I build a novel geo-coded firm-level panel dataset that, for the first time, allows for the evaluation of the implications of Special Economic Zones (SEZs) on the distribution of firms. SEZs are locations where firms enjoy reduced corporate taxes and enhanced access to credit, contingent upon maintaining a minimum scale. First, I empirically show a new set of stylized facts summarized by SEZs being associated with better firm selection, higher capital and better within-firm resource allocation across time. Specifically, with the aid of the new panel data, I show that: (1) firms born in SEZs are permanently more productive (better selection), invest more (higher capital), and show a higher within-firm correlation of TFP and capital over time (better within-firm resource allocation across time) compared to firms born outside the SEZs; and (2) firms that move to SEZs, having originated outside, manifest similar advantages post-relocation (better selection, higher capital and better within-firm resource allocation across time) compared to firms that stay outside the SEZs. However, even if SEZs were randomly allocated across China (which is not the case), we cannot attribute the significant differences in firm-level performance across zones solely to SEZs. This is because the decision to enter an SEZ, either through birth or relocation, is endogenous (with 55% of firms in SEZs being born in SEZs and 45% relocating from non-SEZs). Moreover, I am interested in understanding the effects of SEZs on aggregate productivity and output and, ultimately, determining the optimal size of SEZs—and notably, an empirical analysis for these two aspects is not possible. To address these questions, I build a firm dynamics model that incorporates endogenous entry, exit, and location (zone) choices. The model is carefully calibrated to match my empirical findings. Then, through counterfactual experiments I find that the SEZ increases aggregate TFP by 25.7%. Looking into the mechanisms I find that the rise in aggregate TFP is due to both better selection (average firm TFP increases by 25.1%) and better resource allocation (the correlation between capital and TFP increases by 88%). Isolating the role of financial frictions, I find that approximately half of the increase in the aggregate TFP is generated by the reduction of financial frictions in SEZs, which induces better selection (in births and movers) and better resource allocation for the endogenous distribution of firms.