Using both regression- and VAR-based estimates, the paper finds that the exchange rate pass-through to import prices for a large number of countries is incomplete and larger than the pass-through to export prices. Previous studies have reported similar results, which give rise to the puzzle that while local currency pricing is needed to account for incomplete import price pass-through, it would not imply a lower export price pass-through. Recent explanations of this puzzle have emphasized markup adjustment in response to exchange rate changes. This paper suggests an alternative explanation based on the presence of both producer and local currency pricing. Using a dynamic general equilibrium model, the paper shows that a mix of producer and local currency pricing can explain the pass-through evidence even with a constant markup. The model can also explain the observed exchange rate and inflation variability as well as the fact that the regression and VAR estimates tend to be similar.