I don't have stata but what does the command pvarirf do? Remember in a homogenous pvar conditional on the effects the slope coefficients do not vary across panel entries. The implication is that the irf has a similar shape for individual countries. If this is not expected, homogeneous pvar is not an appropriate methodology. If you wish to learn about pvar, chapter 5 of Hsiao (2022), analysis of panel data, Econometric Society, is a good reference. You have chosen a challenging area to analyze.
Added on 30 June.
Are you aware of the fact that a VAR is the reduced form ar some structural model? Impulse response functions are the estimates of the effects of structural shocks to the variables in the system. In order to estimate these shocks you must have a well-defined structural model or provide identification restrictions to your VAR model. This can be achieved by
Theory-based identification in which sufficient restrictions are imposed on the VAR to identify it
Impose external restrictions on the model
Recursive identification - In this case, the variables are ordered in such a way that the explanatory variables for the first variable do not include any other contemporaneous variables. The explanatory variables for the second include contemporaneous values for the first only..... for the nth contemporaneous values for the previous n-1. In software, this is usually determined by the order in which the variables are entered in the command. Experiment by changing the order of the variables in the command you used in your software.
Even in a simple VAR identification requires a considerable knowledge of the economics of the process and is generally very difficult. It is an economic matter that you should deal with before you start your econometrics. Econometrics can not solve identification problems. The unidentified VAR that you have estimated could have arisen from many different structural models, each of which will give rise to different causal effects and impulse response functions.