Abstract:  This paper documents the paradox of Chinese individual patent filings, namely, individual inventors' patents are widely regarded as of lower quality compared to firm-filed patents; while on the other hand, they are maintained significantly longer than their firm counterpart. We point to the inefficiency of the patent market to explain the paradox: because of the high search cost and information asymmetry, it takes longer time for individuals to license their patents to a potential commercialized firm, without which the individual inventors cannot reap the benefit and cover the investment cost. The individual inventors choose to wait for potential buyers of their invention until the invention becomes almost obsolete. We construct a dataset on Chinese patents for empirical analysis, incorporating crawled data from Google Patents to include the number of claims and citations for the patents, and then match with SIPO data. These two variables will be instrumental in studies that focus on patent quality and knowledge flows. The ex-ante quality measures (claims, number of subclasses, number of inventors, grant lag, grant rate, etc) suggest that the individual patents are of much lower technical quality, compared to the firm or academic patents. These statistics show that individual applications are not simply disguised form of firm applications: they indeed have significantly different quality measures, even conditional on being granted. However, when we look at the ex-post quality measures, we find that individual patents have a much longer life than firm patents: they are on average maintained 10 months longer than firm patents! Since the life-in-force has been regarded as one of the most reliable indicators of patent value, the finding constitutes the paradox of the Chinese individual filings, as a large proportion of individual innovators are non-performing entities (NPEs), holding their patents longer means paying maintenance fees without being compensated. We provide an explanation based on the observation that individual innovators, especially those NPEs, need to license or transfer their patents to other performing entities (like firms) to make their innovation activity a livable business. Therefore, we postulate that individual innovators hold their patents longer not because their patents are more valuable, but because they are waiting for potentially interested firms to license/transfer their patents. They only choose to abandon the patents when the technology embedded in the patents is close to obsolete, or in other words, the patents do not have a very long economic life left even if they are developed and commercialized.We provide four pieces of evidence supporting our hypothesis. The first piece of evidence is that it takes significantly longer time for the individual invention patents to get licensed/transferred, compared to firm patents. We compute the license/transfer lag as the time elapsed between the application date and the first ever license/transfer date. The results suggest that conditional on being granted, the license/transfer lags for individual and firm patents that have been licensed/transferred at least once are around 1651 and 1445 days respectively. It takes around 200 more days for individual patents to get licensed / transferred, compared to firm patents. Higher search cost and information asymmetry for technology transfer between firms and individuals, compared to inter-firm technology transfers lead to individuals waiting longer in hope of eventually getting their patents commercialized. Since the grant of a patent is a crucial event that reduces the uncertainty embedded in the invention, we may expect the difference in the lag is due to more of the firms’ patents being licensed/transferred before they were granted. Following this logic, conditional on the license/transfer occurs after the patent has been granted, the license/transfer lag between individuals and firms should be smaller. In the second piece of evidence, we compare the difference in license/transfer lag between individual and firm patents conditional on grant. Indeed, the difference becomes smaller but is still large and significant at about 130 days. The third piece of supporting evidence is that, conditional on ever being licensed or transferred, patents originally owned by individuals have roughly equal duration in-force compared to patents owned by firms. Individual patents are maintained on average, 300 days longer than firm patents; however, the gap becomes much smaller and drops to only 76 days if we only consider the subsample of licensed/transferred patents. In the last piece of evidence, we show that individual patents are maintained close to the “ideal” economic life of patents when the technology is developed and commercialized. The licensed/transferred patents are in general of very high value, thus it’s expected that they have a longer life than the population average. If we call the life-in-force of a licensed/transferred patent the “maximum economic life in application”, we notice that the average individual patents’ life (2284 days) is much closer to this “maximum life” (2392 days) than their firm counterpart (1987 days compared to 2316 days). It seems the individual innovators hold their patents to the time point when even if they are licensed / transferred, there would be of little value left to be appropriated. It can be explained by the innovators being overly optimistic of the probability, or simply unwilling to accept the sunk cost, i.e., the “sunk cost fallacy”. In other words, individuals will hold on to their patents until they worth little even if transferred.It is well-known that the market for intellectual property (IP) is one of the last large and inefficient markets in economy. IP is the ultimate intangible asset and extremely hard to value. Moreover, there are very high search and transaction costs on both sides of the market (inventors and patent users). Information asymmetry and the risk of litigation make all potential participants even more cautious. Empirical evidence on the consequence of such inefficiencies is, however, scarce. The paper provides some much-needed evidence for the inefficiency of the technology market, based on a new theoretical mechanism to explain the paradox of individual filings in China and four pieces of empirical evidence for the mechanism. We conclude with a final implication of the findings on patent life. If the patent serves a different purpose (i.e., licensing / transferring) rather than production and commercialization, in a market with frictions on the transaction purposes, the patent life may not be a good indicator of its value.

Key words: market inefficiency, patent quality index, patent maintenance