Exchange Rate Regimes and Startup Investment Depth: The Creative Destruction Process

Exchange rate regime choices introduce structural trade-offs between financial stability and innovation incentives that shape the startup ecosystem. Utilizing a panel of 40 emerging economies from 2010 to 2024, we overcome data sparsity by deploying non-parametric Machine Learning imputation before constructing customized indices for financial infrastructure, depth, and openness channels. We employ Double Machine Learning Local Projections (DML-LP) to execute a nonlinear causal inference strategy. To prevent traditional fixed effects from mechanically absorbing the treatment variance of sticky exchange rate regimes, we embed the Hybrid-Mundlak controls directly within the DML-LP nuisance equations. Our findings reveal that unexpected transitions toward floating regimes catalyze a creative destruction process that expands financial infrastructure, whereas peg defenses reinforce structural stagnation. Paradoxically, startup depth responds positively to both financial infrastructure expansions and contractions. These dynamics highlight systemic reductions in financial frictions and powerful technological leapfrogging effects. Ultimately, we demonstrate that financial infrastructure serves as the underlying transmission mechanism through which floating regime shifts optimize private capital absorption and mobilization within startup ecosystems.