Відео

屆時將遷入展城館！

Thanks!

Wow! Your message filled me with so much joy and gratitude! It's incredible to hear that the video had such a profound impact on your neighbor, your town, and you. Moments like these remind me why I create and share content. Thank you for sharing your inspiring story and for being a part of such a meaningful experience. Your heartfelt message means the world to me. Wishing you and your community continued happiness and inspiration.

C_t in step 4 is the C_t in step 1. In the first part of step 4, we are not trying to extract C_t, but we are trying to obtain an estimate for the coefficients b_s and b_y. Then, in the second part, we use C_t, b_y (from the first part), and Y_t to sort of back out F_t (the latent factors). The idea is that we are trying to determine the part of C that is not spanned by Y_t, such that we can identify the remaining part to be the "slow-moving part". You can refer to part II C (Estimation) of the paper for more information:) (faculty.wcas.northwestern.edu/lchrist/finc520/QJE.pdf)

Thank you very much. I am currently in the process of writing my master's thesis, and a significant portion of it is based on BBE (2005). I realiz now that I may have misunderstood step 3, particularly with regards to the distinction between PC and F^s_t. They are the same :) However, I found clarity on this matter about step 4 on page 405 of BBE III A. Your explanation aligns with what I discovered. Thanks! I am also interested in reading the paper you recommended, but unfortunately, the provided link is not functional. Could you kindly share the title of the paper instead?

@@wiztech2732 Good luck on your thesis! The link I shared is just BBE (2005) itself - faculty.wcas.northwestern.edu/lchrist/finc520/QJE.pdf (I think it should work now). Or you can read this - which provides a clear explanation regarding FAVAR modelling: www.diva-portal.org/smash/get/diva2:1215768/FULLTEXT01.pdf

It's the fed fund rate and it is given in the dataset with the name: FEDFUNDS

Thank you so much@@vincentlo812

S1-1 Row C最右邊

@@vincentlo812 谢谢

其實點transform好睇情況，呢度冇記錯嘅話就係就咁take ln 嘅first difference～ 另外平時transform做stationary (for regression )都要留意吓啲variables會唔會有機會Cointegrated ~如果有唔理Cointegration就咁take first difference 都係會有問題嘅～

Just a bit more on this: A vast number of well-regarded published papers on forecasting with multiple economic time series used different classes of VAR/SVAR, and these economic time series are all very likely to cointegrate, but still a VAR were used... To give just an example Kilian's Econometric of oil market even wrote: ...from a Bayesian point of view, inference remains valid whether the real price of oil is I(0) or I(1)... (where applying log-difference or not is not relevant for inference) You normally see VAR is used without too much concern of Cointegration (as you suggested) why is that the case? @@vincentlo812

Hello. You can get the data here: qed.econ.queensu.ca/jae/2019-v34.5/luvsannyam-batmunkh/

@@vincentlo812 Thanks a lot..

你是怎么听懂的

@@人荣肖 我不太明白你的问题

例如網上認識一位女子~

Thanks!

Didn't film that sadly :(

Yes (in theory) as shown in the video. But in practice identifying the direction of the bias is very complicated as there could be many confounding factors.

是！