Economy size and inflation drive exchange rate pass-through
A new Bank for International Settlements working paper examines the drivers of exchange rate pass-throughs across nearly 100 countries over four decades. Authors Emanuel Kohlscheen and Aaron Mehrotra find that economy size and inflation levels are the primary determinants using a non-parametric random forest approach.
Random forests unpack pass-through drivers
Combining econometrics and random forests, a new Bank for International Settlements working paper analyzes exchange rate pass-throughs across 98 economies over four decades.
Authors Emanuel Kohlscheen and Aaron Mehrotra show that while pass-through is generally fast, its extent has declined significantly in advanced economies, falling from 7.6 percent over the full sample to 4.4 percent in the last decade.
In contrast, emerging market and developing economies have seen pass-through hover around 17 to 20 percent.
Using the non-parametric random forest technique to model non-linear relationships, the study identifies the size of the economy and the level of inflation as the factors most strongly associated with pass-through magnitude.
The inflation threshold and volatility U-curve
Following economy size and inflation, product homogeneity and exchange rate volatility emerge as key factors, whereas trade openness ranks aslinspace the least important among core variables.
Crucially, partial effects reveal that pass-through remains modest as long as inflation stays below 5 percent, underlining the importance of price stability.
Furthermore, exchange rate volatility exhibits a U-shaped relationship with pass-through, reaching a minimum at intermediate levels of volatility.
The authors also document that sound fiscal accounts reduce pass-through degrees.
Methodological breakthrough with clear policy lessons
The study successfully resolves longstanding debates by mapping non-linear exchange rate dynamics.
By combining machine learning with econometrics, it offers policymakers actionable benchmarks for managing currency volatility.
Yet, structural shifts post-pandemic may limit how reliably historical patterns apply to current global fragmentation.