Have you tried the Stata command "threg" to "fit threshold regression model"?

Sorry sir i have not tried as i'm a new user in STATA. i'll try for that

It seems to me (with more than 95% confidence level) that the Stata command "threg" DOES NOT perform Hansen's threshold regression. It seems to be used for survival analysis, not the common (non)linear regression. Please visit Hansen's website for Gauss, Matlab and R codes. http://www.ssc.wisc.edu/~bhansen/progs/progs_threshold.html

I agree with Huang that "threg" command does not work.I have myself tried it many times but it does not work. To my knowledge, research use STATA to perform Hansen's threshold regression but it is not clear how do they develop STATA command for this purpose.

Professor Hansen now provides a preliminary Stata code for implementing the threshold regression model. Please see: http://www.ssc.wisc.edu/~bhansen/progs/ecnmt_00.html

I have tried lot of time in stata. plz let me know the proper code with example of Threshold Regression.. It will be my humble request to you. Thanks

hi, when i use the command "xthreg" for Hansen , the estimation parameters is in only in one regime !!!!!

But the estimation  parameters should be in two regimes

Use the STATA command "thresholdreg" to estimate the 2 regime parameters after "thresholdtest" to test for the presence of a threshold.

Thank you Muazu Ibrahim, yes i am resolving this problem

I want to find the Budget deficit Threshold on growth,

I've tried to use "thresholdtest" command to test for the presence of a threshold, and "thresholdreg" command to estimate the 2 regime parameters (by hansen 1996) but the result is like this :

. thresholdtest rgdpr bd lner inflasi m2, q(bd) trim_per(0.15) rep(2000)

Test of Null of No Threshold Against Alternative of Threshold

Allowing Heteroskedastic Errors (White Corrected)

______________________________________________________________________

Number of Bootstrap Replications: 2000

Trimming Percentage: .15

Threshold Estimate: .51935041

LM-test for no threshold: .

Bootstrap P-Value: 1

______________________________________________________________________

.

. thresholdreg rgdpr bd lner, q(bd) h(1)

Global OLS Estimation, Without Threshold

______________________________________________________________________

Heteroskedasticity Correction Used

______________________________________________________________________

Independent Variables Estimate St Error

______________________________________________________________________

Intercept . .

bd . .

lner . .

Observations: 53

Degrees of Freedom: 50

Sum of Squared Errors: .

Residual Variance: .

R-squared: .

Heteroskedasticity Test (P-Value): .

Threshold Estimation

______________________________________________________________________

Threshold Estimate: -.227700159

.95 Confidence Iterval: [-.227700159,.]

Sum of Squared Errors .

Residual Variance: .

Joint R-Squared: .

Heteroskedasticity Test (p-value):.

______________________________________________________________________

Regime1 q-.227700159

Parameter Estimates

______________________________________________________________________

Independent Variables Estimate St Error

______________________________________________________________________

Intercept . .

bd . .

lner . .

.95 Confidence Regions for Parameters.

Independent Variables Low High

______________________________________________________________________

Intercept . .

bd . .

lner . .

Observations: 52

Degrees of Freedom: 49

Sum of Squared Errors: .

Residual Variance: .

R-squared: .

.

Can someone help why is it happened