Hello people,
I'm trying to evaluate the impact of remittances on expenditure using propensity score matching (teffects). I'm at the initial stages of my analysis and wanted to make sure STATA and the PC would compute results without problems. My outcome variable is expenditure in education (continuous), dependent variable receiving remittances and independent variable is area (dichotomous rural or urban, namely 1 and 0). Again, at this instance, I just wanted to make sure STATA could do the numbers and that's why I only have area as independent dichotomous variable. However I encountered problems.
Code:
 teffects psmatch (gmed) (remittances area)
_matchnn_neighbors_u (): 3900 unable to allocate real tmp [25309,1]  
_matchnn_neighbors_u ():        - function returned error
 
_matchnn_neighbors_u (): 3900 unable to allocate real tmp [8663,1]  
_matchnn_neighbors_u ():         - function returned error
_matchps_estimates ():                - function returned error
Istmt                                                    - function returned error
This is from household statistics database containing around 74k total observations of which 20k where treated (remittances) and 54k non treated. I'm running STATA v14, hp Corei7 processor with 8gb ram.
Any suggestions?

My best,