This paper has developed and implemented a two-stage least square (2SLS) simultaneous equation model based on Hayashi and Sims (1983). First, we have applied autoregressive distribution lag (ARDL) bounds testing approach for estimating the macro-economic determinants of foreign exchange rate (FER). The presence of long-run stable co-integrating relationship between FER and indirect macro-economic variables is established. The second stage examines the impact of pre-determined exchange rate and other (direct) macro-variables (like IIP, inflation, interest rate parity, oil trade index, and so on) on stock return (SR) in an OLS framework. In line with theory, we find a significant impact of predicted exchange rate, gold prices, risk free returns and international security returns on stock returns on the Index. We conclude that the relationship between FER and SR is a complex simultaneous one and it cannot be captured by a single equation model as has usually been done in extant literature.