Unveiling the J-curve: How Intangibles Drive Productivity Mismeasurement

This paper identifies a firm-level Productivity J-curve induced by intangible investments. Using novel microdata on business-to-business transactions for Belgian firms, we construct a comprehensive measure of intangible investment covering software, R&D, design, training, and organizational capital. Our analysis shows that returns on intangibles substantially exceed those of traditional production factors, highlighting their central role in value creation. However, because intangible expenditures are rarely capitalized and are often recorded as intermediate inputs, they are not properly accounted for in conventional measures of total factor productivity (TFP). This misclassification creates systematic mismeasurement, whereby inputs are overstated relative to output in the short run, leading to an underestimation of TFP. Exploiting the lumpy nature of intangible expenditures within a difference-in-differences event-study framework, we document that such mismeasurement results in a persistent underestimation of TFP, by about 3% over a seven-year horizon. Given average measured TFP growth of 1% annually, this represents a substantial distortion. The bias is strongest among small, young, and low- capital intensive firms, reflecting slower absorption of intangible assets. By clarifying how intangible capital and emerging technologies such as AI systematically distort measured productivity, our findings provide new empirical insights into the productivity slowdown and the role of mismeasurement in modern economies.