The Life-Cycle Growth of Plants: The Role of Productivity, Demand and “Distortions”

Abstract

Using rich product level data on prices and quantities at the establishment level for Colombia, complemented with information on input use, we develop a methodological approach to decomposing output and sales growth over an establishment's life cycle. Our approach measures output at the plant-level using plant-level deflators that capture within plant changes in quality/appeal from product turnover. Idiosyncratic variation in measured fundamentals (e.g.,technical efficiency (TFPQ), between plant variation in demand/appeal, and input prices) explain more than 100% of the variability of output and sales relative to birth level, so that implied wedges dampen actual growth relative to that implied by fundamentals. This is consistent with either adjustment frictions or distortions that implicitly tax plants with strong fundamentals. About two thirds of the contribution of fundamentals to output growth is from technical efficiency and one third from idiosyncratic variation in demand/appeal, while for sales this latter component explains over 80% of the contribution of fundamentals. A reduced form decomposition of output and sales growth volatility implies that uncorrelated wedges account for about 20% of output and 10% of sales growth volatility. The (absolute) contribution of both correlated and uncorrelated wedges is substantially greater for young compared to more mature plants. We also show that our inferences about the respective role of fundamentals and wedges are quite different from those drawn from revenue and input data alone. Use of the latter (which is the common approach in the literature) yields a substantial overstatement of the contribution of role of wedges in accounting for sales growth volatlitiy.