i am writing a paper analyzing What is the effect of corruption on Government expenditure (education,health,military and infrastructure). I have data for 30 countries for time period 14 years. So it will be a panel unbalanced dataset. Data on 4 government expenditures and plus data on corruption index.

i need help what model to use for regression, can I use nominal GDP as control variable or is GDP needed here? Or should i use a population as a control variable and not include at all nominal (current) GDP? I want to do simple regression with not too many variables so probably i will have 4 equations (one for each indicator of government spending) Please any suggestion! thanks a lot