Comment

Understanding the determinants of innovation and productivity growth is a core open area of economics research. Although enormous progress has beenmade in theoretical models of innovation accompanied by an increasing use of firm-level data to quantify the nature of innovation and productivity, many challenges remain. A key challenge is that much of the research using firm-level data has focused on firms with observable measures of the inputs into innovation (e.g., research and development [R&D] expenditures) and direct measures of the success of innovations (e.g., patents). This approach focuses on a relativelynarrow subset offirms and sectors where such observables are relevant. Most firms do not report R&D expenditures or patents. It is implausible that only the firms with these observable measures of innovation are responsible for the observed fluctuations in productivity growth from innovation. As evidence of this, a National Academy of Sciences report (Brown et al. 2005) highlighting the limitations of R&Ddata reported that one of themost innovative firms in retail trade, Walmart, reports no R&D expenditures in its 10-K reports. This paper takes an indirect approach to identifying innovation activity that overcomes these limitations. Using an innovative growth accounting framework that is motivated by a quality laddermodel of innovation, this paper uses data on the employment growth rate distribution for the universe of private sector, nonfarm (hereafter private sector for short) establishments to quantify the contribution of creative destruction (CD), own innovation, and new varieties. The authors accomplish this important objective by using the Longitudinal Business Database (LBD) that tracks the employment dynamics, including entry and exit,firm size, andfirmage of