.png){kind=icon}
.png){kind=icon}
Return
Rounding error and misallocation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.econlet.2026.113038.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Measurement error poses a challenge to measuring misallocation. Using Chinese manufacturing data (1998-2007), we exploit rounding - a directly observable form of measurement error - in firms' reported labor, capital, and value added. The share of values ending in zero exceeds 20% for each variable, more than double the 10% uniform benchmark. Within industry-year cells, the standard deviation of log labor productivity is 0.021 higher among firms whose reports end in zero than among firms ending in other non-rounding digits, with a sharply larger differential in small-firm industries. At the aggregate level, including rounded observations lowers estimated allocative efficiency by 1.6 percentage points.
Keywords:
Rounding
Misallocation
Measurement error
Productivity dispersion
Journal
E
IF:
1.8
Papers:
345
Citations:
0