Deadlines, Work Flows and Work Quality

Abstract

We examine work flow patterns and its relation to work quality using data on patent applications. Because the filing date is critical for determining the priority date for a patent and because ideas likely arrive randomly, the distribution of patent filings can be expected to be uniform over time. Analyzing data on almost 3 million utility patents granted between 1976 and 2009 and about 1.9 million applications applied and published between 2001 and 2010, we find striking deviations from uniform filing volumes, with significant month-end, quarter-end and year-end clustering of patent applications. Results from a number of empirical tests, as well as from interviews with patent attorneys, strongly suggest that these patterns are induced by reporting and planning deadlines at corporates and patent law firms. We then examine the quality of work output during month-ends using three proxy measures of work quality, viz probability of approval, probability of obtaining an 'application incomplete' notice from the USPTO, and duration of review for successful applications. We find, relative to other applications, a small but robust decline in work quality of applications filed at month-ends on all three measures. Interestingly, we do not find any evidence that this drop in work quality is driven by the sorting of lower quality ideas towards month-ends. On the contrary, the evidence strongly suggests that more complex applications get filed close to the end of the month. We conclude by presenting a simple model which reconciles all of the observed patterns as arising from rational responses to discontinuous incentives around month-end deadlines.

Keywords: Innovation, patents, incentives, routines, productivity

JEL Classification: O30, O31, O34, O33, L25