teffects psmatch: logit failed to estimate propensity scores

[Sarah Marilles]

Hello,

I am trying to use teffects psmatch, my code below is specifically:

foreach y in 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 ///
2011 2012 2013 2014 2015 2016 2017 2018 2019 {
teffects psmatch (av_party_choice`y') (dstreat ///
av_age`y' av_cohab_status`y' av_educ`y' av_year_spouse`y' av_satis_spouse`y' ///
av_satis_health`y' av_health_prob`y' av_occup`y' av_satis_income`y' av_satis_finan`y' ///
av_pol_interest`y' av_satis_democ`y' av_trust_govn`y' av_vote_turn`y' av_pol_position`y' ///
av_opin_army`y' av_opin_social`y' av_opin_EU`y' av_opin_foreign`y' av_opin_env`y' av_opin_tax`y' ///
av_opin_nucl`y' av_relig`y' av_satis_leisure`y' av_health_stat`y' ///
av_hh_asset`y' av_hh_priv`y' av_hh_pub`y' av_hh_pension`y' av_labour_inc`y' ///
av_windfall`y' av_hh_windfall`y' av_hh_labour`y' av_hh_inc`y' av_satis_social`y' ///
av_birthy`y' av_sex`y' av_move_house`y' av_satis_life`y' av_satis_fin_chg`y' ///
av_age2`y' av_length_mar`y'), atet
}

However, i get the error message 'logit failed to estimate propensity scores; computations cannot proceed'.

I have tried to compute the below and it works perfectly fine.

foreach y in 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 ///
2011 2012 2013 2014 2015 2016 2017 2018 2019 {
logit dstreat ///
av_age`y' av_cohab_status`y' av_educ`y' av_year_spouse`y' av_satis_spouse`y' ///
av_satis_health`y' av_health_prob`y' av_occup`y' av_satis_income`y' av_satis_finan`y' ///
av_pol_interest`y' av_satis_democ`y' av_trust_govn`y' av_vote_turn`y' av_pol_position`y' ///
av_opin_army`y' av_opin_social`y' av_opin_EU`y' av_opin_foreign`y' av_opin_env`y' av_opin_tax`y' ///
av_opin_nucl`y' av_relig`y' av_satis_leisure`y' av_health_stat`y' ///
av_hh_asset`y' av_hh_priv`y' av_hh_pub`y' av_hh_pension`y' av_labour_inc`y' ///
av_windfall`y' av_hh_windfall`y' av_hh_labour`y' av_hh_inc`y' av_satis_social`y' ///
av_birthy`y' av_sex`y' av_move_house`y' av_satis_life`y' av_satis_fin_chg`y' ///
av_age2`y' av_length_mar`y'
}

I have tried to add in 'mlogit' but I get 'option mlogit not allowed'. Am I allowed to use multinomial logit regressions with teffects seeing as my outcome variable is categorical.

Thank you!

Sarah