IV results interpretation

#1
Hello,

I have an endougenous variable x and an instrument for it, variable z. The instrument is (significantly) negatively related to x. My hypothesis expects a positive effect of my endogenous variable x on y, my dependent variable. When using 2SLS ivreg in Stata does this mean that (because of the negative correlation between z and x) the sign of the betacoefficient for x in the second stage should be negative in order for my hypothesis to be confirmed? Or should it still be a positive beta to be consistent with my hypothesis?

Thanks a lot in advance for your help,

Best,

Carmen