An Exploration of Technology Diffusion
read more
Citations
The New Kaldor Facts: Ideas, Institutions, Population, and Human Capital
Misallocation and productivity
Nonhomotheticity and Bilateral Trade: Evidence and a Quantitative Explanation
Social Connectedness: Measurement, Determinants, and Effects
The Facts of Economic Growth
References
Endogenous Technological Change
Why Do Some Countries Produce so Much More Output Per Worker than Others
The role of human capital in economic development Evidence from aggregate cross-country data
International R&D Spillovers
Investment in humans, technological diffusion and economic growth
Related Papers (5)
Frequently Asked Questions (16)
Q2. What future works have the authors mentioned in the paper "Nber working paper series an exploration of technology diffusion" ?
The line of research developed in this paper leaves several doors open for future research. 19 Finally, it will be interesting to extend this analysis to other technologies and countries. 18One potential difficulty of pursuing this route at this point is the quality of sectoral TFP data.
Q3. What are the main determinants that the authors allow for in the vector with explanatory variables?
The main determinants that the authors allow for in the vector with explanatory variables, xτ ,t, can be classified into four groups: (i) human capital, (ii) openness and trade, (iii) quality of institutions and (iv) relative level of overall advancement.
Q4. What are the determinants of the adoption lags?
In terms of the determinants of the adoption lags, the authors find that technologies such as PC’s, robots and electricity are complementary to human capital in the sense that human capital reduces the adoption lags for these technologies.
Q5. Why is the speed of diffusion fast for some technologies?
for others, such as electricity and robots, the speed of diffusion has been fast both because of the rapid productivity growth embodied in new vintages and because of the increase in the number of varieties.
Q6. What is the common reason for the differences in levels of output per capita?
bart.hobikn@ny.frb.orgMost cross-country differences in levels of output per capita are due to differences in the level of total factor productivity (TFP), rather than differences in the levels of factor inputs.
Q7. What is the effect of trade openness on the adoption of new technologies?
Trade openness tends to reduce the adoption lags of transportation technologies, such as passenger and cargo aviation as well as sail and motor shipping.
Q8. What is the effect of secondary enrollment on the quality of the adopted vintage?
The marginal effect of secondary enrollment is that a 1 percent increase in secondary enrollment reduces adoption lags such that the quality of the best adopted vintage increases by 0.45%.
Q9. How do the authors avoid the problem with D (s) ,t and ?
The authors avoid the problem withD (s) τ ,t and γ (s) τ by linearizing (36) around the immediate adoption path in which D (s) τ ,t = 0 for all t.
Q10. What is the average detrended R2 for each technology type?
For those technologies with fewer observations, like textiles, steel and robots, the authors obtain a slightly higher detrended R2 and those technologies with a bad fit of the curvature, like radios and TV’s, have a detrended R2 lower than0.5.
Q11. Why does the logistic diffusion curve differ from the empirical literature?
The logistic diffusion curve in their model differs from that in the empirical literature because it results from the optimizing behavior of agents.
Q12. What is the number of workers that use vintage v?
The number ofworkers that use vintage v is then given byL (τ) v,t = S (τ) v,t L (s) τ ,t (9)The corresponding level of output produced with vintage v isY (τ) v,t = CL (s) τ ,t ³ Y (τ) v,t ´³ S (τ) v,t ´1−µ (10)where the constant C depends only on µ.
Q13. What is the effect of new vintages on productivity?
The authors find that for several of their technologies, such as computers, robots, planes, electricity and steel, new vintages embody significantly more productivity than old vintages.
Q14. what is the level of output that worker l would produce using vintage v?
Let the level of output that worker l would produce using vintage v at time t equalY (v) l,t = ³ Z(τ)v e ε (v) l,t ´1−α ³ K (v) l,t ´α , where 0 < α < 1 (1)where K (v) l,t is the number of units of the vintage specific capital good, Z (τ) v is the level of productivity embodied in capital of vintage v and ε (v) l,t is an idiosyncratic, time, worker, and vintage specific productivity shock.
Q15. What factors might affect the intensity with which technologies are used?
Other factors, such as the degree of democracy, might still affect the intensity with which technologies are used, but do not seem to be very important in explaining which technology vintages are being used.
Q16. what is the probability that vintage v satisfies?
The probability that vintage v satisfies (4) and that the associated productivity shock equals ε (v) l,t isπ ³ v, ε(v) l,t´ = 1µ exp" − ε (v) l,tµ − expà − ε (v) l,tµ !Z v0∈V (τ)t exp à − y (τ) v,t − y (τ) v0,t µ ! dv0 # (7)Because there is a continuum of vintages, this is not a proper probability but can better be interpreted as their continuous vintage approximation to the finite number of vintages case.