Innovation is traditionally thought of as a process that takes place inside of a firm, university or government lab. Private households have often been overlooked as sources of invention and innovation. Increasingly, evidence has begun to emerge that shows the importance this sector plays in this regard (e.g. von Hippel et al. 2012, Arora et al. 2016). These studies are based on survey data which provide wide-ranging detail, but issues of low response rates and challenges identifying innovations remain a concern when interpreting the results.
In a recent paper (Miranda and Zolas 2018), we take a different approach. We use administrative data from the US Patent and Trademark Office (USTPO) and the US Census Bureau to describe patented household innovations in a systematic way. We link USPTO patent data to the Integrated Longitudinal Business Database (iLBD), a panel database of all non-employer businesses in the US, and the Census-NUMIDENT file, a database containing demographic information. The combined dataset provides us with a longitudinal directory of all self-employed businesses, their patented innovations and the characteristics of the inventors associated with them starting in 2000 and through 2011. While patented household innovations arguably represent but a very small slice of household innovations, it is perhaps the most valuable one (Arora et al. 2016).
Who are the household innovators?
Our combined dataset reveals information about the characteristics of household inventors. These are US-based inventors identified in the assignee document, who can be tied to a self-employed business or whose patents are not assigned. We compare them against the population of US-based salaried inventors whose patents are assigned to firms with employees. We identify 1.48 million US-based inventors in the patent database. These inventors are associated with 1.29 million patents. Compared to US-based salaried inventors, we find household inventors are disproportionately US-born (82% versus 66%). They are relatively white (85% versus 74%) and equally likely to be male. Inventors under the age of 25 and over 55 are disproportionally household inventors (34% versus 20%), consistent with the idea household innovations are associated with students and older inventors (von Hippel et al. 2012). Household innovations are also a white-male dominated activity. Across the board, we find a deficit in female and black inventors relative to the overall working age population. This is consistent with Bell et al. (2015), who look at innovation more broadly.
Figure 1 Inventor demographics by assignee type
What do household innovators innovate?
We can also say something about the types of innovations that emerge from households. Household inventors work on technology classes disproportionately tied to consumer products such as Design and Mechanical and Other. This is consistent with survey data that show independent inventors contribute substantially to consumer product innovations (von Hippel et al. 2012, Shah 2000). Not surprisingly these consumer product innovations often include design elements. Analysis of the top 50 companies having been granted design patents shows these are dominated by technology, automotive, and consumer product companies.
In terms of complexity and sophistication, evidence from surveys and product studies suggest that the complexity and knowledge embodied in household innovations might not run very deep. We follow Jones (2009) and use inventor team size as a proxy of the complexity and depth of knowledge associated with a particular innovation. We find team size for patents matched to non-employer business tend to be significantly smaller on average than patents matched to employer firms, having nearly one less team member (1.65 team members versus 2.6).
Looking at innovative value, household innovations tend to have fewer citations (27-33% fewer on average), but their impact remains remarkably high. The difference in average citation counts is driven in part by composition effects, technology classes and sophistication. For instance, design patents, which non-employers are more than twice as likely to take out relative to employer firms, receive far fewer citations than utility patents. We find patents have a mean citation count of 16.4, 13.4, and 11.4 respectively for patents associated with employer businesses, non-employer businesses and no business activity. Despite the difference in citations, the mean number of citations and proportion of ‘radical’ patents - defined as novel, unique, and adopted (Dahlin and Behrens 2005) – for household innovations is relatively high. The bulk of radical household innovations are found in Computers & Communications, Design, and Drugs & Medical. They include a system for providing traffic information to a plurality of mobile users connected to a network, a system for organising virtual offers from the internet, a method and apparatus for securing a suture and a flash memory drive with quick connector. All these technologies had broad impacts in their fields.
Figure 2 Patent technology class shares and mean citations by assignee type
What are the economic impacts of these innovations?
The bigger question underlying these measures of household innovation is whether the innovators are able to monetise their innovation without relying on the corporate structure of the firm. To address this question, we look at the population of non-employer businesses associated with these household innovations. There are relatively few household inventors that start a business. Only 19% of patented household innovations are associated with a business. The equivalent rate for patents with a declared business assignee is 93%. Household innovators that do not start a business might be able to monetise their innovations in other ways, either through direct payments or from selling or licensing the use of the patent.
