Because editors of top economic journals, published only papers which were in line with neoclassical models.

Because economists have (or used to have, at least), strong theoretical presumption about causal effects they have investigated, which affected their choices of empirical specifications.

Because economists manipulate their results.