EVIEWS AR forecasting

แชร์
ฝัง
  • เผยแพร่เมื่อ 8 พ.ย. 2024

ความคิดเห็น • 33

  • @niibarnor4209
    @niibarnor4209 9 ปีที่แล้ว

    Thanks Prof. Becker. I av learnt a lot from your presentation.

  • @ojokwilfred8327
    @ojokwilfred8327 10 ปีที่แล้ว

    Thanks prof, I have learnt some thing from your presentation

  • @farabiahmedulah4496
    @farabiahmedulah4496 8 ปีที่แล้ว +1

    great video .....helped me out a for my thesis on hyperbolic discounting :)

  • @kamrulhassansunon5438
    @kamrulhassansunon5438 10 ปีที่แล้ว +2

    Dear Sir. Thanks a lot for the videos. I'm an MSc student in MBS. I find your videos very helpful since I have very little idea about EViews. Sir I want to know one thing that is if we have out-layers in the sample that we want to omit how can we do it? I mean if the out-layers are in the middle of the sample data, not in the beginning or ending? Thanks in advance.

  • @tzett0011
    @tzett0011 9 ปีที่แล้ว +2

    thx a lot for the great video Mr. Becker,
    is there a way to keep the length of the estimation period constant, so that eviews would always use the 10 latest datapoints for the estimation?

    • @RalfBecker
      @RalfBecker  9 ปีที่แล้ว

      JohnDot yes there is, you will have to learn how to use the in-build programming language of EVIEWS. Look at the document you can download from the "Customising EVIEWS" section on this page eclr.humanities.manchester.ac.uk/index.php/EVIEWS for a start.

    • @tzett0011
      @tzett0011 9 ปีที่แล้ว

      Ralf Becker thank you very much for your help

  • @elsiddigobaid
    @elsiddigobaid 9 ปีที่แล้ว

    Thanks Prof. Is it possible to use equations created by the VAR system for forecasting? That's to say, first estimate the VAR then we go to "Proc" and select "make model" then we get the equations and subsequently we use these equations for forecasting.

  • @czjcastillo
    @czjcastillo 8 ปีที่แล้ว

    Great tutorial! May I ask, if you used quarterly data adjusted for seasonality and differeced, do you reseasonalize the forecast and undo the differencing?

  • @kcaron01
    @kcaron01 10 ปีที่แล้ว

    Ralf, these are great videos. I'm really wondering how to populate a forecast into the future. I understand breaking the full sample into two parts and using the early part to build the model, then test it on the later part of the sample. What is making me nuts is that I can't figure out how to populate EViews fields from, lets say NOW, out through the next 3 to 6 months or so. Static seems to work better and I can get it to go forward one month, but then I want to go beyond that and can't seem to figure that part out... Any help you could offer would be greatly appreciated! Thanks for the great videos!

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      Remember, to produce the static forecasts EVIEWS will attempt to use the actual observations as you go forward. Say you have quarterly data and the last available observation is Q3 in 2014. Now you can use static to forecast Q4 in 2014, but if you attempt to forecast Q1 2015 EVIEWS will attempt to use the observation for Q4 2014. But of course, that is not available. This is why the static forecast mode will only work one step into the future.

  • @jackieurrutia1789
    @jackieurrutia1789 6 ปีที่แล้ว +1

    hi sir. i just want to ask if y c y(-1 to -4) is the same as y c ar(1) ar(2) ar(3) ar(4) in eviews. thanks a lot

  • @fateneclarke3173
    @fateneclarke3173 9 ปีที่แล้ว

    Thank you for this very useful video. I need help with value at risk based garch model. I know the theory but I am wondering how to compute it in Eviews 6.
    your advice will be helpful.

  • @elninou3174
    @elninou3174 10 ปีที่แล้ว

    Dear Ralph, first off thank you for this great video, very useful and very well explained. Then, I'd like to know what is the difference between uk_cpi and ps_uk_cpi?

