess wouldnt change much
You are amazing thank you very much
Is there a way to find out corrected f-statistic through this. As t sq follows f distribution, we can calculate f-statistic through the t-statistic in simple regression model. Is there a way to do so in multiple regression as well?
I ussually deal with Regression analysis tools and this is a bit confusing, can you show me how to use regression to correct the serial correlation in my data? I have DW less than 2 meaning that the present of serial correlation exists!
Hello! Thank you for the amazing video! Can you tell me a paper or similar source that explains the exact formula you use for "w" in your video?
Hi, and glad you enjoyed the video! The "W" is a weight matrix which is quite universal across all robust standard error estimators. This particular one comes from Newey and West (1987), and here is the non-paywalled working paper PDF: www.nber.org/system/files/working_papers/t0055/t0055.pdf
Hi I really liked your video. I have to write about this topic for the university. Therefore, could you name me the source of the algorithm to calculate the weights? With the max() expression.
Hi, and thanks for the question! The source for this approach is Newey and West (1987) - one of the most heavily cited papers in econometrics.
hi. plz, make a video on volatility and higher-order moments timing using mutual fund example.
Hi Ghulam, and thanks for the suggestion! I have got several video on higher-order moments and their application to investment management, for example here I discuss MVaR for performance evaluation (ruclips.net/video/qvQ4gUiC1yU/видео.html), and here I show the impact of skewness and kurtosis on investor utility (ruclips.net/video/skmYLg7vk3g/видео.html).