Thank you for this video. I have a question: what about Chow test under heteroschedasticity in Stata? How can I perform the test if my data show heteroschedasticity?
This video helps me a lot to understand the logic of testing! By the way, my data is unbalanced panel data with 10 columns of dummy variables that do not relate among themselves. I want to test the structural break model. My question is ‘for the if condition, which syntax do I have to use? AND(&) or OR (I) for all 10 dummy columns’ could you please suggest to me? I really appreciate any help you can provide.
Thank you so much @Mike Jonas Econometrics. Would you say I can apply your explanation to a regression where I need to split the sample into two types of sectors? I mean, firms in sensitive and non-sensitive industries.
I am using the xtbreak command to use the Bai-Perron test for multiple structural breaks in the data of granted patents in the US. I expected one in the 1980's because the patent law was changed and maintanance fees were introduced. But in the end I find 5 structural breaks which happens to be also the maximum amount of breaks the test can detect. Does this sound like useable results?
To piggy back off this question. I'm trying to see if increased state funding for k-12 education in Illinois (after the passage of new funding legislation in 2017) is going to its intended purpose (instruction and services for students) or other purposes (e.g. debt service). My thought was to use a structural break test to see if the relationship between state funding (independent variable) and various expenditure funds (running separate models for each expenditure fund as a dependent variable) is meaningfully different before and after 2017. For example, does the state funding coefficient for debt service funds increase/decrease meaningfully before and after 2017. I too have panel data. I was thinking of following your method (creating a break variable and an interaction effect between the break variable and state funding) and running a time fixed-effects model in order to account for variation in time and place. 1) In your opinion, is this an appropriate use of the chow test and 2) is this how one could go about applying the chow test in panel data.
@@mikejonaseconometrics1886 In time series content, many videos explain how to detect structural breaks and how to test for unit roots with this issue, but it's still hard to find somewhere explaining how to consider this information within our estimations, for example, in a VAR model...