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      Marc Jack pc_uk_cpi is the percentage change of uk_cpi, i.e. pc_uk_cpi = (uk_cpi-uk_cpi(-1))/uk_cpi(-1)

  • @rosianakurniawan910
    @rosianakurniawan910 8 ปีที่แล้ว

    how to manually calculate the beta coefficient (intercept and the autoregressive coefficients)? by what formula is that to get those beta value sir? thank you

  • @Bklyn2005
    @Bklyn2005 10 ปีที่แล้ว

    Thanks for a good learning video. I have one question where do you get Y(-1 to -4). is it because you have 4 quarters And is it always going to be negative when you are forecasting. Thanks Ralf

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      Bklyn2005 the -1 to -4 indicate that we are using lags of the dependent variable. It is always going to be like it when you build AR models.What order you should use depends on the data. But if yo have quarterly data you should certainly consider 4 lasgs at least. But check this clip for how to use information criteria to determine the optimal lag: th-cam.com/video/PmXaLTLv_rE/w-d-xo.html

  • @AMF2EG
    @AMF2EG 10 ปีที่แล้ว

    Hi Ralph, thx for the videos, please guide how to check for serial correlation and heteroskedasticity for Panel data in Eviews ? thanks again

  • @adebayoyinka4295
    @adebayoyinka4295 9 ปีที่แล้ว

    thanks a lot for this video

  • @shawnrobert1619
    @shawnrobert1619 10 ปีที่แล้ว

    Dear Prof. Becker, First of all, thank you for your wonderful tutorial video. I have a question for you. I tried to use the static forecast for forecast 4 quarters ahead, but only got one quarter forecast instead. Can you help me with this. My data set is from 1951Q1 to 1978Q2 and I tried to forecast 1978Q3 to 1979Q2, but only had forecast of 1978Q3 data. But if I choose dynamic forecast, there will be no problem to get the forecast value of 1978Q3 to 1979Q2. Can you help me with this? Thank you.

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      If you only have data till 78Q2 then you can use the static forecast only to forecast one quarter ahead. You would need an observation in 78Q3 to produce a static forecast for 78Q4 (see min 7:00 onwards). I should note that I don't find the EVIEWS terminology overly intuitive, but it is what it is.

    • @shawnrobert1619
      @shawnrobert1619 10 ปีที่แล้ว

      Ralf Becker
      Prof. Becker, Thanks! I am not an econometrican, but I am a bit of confused by Eviews. Do you know where I could get familiar with writing a loop to do the static forecast in Eviews? Since I have lots of these equations to conduct static forecast.

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      Shawn Robert You may want to check the file available in the "Customising EVIEWS" Section on eclr.humanities.manchester.ac.uk/index.php/EVIEWS.

    • @shawnrobert1619
      @shawnrobert1619 10 ปีที่แล้ว

      Ralf Becker
      Thanks and I'll try it.

  • @tayloryakuza
    @tayloryakuza 7 ปีที่แล้ว

    what the relation between the signification of coefficient(-4) and saisonality?

  • @avrammadalina28
    @avrammadalina28 7 ปีที่แล้ว

    hello, there is any posibility to write different values for range and sample? For example: the range to be monthly and the sample quarterly.

    • @RalfBecker
      @RalfBecker  7 ปีที่แล้ว

      i am not sure that I understand your question. If your data are quarterly, it doesn't really make sense to have a sample that is down to the month. But I may misunderstand your query.

  • @alimarseet8533
    @alimarseet8533 10 ปีที่แล้ว

    hi
    when estimate forecast which data should be used (stationary or non-stationary data)?

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว

      This really depends on what you want to do and where the non-stationarity comes from. A good reading starting point may be this: mitpress.mit.edu/books/forecasting-non-stationary-economic-time-series

  • @vengeefulspirit
    @vengeefulspirit 9 ปีที่แล้ว

    why DW is so low?

  • @vuang3088
    @vuang3088 10 ปีที่แล้ว +1

    this is a model regression and it isn't a model ar :)).
    model ar : y c ar(1 to 4)
    :)).

    • @RalfBecker
      @RalfBecker  10 ปีที่แล้ว +1

      Dear Vu, if you estimate Model1: y c ar(1 to 4) and Model2: y c y(-1 to -4) you are really estimating the same model. If you estimate both you can see that the AR coefficients are identical as are the regression statistics (like the log-likelihood function or the information criteria). The only thing that is different is the constant. You will find that the constant from Model2 is equal to the constant from Model 1 * (1 - sum of all AR coefficients).
      Thios is best explained in Ben Vogelvang's book, Econometrics, Theory and Applications with EVIEWS, pages 346-347.