Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Many thank Professor River Huang

    I still have a little question. That is:
    I have read some papers that employed PTR that exist 2 significant thresholds, they always showed their findings was Threshold 1 < Threshold 2 !
    But In this result, the Threshold 21 > Threshold 22

    I wonder could i interprete the result that Threshold 1 = Th-22 = 10.1109 / Threshold 2 = Th-21 = 10.1220 (because Fstat in Double Threshold Model > Fstat in Single Threshold Model) ? And Therefore, the Dependent-Regime coefficients (_#catc.OEAR 0,1,2 in 2 threshold model estimation will belong to regimes that lower, between, higher 10.1109 and 10.1220 respectively) ?

    With my highly appreciation.,
    Last edited by Le Trung; 15 Apr 2019, 21:46.

    Comment


    • #32
      dear Le Trung

      i hope you got some answers about how to interpret your results
      if so, i hope you can give me the main idea, because actually i face the same problem

      kind regards
      sedki

      Comment


      • #33
        Dear Le, The literature is right. In all cases, "Threshold 1" is the smaller estimated threshold value and "Threshold 2" is the larger one.

        Ho-Chuan (River) Huang
        Stata 17.0, MP(4)

        Comment


        • #34
          dear all

          please how to get the estimation before and after the peak as follow :
          i mean which command should be used

          Click image for larger version

Name:	sub regime.png
Views:	1
Size:	54.2 KB
ID:	1497812



          Best regards
          sedki

          Comment


          • #35
            Dear sedki, Try something like
            Code:
            xthreg y lnOPEN lmIND, rx(lnFD lnFE) qx(lnPGDP) thnum(1) trim(0.05) bs(100)
            Ho-Chuan (River) Huang
            Stata 17.0, MP(4)

            Comment


            • #36
              Dear River Huang

              thank you very much
              i would like to share you my work and i hope you can help me about some points

              actually, i study the non-linearity between financing constraints ( FC, measured by SA index) and CEO stock option remuneration , and i will take the FC variable as the threshold variable.

              i run the model i get the following outputs

              after test, i get single threshold
              Click image for larger version

Name:	thres.PNG
Views:	1
Size:	4.9 KB
ID:	1497951



              the regression gives me the following output:
              Click image for larger version

Name:	thre2.PNG
Views:	1
Size:	27.5 KB
ID:	1497952





              OK the threshold effect exists ( which means no linear relation between FC and Y) but how to interpret the result more ?? have i to focus on "cat#c SAindex" (0 and 1 ) coef ?


              best regards
              SEDKI

              Comment


              • #37
                Dear sedki, I believe that you need to read some applied papers (for explaining the results) of panel threshold model.
                Ho-Chuan (River) Huang
                Stata 17.0, MP(4)

                Comment


                • #38
                  Originally posted by River Huang View Post
                  Dear sedki, Try something like
                  Code:
                  xthreg y lnOPEN lmIND, rx(lnFD lnFE) qx(lnPGDP) thnum(1) trim(0.05) bs(100)
                  Dear Prof. River Huang

                  As i see The Prof's command above, it is applicable for case "Simultaneously Many Regime-Dependence Variables and one Threshold Variable". That mean we can upgrade to threshold models for many regime-dependence variables ? with this xthreg ?

                  Many thanks Prof !
                  Trung.,

                  Comment


                  • #39
                    Dear Le, What exactly is your problem? Are you asking that if -xthreg- allows for multiple regime-switching variables? If this is your problem, the answer is YES.
                    Ho-Chuan (River) Huang
                    Stata 17.0, MP(4)

                    Comment


                    • #40
                      Many thank Prof. River Huang

                      That's exactly my question. I will have to learn more for chosing and interpreting this case of PTR

                      Le Trung.,

                      Comment


                      • #41
                        Hi all,

                        are there any codes to conduct (non)linearity test as well to produce GIRFs?

                        Thanks,

                        Mateo

                        Comment


                        • #42
                          Hello Prof River Huang ,

                          I'm running a TR to estimate an optimal cash holding and I'm using the following code

                          Code:
                           xthreg ROA Cash Size_TA MB LEV, rx(Cash) qx(Cash) thnum(1) trim(0.1) grid(400) bs(300)
                          However, I always get the r(3200) error:

                          Could you offer some insights into this?

                          Thank you!

                          Comment


                          • #43
                            Dear Abdessamad, Please "help xthreg" and try the following:
                            Code:
                            xthreg ROA Size_TA MB LEV, rx(Cash) qx(Cash) thnum(1) trim(0.1) grid(400) bs(300)
                            That is to say, CASH can not be "regime-invariant" and "regime-variant" at the same time.
                            Ho-Chuan (River) Huang
                            Stata 17.0, MP(4)

                            Comment


                            • #44
                              [QUOTE=River Huang, Hello Professor kindly guide me about Threshold regression command for my variables. I want to check the threshold effect of GDP growth rate on Income inequality using panel data. Kindly also guide about the importance and usage of regime dependent variable which is in the syntax of Stata command for panel threshold regression. Sir please guide me, it will be very helpful for me. I shall be really grateful to you for this favor. Thanks.

                              Comment


                              • #45
                                Hello Dear Stata Community, kindly guide me about Threshold regression command for my variables. I want to check the threshold effect of GDP growth rate on Income inequality using panel data. Kindly also guide about the importance and usage of regime dependent variable which is in the syntax of Stata command for panel threshold regression. Please guide me, it will be very helpful for me. I shall be really grateful to you for this favour. Thanks.

                                Comment

                                Working...
                                X