Would a pre-trends test be appropriate?

[Christopher Ujma]

Hi there,

I am looking to examine the impact of a period of civil unrest in a region of South Africa in July 2021. Essentially, I am looking to conduct a difference in difference estimate - where the log of the total monthly export value in a region is regressed on a number of variables and the treatment variable which takes a value of 1 if the observation was recorded in a region of unrest (in this case, a province in South Africa called KwaZulu Natal) during the time of unrest (July 2021) and 0 otherwise. Essentially, I am looking to determine whether the province of the Western Cape is a good control group and counterfactual for how export values in the affected region would have evolved had the unrest not occurred.

I was wondering, would performing a pre-trends test for export values between KwaZulu Natal and the Western Cape be appropriate in this case? And how would one go about it?

My data looks as follows:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float lv str19 province float DD
10.160206 "Western Cape" 0
 6.687344 "Western Cape" 0
 8.906497 "Western Cape" 0
 7.114627 "Western Cape" 0
 7.398255 "Western Cape" 0
3.8665526 "Western Cape" 0
 3.911809 "Western Cape" 0
 4.745103 "Western Cape" 0
 6.958226 "Western Cape" 0
 4.273046 "Western Cape" 0
2.0654907 "Western Cape" 0
 6.580245 "Western Cape" 0
 8.545335 "Western Cape" 0
11.385674 "Western Cape" 0
 6.805698 "Western Cape" 0
 5.382561 "Western Cape" 0
  8.79986 "Western Cape" 0
 5.376102 "Western Cape" 0
 7.671155 "Western Cape" 0
 4.155391 "Western Cape" 0
 5.034706 "Western Cape" 0
 5.788817 "Western Cape" 0
 6.176671 "Western Cape" 0
 7.530075 "Western Cape" 0
 5.425347 "Western Cape" 0
  5.65393 "Western Cape" 0
  5.03806 "Western Cape" 0
 4.397544 "Western Cape" 0
 6.234399 "Western Cape" 0
4.1983843 "Western Cape" 0
 6.407596 "Western Cape" 0
 8.087784 "Western Cape" 0
 5.654188 "Western Cape" 0
 4.988049 "Western Cape" 0
 6.738582 "Western Cape" 0
 4.883191 "Western Cape" 0
 8.410487 "Western Cape" 0
  8.95392 "Western Cape" 0
 3.560178 "Western Cape" 0
 6.717937 "Western Cape" 0
 10.06614 "Western Cape" 0
3.5643625 "Western Cape" 0
 10.82003 "Western Cape" 0
 6.215573 "Western Cape" 0
  8.68768 "Western Cape" 0
 5.562564 "Western Cape" 0
 6.288291 "Western Cape" 0
 6.660422 "Western Cape" 0
 8.076735 "Western Cape" 0
 8.175714 "Western Cape" 0
 8.815947 "Western Cape" 0
 6.498086 "Western Cape" 0
 7.892373 "Western Cape" 0
11.757206 "Western Cape" 0
10.035756 "Western Cape" 0
8.4584875 "Western Cape" 0
 8.235003 "Western Cape" 0
 4.575014 "Western Cape" 0
4.6203246 "Western Cape" 0
 6.820771 "Western Cape" 0
  5.43793 "Western Cape" 0
 6.261779 "Western Cape" 0
 7.556087 "Western Cape" 0
 9.544616 "Western Cape" 0
  6.10887 "Western Cape" 0
 7.723115 "Western Cape" 0
 8.265488 "Western Cape" 0
 8.739627 "Western Cape" 0
8.3882065 "Western Cape" 0
 9.581975 "Western Cape" 0
  7.15683 "Western Cape" 0
11.196548 "Western Cape" 0
 6.040884 "Western Cape" 0
4.6368895 "Western Cape" 0
 1.304685 "Western Cape" 0
 3.648412 "Western Cape" 0
 8.993067 "Western Cape" 0
 9.082662 "Western Cape" 0
 8.244229 "Western Cape" 0
 6.497642 "Western Cape" 0
 6.339427 "Western Cape" 0
 -.304753 "Western Cape" 0
 3.869634 "Western Cape" 0
 4.358686 "Western Cape" 0
 4.593833 "Western Cape" 0
 4.692459 "Western Cape" 0
 5.274977 "Western Cape" 0
 8.099204 "Western Cape" 0
  7.76136 "Western Cape" 0
 8.908922 "Western Cape" 0
 6.769872 "Western Cape" 0
 6.903107 "Western Cape" 0
 5.958645 "Western Cape" 0
9.5051155 "Western Cape" 0
 5.130278 "Western Cape" 0
 2.403297 "Western Cape" 0
 6.799309 "Western Cape" 0
 4.813241 "Western Cape" 0
11.510902 "Western Cape" 0
 4.126064 "Western Cape" 0
end

Thank you!

Kind regards,
Chris