The Chinese, all manufacturers agree, love to “steal” machines. Not the actual machine, of course, but the idea of a given machine. Reverse engineering can often be fairly straightforward. Many nations, especially Germany with its enviable machine tool industry, are pushing the WTO to press the Chinese and other developing nations on this margin. But we know from theory that convergence across nations in income can often rely greatly on “learning by imitating” – as any teacher knows, building simple existing devices creates the knowledge base on which novel ideas can grow. But what of history? Is such learning by imitation historically important, as claimed by people like David Landes?
Richter and Streb consider the German machine tool industry in its own early days: 1877 to the 1930s. A quick case study of J.E. Reinecker, a major tool producer, show that firm buying machines from more advanced US firms from 1873 onward, first replicating parts then replicating whole machines. German law at the time did not give patent protection within Germany to foreign inventions not manufactured in Germany. By the late 1800s, Reinecker was creating many novel inventions on its own, applying for patents in both Europe and the US. During World War I, nonmilitary R&D essentially came to a halt in Germany, and the German firms fell behind once again, leading to a decade where imitation of foreign machines once again predominated. When German firms were innovative, the German government assisted them in getting protection overseas: for instance, after 1909, American toolmakers were exempt from the rule that they had to manufacture in Germany, a rule that presumably would lead to more favorable treatment of German patents by US authorities. Further, the extra delay incurred by patent applicants from overseas versus German firms waxed and waned depending on the innovativeness of German firms; at times of imitation, there were long delays for foreign applicants, while in times of novel invention, foreign firms were not treated so harshly.
Using a custom dataset of German tool patents that are “valuable” (renewal fees were paid for at least 10 years), Richter and Streb show a similar pattern across toolmakers as a whole in Germany: imitation early on, then a series of valuable novel patents, then a collapse in WW1 followed by imitation for another few years. Firm level data is not specific enough, except at the level of a case study, to know how important imitation was to the growth of future successful German tool firms, but the evidence presented is at least suggestive of the fact that German firms who imitated (as listed in a complaint by a US industry lobby) produced many of the future valuable patents in their industry.
This paper is just historical evidence, but it does provide a great example of a more general rule: intellectual property rarely follows the same logic as traditional Ricardian trade. There is no particular reason in standard trade theory why IP rules need be harmonized. Indeed, were I (or essentially any economist who has looked at IP) were to advise a third world government, I would absolutely tell them to enforce much weaker IP than that in Europe or the US. Learning by doing matters.
http://eh.net/eha/system/files/Richter.pdf (Aug 2010 Working Paper – final version in most recent issue of the Journal of Economic History)
A Fine Theorem 
When you say “intellectual property rarely follows the same logic as traditional Ricardian trade”, do you mean that technological innovation and transfer is not a one-way transfer from “more developed” countries (those that are better at innovating) to “less developed” countries?
I mean something simpler. Ricardian trade is, at the national level, welfare-improving for both countries. Trade agreements that include IP treaties (like TRIPS) are not necessarily so: *even if* a global IP agreement is globally welfare improving, it is not necessarily so for all nations, and in general, the greatest harms accrue to countries which import IP rather than invent – that is, the poor countries.