Estimating the speeds of long-run technological catch-up and growth of total factor productivity for countries

ABSTRACT

This paper constructs a dynamic production frontier function under the framework of a forward-looking rational expectations model, taking the effect of quasi-fixed inputs into account. The sample comprises balanced panel data of 36 countries over the period from 1990 to 2009. Evidence is found that all of the four country groups show the technological catch-up phenomenon in the long run and experience total factor productivity (TFP) growth. Their TFP growth is primarily driven by long-run technical efficiency improvement, followed by technological progress. The Non-G7 and Non-NICs groups are the top two groups in terms of speeds of long-run technological catch-up and rates of TFP growth.

KEYWORDS:

• Dynamic production frontier function
• adjustment speed
• technological catch-up
• total factor productivity

JEL CLASSIFICATIONS:

• C23
• E23
• O47

Acknowledgments

The author is indebted to the Ministry of Science and Technology of the Republic of China, Taiwan, for their financial support of this research under Contract Nos. NSC 99-2410-H-264-006. The author is most grateful for the anonymous referees’ valuable comments and constructive suggestions. Remaining errors are the responsibility of the author.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Solow (1957), Lucas (1988) and Romer (1986) proposed that technological progress is the main reason behind continuous economic growth.

2. Färe et al. (1994) and Ray and Desli (1997) estimated the Malmquist productivity index, and decomposed the change in productivity into technical and efficiency changes. In recent years, Chen, Huang, and Yang (2009) and Oh and Lee (2010) measured the decomposition of Malmquist productivity index by using a metafrontier approach, to investigate the differences in productivity growth across regions or countries operating under different technologies.

3. Kumbhakar and Wang (2005) used a stochastic frontier analysis to estimate world production frontiers econometrically, taking country heterogeneity into account, and obtained estimates of technical change, efficiency improvement and scale-related components of TFP growth. This analysis was used by Garcia, de Souza, and Pires (2008), Shahabinejad, Zare Mehrjerdi, and Yaghoubi (2013) and Arazmuradov, Martini, and Scotti (2014).

4. Banker and Morey (1986) is the first study to evaluate efficiency performances of decision-making units when some of the inputs or outputs are exogenously fixed and the discretionary control, e.g. level of advertising, median income in service area, number of competitors.

5. They exemplified that hospital size may change only after a very long delay, resulting from the number of decision-makers involved in the process and from the financial burden associated with alteration in hospital scale.

6. In order to generate an estimable model, I have to specify an expression for the desired output, namely Y* _it = α_0i + α_0t + y* _it , where y* _it is a output function with unknown form for the time being, α_0i refers to all those factors that are country-specific but time invariant, and the time effect α_0t captures factors common to all countries, such as technological diffusion and technical change.

7. This reduced form is on the basis of assumptions that the stochastic processes generating y* _it and α_0t in which α_0t = ρ₁α_{0t − 1} + ϵ_1t and y* _it = ρ₂ y* _{it − 1} + ϵ_2t where ϵ_1t and ϵ_2t are white noises. More complex specifications are legitimate, but there may be more lags in the equations, which in turn lead to more terms of lagged dependent variables in Equation (4)(4) .

8. For example, α _i = (1 − λ)α_0i , α _t = [(1 − λ)(1 − Dλ)/(1 − Dλρ₁)]α_0t and α₁ = (1 − λ)(1 − Dλ)/(1 − Dλρ₂).

Additional information

Funding

Ministry of Science and Technology of the Republic of China, Taiwan [grant number NSC 99-2410-H-264-006].