This paper studies conversion factors based on the expenditure approach and evaluates the appropriateness for international comparisons of output levels in manufacturing. We apply a consistency check based on the insight that relative productivity levels should be invariant to the choice of base year. Consequently, convergence parameters and dispersion of productivity across countries should also be unaffected by this choice. The results are disappointing: relative measures of productivity depend heavily of the choice of base year and change systematically as the base years roll forward. The conclusion is insensitive to the applied method for developing conversion factors. The implication is that we cannot measure relative productivity levels in manufacturing across countries using the expenditure approach.