The canonical New Keynesian model specifies inflation as the present-value of future real marginal cost. This paper tests this New Keynesian Phillips Curve and exploits projections of future real marginal cost generated by VAR models to assess the model's ability to match the behavior of actual inflation. In accordance to the literature, the model fits Euro data well at first sight. However, analyses of this kind disregard the considerable degree of uncertainty surrounding VAR forecasts. A set of bias-corrected bootstrapped confidence bands reveals that this result is consistent with both a well fitting and a completely failing model. Allowing for inflation inertia through backward-looking indexation narrows confidence bands around measures of the model's fit but, still, cannot generate sufficiently precise estimates. Hence, we cannot say whether the model fits or fails.