To measure the economic impact of these inventions we first look at the revenue generated by these household businesses and compare these flows against the revenue generated by similar household businesses without patents. Non-employer businesses with a patent are rare. Out of more than 20 million non-employer firms in a given year, between 5,000 and 10,000 firms will seek out a patent (less than 0.03%). Approximately 43.6% of non-employer firms that are granted a patent apply for the patent prior to starting their non-employer business activity. For many other businesses, the birth of the business coincides with the patent application year. A sizable number of patent applications, 18%, are filed three or more years after starting the business activity suggesting that a non-trivial number of businesses are started perhaps with the intent of generating income from the innovation.
We find non-employer businesses with patents generate higher revenues than those without a patent (approximately $10,200 in annual revenue on average versus $9,700 for non-employer businesses who hold no patents) and have longer survival rates (3.95 years versus 2.72). This is relatively small when compared to the revenue generated by the average innovative employer businesses of $1.2 million, although evidence suggests the costs of development are also much smaller. We find a fairly wide revenue distribution amongst innovative non-employer businesses, with a standard deviation of $97,500.
In specifications controlling for selection which also include demographic characteristics (male, US-born, race, age) and non-employer characteristics (industry, zip code, year), we find that a doubling of granted patents is associated with a 3-4% increase in revenue, while a doubling of citations is associated with a 0.3-0.6% increase in revenue.
Our analysis indicates patented household innovations granted between 2000 and 2011 generate a revenue flow of $1.7 billion. Extending this to household innovations, we were not able to match to our databases, and assuming similar revenue flows suggests total revenue flows of $5.0 billion in 2000 dollars.2
To conclude, patented household innovations have both impact and value. Many of them are radical and represent breakthroughs in their fields and the economic benefits are quite sizable. Despite efforts to understand their role in the economy our knowledge of innovations and their inventors remains limited due to data constraints. Administrative data helps shed light on this population and their impact, but a more targeted survey based on the population of patented household innovators as identified through administrative records such as the ones used here would go a long way to expand our knowledge in this area.
Authors’ note: Any opinions and conclusions expressed herein are those of the authors and do not necessarily represent the views of the US Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.
Arora, A, W M Cohen and J P Walsh (2016), “The acquisition and commercialization of invention in American manufacturing: Incidence and impact”, Research Policy 45(6): 1113-1128.
Bell, A, R Chetty, X Jaravel, N Petkova and J Van Reenen (2015), “The lifecycle of inventors”, mimeo, Harvard University, 5 June.
Dahlin, K B, and D M Behrens (2005), “When is an invention really radical? Defining and measuring technological radicalness”, Research Policy 34(5): 717-737.
Jones, B F (2009), “The burden of knowledge and the death of the renaissance man: Is innovation getting harder?”, The Review of Economic Studies 76(1): 283-317.
Miranda, J, and N Zolas (2018), “Measuring the impact of household innovation using administrative data”, NBER Working Paper 25259.
Shah, S (2000), “Sources and patterns of innovation in a consumer products field: Innovations in sporting equipment”, Sloan School of Management WP-4105, MIT.
von Hippel, E (1976), “The dominant role of users in the scientific instrument innovation process”, Research Policy 5(3): 212-239.
von Hippel, E, J P J de Jong and S Flowers (2012), “Comparing business and household sector innovation in consumer products: Findings from a representative study in the United Kingdom”, Management Science 58(9): 1669-1681.
 Appropriate caution must be taken with these back of the envelope calculations. Our work does not attempt to place an economic value to innovations that are not known to the patent system which are presumably the bulk of all household innovations. In addition we had to make some strong assumptions to arrive at this number the most important being the revenue generated by business started by household inventors themselves is the same as the revenue generated by household innovations whose outcomes we are not able to observe including those sold to or appropriated by existing businesses. A different set of assumptions would undoubtedly yield different numbers. We also assume the innovative nonemployer businesses we observe would not have existed were it not for their innovation.