Research papers of Peter Benjamin Meyer
Active research works:
Network of tinkerers: a model of open source invention

Nov 2007
Airplanes were invented by hobbyists and experimenters, and some personal computers were as well. Similarly, many open-source software developers are interested in the software they make, and not focused on profit. Based on these cases, this paper has a model of agents called tinkerers who want to improve a technology for their own reasons, by their own criteria, and who see no way to profit from it. Under these conditions, they would rather share their technology than work alone. The members of the agreement form an information network. The network's members optimally specialize based on their opportunities in particular aspects of the technology or in expanding or managing the network. Endogenously there are incentives to standardize on designs and descriptions of the technology. A tinkerer in the network who sees an opportunity to produce a profitable product may exit the network to create a startup firm and conduct focused research and development. Thus a new industry can arise.
The airplane as an open source invention

In progress, Oct 2007
Discusses the network of experimenters and the information available when the Wright brothers began work. Makes the analogy to open source software.
Turbulence, inequality, and cheap steel

February 2005
According to new measures in this paper, earnings inequality rose in the iron and steel industries during this turbulent time when new mass-production technologies were adopted. This paper addresses why that would be, and discusses historical evidence that this rise in inequality resulted from a chaotic environment which resulted from technological uncertainty and opportunity. The players could not predict the future. They constructed new skills and organizations.
Episodes of collective invention

August, 2003
The process of developing technology through open discussion has been called collective invention. Open source software projects have this form. This paper documents two earlier episodes of collective invention and proposes a general model based on search theory. One episode was the development of mass production steel in the U.S. (1866-1885), and the second with early personal computers (1975-1985). Technical people openly discussed and shared these developing technologies between firms. Collective invention episodes begin with an invention or a change in legal restrictions. Hobbyists and startup firms experiment with practical methods of production and share their results through a social network whose members gradually form a new industry. The network itself may disappear if the firms then keep their R&D secret. A model of an innovation search can describe this process if it is expanded to include independent hobbyists and consultants as well as profit-seeking firms.
Technological uncertainty and superstars: two sources of inequality within occupations (1968-2003) Some of the rise in income inequality in the U.S. from 1973-1996 has been traced to changes associated with the new information technology such as personal computers. Measurements of earnings inequality within occupations show the degree to which this technology seems to cause detectable effects associated with job role. Two effects are shown. First is a rise in earnings dispersion within the professions whose core work content included using and adapting to new semiconductor and information technology. This could be an effect of technological uncertainty and differences in adaptability between workers (as in Greenwood and Yorukoglu (1997)). The second effect is the "superstars" phenomenon named by Rosen (1981). Earnings dispersion increased in performance occupations like athletics, in which the top producers gain distribution or marketing advantages from new information and communication technologies, expanded trade, and globalization.
Proposed category system for 1960-2000 Census occupations by Meyer and Osborne, 2005 This paper defines a detailed, consistent category system for occupations in the U.S. Census of Population data from 1960 to 2000 and Current Population Survey data since 1968. The categories are based on the 1990 Census coccupation definitions, coarsened somewhat into 389 categories to achieve better consistency over time. We summarize employment levels, average earnings levels, and earnings variance in these occupation catgories over time, compare these to similar trends for occupations defined by the occ1950 classification at, and test both classifications for consistency over time.
Preliminary MFP measurement (aka, quick MFP) by Meyer and Harper, 2005 We measure US private business sector multifactor productivity based on preliminary data. This method is now used to produce the earliest official MFP from BLS.
Stylized facts of the economics of major inventions Under development. Show a baseline pattern of descriptive statistical facts about each takeoff of several radically new technologies (Bessemer steel, microcomputers, cars, transistors, planes, phones, two versions of steam engines) which an algebraic theorist of technological change could try to match. Oddly this seems not to have been tackled on such a wide scale although it is extremely interesting. Among these are: a rise then fall in earnings inequality in the affected workforce; an inflow of new producing firms followed by a shakeout; an inflow of relevant immigrants; and there are many others. This is a synthesis of the findings above, whose development was motivated to support this more general work.
Weeks report data set from 1880 Census report The Weeks Census report contained results of a large 1880 survey of firms and the wages they paid to manufacturing workers. With assistance, I made an electronic data set of them. I do not know of another one. Freely copy it and cite it. I created it in order to analyze the question of whether income inequality had risen in response to Bessemer steel (the paper at the top of this list).
The TAIME project Carol Siri Johnson and I have scanned the Transactions journal of the American Institute of Mining Engineers for 1871-1885. We're looking for markers in the text of uncertainty in steel and other aspects of the important technological changes occurring then -- e.g. that over that period production of steel went from tens of thousands of tons per year to something like 2 million tons a year. And the rise of Big Steel.
A model of collective invention

Latest presentation of tinkerers paper, June 14, 2007, OSSEMP 2007

This is an early draft of a micro model and simulations which make the assumption that each of two players wants to (a) use a technology for some exogenous reason, and (b) advance that technology in order to lower costs. Then this explains (c) why these players would choose to share innovations with one another. There are important historical cases where this seems to have happened, such as the early developments with Newcomen steam engines, the post-1800 developments with Watt steam engines, the invention of the airplane, the invention of the personal computer, and all open-source software projects.
Employment relationships and the labor market during technology takeoffs This was my dissertation proposal. It includes versions of the steel and microprocessor/PC papers at the top of this list.
Private information in government bonds In this class paper I tried to show that some national governments of industrialized countries have private information about the future of inflation at the time the national governments issue inflation-indexed bonds. The evidence took the form of a decline in inflation after the issue ("Granger-causation") in these countries: U.S., U.K. (dramatically), Australia, Canada, and Sweden. The hypothesis is true in each case in this small sample. Seeking stronger evidence I fit the inflation series to an AR-GARCH specification with another indicator variable which marked the post-issue period. The coefficient had the predicted sign but lacked statistical significance and there were other econometric problems. Yet the result is (a) interesting and (b) very different from anything I've seen in the research literature on *government* behavior, and (c) likely to be true, on the basis of other evidence. I will return to this subject when I find the right coauthor or venue.

Back up to Peter B. Meyer's homepage