Abstract: Is growth always efficient? We examine the effects of transaction costs on accumulation-based and efficiency-driven growth regime building on the novel augmented neoclassical growth model with non-zero transaction costs. Higher transaction costs always lead to sub-optimum resource allocation discouraging efficiency improvements. The persistence of high-cost equilibrium tends to foster the departure from the productivity frontier. Constructing a large dataset encompassing 139 countries in the period 2003-2014, we tackle the contribution of transaction costs to the efficiency- and accumulation-based growth paths by developing a novel series on total factor productivity paths over time. Using latent variable methods, three internally consistent components of costs are identified: (i) cross-border, (ii) procedural, and (iii) administrative. The evidence suggests lower administrative costs encourage stationary TFP performance while lower procedural and cross-border costs exhibit a strong positive impact on TFP growth. Both effects are robust to the measurement error across multiple subsamples, and are not susceptible to time-invariant heterogeneity bias and unobserved technology shocks. Employing country-specific stationarity tests, we examine the break points in TFP performance for each individual country and use the variation in transaction costs to explain (i) efficient growth accelerations and disaster episodes and (ii) establish the conditional probability of entry to and exit from both types of growth regimes. Lower administrative transaction costs tend to produce unsustainable accumulation-driven growth episodes that eventually fizzle out while lower procedural and cross-border costs are significantly more likely to create sustained growth accelerations. The variation in transaction costs accounts for a substantial fraction of cross-country differences in TFP.