Exploratory study of the influence of patent latent variables and firm variables on price of auctioned singleton patents

Abstract

The expected value that a patent might deliver in the market is important information that impacts an organisation抯 revenue. Patent valuation has attracted the attention of researchers since the 1980s. Studies have suggested that a good indicator of patent value is the price of a patent. However, lack of public data on patent price has hampered research in understanding the dynamics of patent pricing. In this thesis, we used patent price data from Ocean Tomo, a patent auction firm, to understand the dynamics of patent pricing. Between 2006 and 2008, Ocean Tomo listed 349 singleton patents for auction. Out of the 349 patents listed, 167 patents which were finally sold form the sample of our study. Pricing a patent is complex and in this thesis we attempted to determine some of the variables that influence the price of a patent. Endogenous and exogenous variables are used as predictor variables. Patent price is the outcome variable used throughout the study. Initially, we study the relationship between endogenous variables (predictor) and patent price (outcome). In this thesis, we use the term patent latent variables to refer to endogenous variables. From existing literature, ten patent latent variables namely age, forward citations, backward citations, number of claims, number of classes, technology field, generality, originality, foreign filings and number of inventors are identified as having an influence on patent value. Data on the variables were extracted from Ocean Tomo, USPTO and NBER databases. After understanding each variable individually, a series of transformations were attempted. We transformed price and patent latent variables to their logarithmic value. The OLS estimates revealed that forward citations, foreign filings and technology field were positively correlated to price. The three variables jointly explained 19.80% of the variance in patent pricing. We did not find sufficient evidence to come up with any definite conclusions on the relationship between price and variables such as age, backward citations, claims, class, generality, originality and number of inventors. We used the Heckman two-stage sample selection model and confirmed the absence of sampling bias. To determine the extent of influence of exogenous variables, we attempted to study the relationship between exogenous variables (predictor) and patent price (outcome). We identified assignee type (individual inventor or organisation) and entity size of the seller (large and small firm) as the exogenous variables. Data on exogenous variables were extracted from the USPTO database. Price was transformed to its logarithmic value. The exogenous variables were introduced into the model as dummy variables. We did not find any conclusive evidence on the relationship between exogenous variables and patent price. Given our results from the initial analyses of endogenous variables, we focused on variables: forward citations, foreign filings and technology field. We divided the sample into patents with foreign filings and patents without foreign filings and continued our analyses. We found that the price of patents owned by small firms increased when they include foreign filings. Patents of small firms get cited more often than patents of large firms, revealing that they have higher "public value". We found that patents of individual inventors get a higher price than patents owned by organisations, when the patents do not have foreign rights. The Inverse Mills Ratio was insignificant. There are several learnings from this thesis. We found that only 15% of the patents sold through Ocean Tomo auctions attracted a price of over one million dollars (11 out of 167). We found that this corroborates the existing prior art that the distribution of patent value is highly skewed and that four-fifths of the total value generated is contributed by one-fifth of the patents. We found that forward citations are good indicators of patent price. Most of the time, a patent with high forward citations attracts a higher price than patents with fewer forward citations. We found that patents which are filed in multiple countries attract higher prices than patents which are not filed abroad. Technology field also has an influence on price. From our study, we found that patents belonging to complex industries (Computers and Communication) attracted a higher price than patents belonging to other industries, including discrete industries (Drugs and Medical). This may be so because valuable patents in the discrete industries are used in-house and may not be put for auction. Earlier research found that due to lack of multi-country filings, small firms often have a low "private value" for their patents. Foreign filings would increase the "private" value of the patents for small firms. Our study revealed that with increased foreign filings, the "private" value of patents belonging to small firms is seen as higher. Individual inventors are more likely to commercialise their patents than organisations, as they lack the necessary resources to utilise the patent optimally. We believe that individual inventors would require the services of a patent broker more than organisations. Therefore, it is possible that patents of individual inventors listed for auction in Ocean Tomo are more valuable than those of organisations when foreign filings are not included. From our data, we understand that organisations engage in more multi-country filings than individual inventors. The individual inventor may not have sufficient resources to engage in foreign filing to maximise the revenue-generating potential of their invention. The thesis proceeds in the following order. Chapter one introduces the thesis. Chapter two reviews the literature on patent valuation, patent price, patent latent variables and firm variables. Chapter three explains the methodology used in the thesis. Descriptive statistics on patent price and patent latent variables are presented in Chapter four. Chapter five explains the relationship between patent price and patent latent variables. Chapter six analyses the relationship between patent price and firm variables. Chapter seven deals with conclusions, limitations and scope for future study. Implications of this thesis for research and practice are discussed.