I am writing a paper on the application of Value-Added TFP to measurement of technological change. Just thinking of throwing in a few variables into my regression. I have ICT capital share (as seen in KLEMS dataset) as another proxy.

I would be grateful to hear your suggestions on other known proxies (if you know papers that relate to those proxies even better).


This paper names some alternatives available:

  • a linear trend (something like assuming constant exogenous growth)
  • ICT capital
  • stock of patents
  • stock of knowledge (proxied by education levels or human capital)
  • R&D
  • energy prices (e.g. see here)

This article mentions the following proxies for technological change:

the NBER productivity series, the Census of Manufacturers series on investment in computers, the industry's research and design to sales ratio, the industry's use of patents, and the number of scientists and engineers employed within the industry.

If you want homogeneous, long and wide coverage, the best option might be the World Bank Development Indicators for science and technology. Some of them have been mentioned above. See the database here, and here is a paper using them to study their importance on technical change.

A dataset which is also quite popular at the moment is that of robots, compiled by the International Federation of Robotics. The logic is the same than that of using ICT (notice both of these could be considered as capital too). For example, see here. The authors state:

The IFR provides data on the number of robots delivered to each industry, in each country and year. We construct the stock of robots based on the deliveries using the perpetual inventory method, assuming a depreciation rate of ten percent. This approach is similar to the EUKLEMS procedure for computing the stock of ICT capital.

  • 2
    $\begingroup$ Also, add to the list a measure of international trade to proxy for knowledge/ideas spillovers effect. In development studies, a measure of access to internet is also used to proxy for the diffusion of knowledge. These two would be relevant if the OP is studying the data for the developing world. $\endgroup$ – london Sep 29 '17 at 15:32
  • $\begingroup$ @london Good point. Spillovers are surely important, particularly for countries far beyond the technological frontier. But is international trade (X+M?) used for this? Sounds like a dangerously endogenous variable (considering that GDP is defined with these variables). $\endgroup$ – luchonacho Sep 29 '17 at 20:10
  • $\begingroup$ I think, a seminal paper by Gorssman and Helpman (1990) uses trade volume nber.org/papers/w3485.pdf and there have been subsequent research. The OP may need to find an instrument for the variable. $\endgroup$ – london Sep 29 '17 at 22:24

I would say these are some proxies


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.