Conditional Logit Matched Pair Analysis

[Michael Hall]

Hi,

I am estimating the probability of financial distress conditional on financial ratios and corporate governance variables. I have created matched pairs based on industry, size (log total assets) and accounting period. The group variable is the matched pairs indicators. However, because of the matching procedure I have distressed observations repeatedly matched to non-financially distressed firms during different years. For example. firm id 22 is matched to firm a non-distressed firm id 33 in year 2012, then firm id 22 is matched to a firm 37 in year 2015. Is the data structure OK for clogit in STATA or do I have to take the fact that distressed firms are re-used into account in the estimation to make sure the standard errors are correctly estimated?

Thank you in advance for your help.

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. I'm also not sure what you mean by high p values.

If you are matching a firm's data in 2000 with another firm's data in 1998, that is a problem. The match is supposed to be very similar so the match takes care of a bunch of factors that influence both firms. This doesn't work in the match is across years.