The Upward Pricing Pressure (UPP) test developed by antitrust economists Joseph Farrell and Carl Shapiro marks a new era in antitrust and provides an alternative to the traditional concentration-based tests in merger analysis. In addition to being free of market definition, the UPP's appeal lies in its ease of use: one simple formula indicates whether a merging firm has an incentive to increase prices post-merger. This paper first establishes the theoretical relationship between the UPP and the standard structural merger simulation, namely, that the UPP is a "single-product merger simulation" that ignores the re-equilibration of all other endogenous variables except that product's own price. To assess the consequence of this simplification, I compute "true" UPP values for a cross-section of airline markets using structurally estimated price elasticities, and confront them with the "gold standard" of a merger simulation. I examine the predictive accuracy of both the sign and magnitude of the UPP. I find that it gives wrong sign predictions to an average 10% of the observations, and its value has an average correlation of 0.92 with the structurally simulated price changes. However, since this test is meant to bypass a complicated demand estimation, I then use the example of a simple logit demand to illustrate the consequence of using inaccurate demand-side inputs in the UPP: the test will give a wrong sign prediction over a much larger range of cost synergies. Lastly, I discuss the pass-through conditions for Farrell and Shapiro's proposition, demonstrate empirically that they are not innocuous, and show that their violation can lead to false positive results (type I errors) in the UPP.