Non‐Scale Endogenous Growth with R&D and Human Capital

Published date01 November 2016
Date01 November 2016
Creina Day*
This paper examines the conditions under which increasing knowledge, encapsu-
lated in ideas for new technology through R&D and embodied in human capital
through education, sustains economic growth. A general model is developed
where, consistent with recent literature, growth is non-scale (not increasing in
population size) and endogenous (generated by factors within R&D and educa-
tion). Recent models feature the counterfactual assumption of constant returns
to existing knowledge and restrict the substitutability of inputs within R&D and
education. We find that non-scale endogenous growth is possible under less strin-
gent conditions. The findings reconcile sustained economic growth with evidence
of diminishing marginal returns in education and R&D, which suggests an
ambiguous role for R&D policy.
Increasing knowledge, encapsulated in ideas for new technology and embodied
in individuals as human capital, has driven unprecedented growth in living
standards for more than a century (Jones and Romer, 2010). A thriving new
growth literature (Funke and Strulik, 2000; Dalgaard and Kreiner, 2001;
Bucci, 2003; Strulik, 2005; Boonprakaikawe and Tournemaine, 2006; Papa-
georgiou and Perez-Sebastian, 2006) predicts that research and development
(R&D) and human capital accumulation will continue to sustain economic
growth under certain conditions. Specifically, recent models draw on constant
marginal returns in knowledge accumulation to predict endogenous growth,
which means long-run growth in the economy is generated within the model
rather than by growth in an exogenous factor, in this case, population.
prediction has relevance for OECD countries facing zero population growth.
This paper contributes to the literature by establishing less stringent
*Australian National University
According to the Concise Oxford Dictionary, ‘endogenous’ means ‘having an internal
cause or origin’. Growth is endogenous in this sense because the processes of R&D and
human capital accumulation, which are internal to recent models, cause technological change
and increasing human capital which sustain long-run economic growth.
Scottish Journal of Political Economy, DOI: 10.1111/sjpe.12115, Vol. 63, No. 5, November 2016
©2016 Scottish Economic Society.
conditions for endogenous growth, thereby opening up this important field to
future research.
To shed light on the origins of conditions for endogenous growth in the
recent literature, it is convenient to classify three broad classes of models.
First generation models of endogenous growth focus on the accumulation of
ideas (Romer, 1990; Aghion and Howitt, 1992) or embodied human capital
(Lucas, 1988; Rebelo, 1991) as alternative explanations of long-run growth.
Constant marginal returns to extant knowledge in the production of new
knowledge allow the R&D or education sector to be the engine of growth.
Because the marginal product of knowledge does not decline as new knowl-
edge is added over time, R&D or education can sustain constant growth in
technology or human capital respectively. In this way, the models generate
endogenous growth. However, these models predict strong scale effects,
whereby an increase in population size permanently raises economic growth
for which there is conflicting evidence.
Second generation models focus on the accumulation of ideas through
R&D (Jones, 1995b; Kortum, 1997; Segerstrom, 1998) and remove strong
scale effects by introducing diminishing marginal returns to extant ideas in
R&D. This assumption implies that R&D requires increasing effort, in the
form of a growing population of researchers, to sustain a constant rate of
technological change. This is the intuition behind semi-endogenous growth.
Jones (1995b) coined the phrase, which means that technological change is
endogenous in the sense that it is generated by R&D, but would eventually
cease without exogenous population growth. Long-run growth is policy invari-
ant and non-scale, meaning it does not increase with population size.
Third generation models consolidate the first generation R&D-based
growth and human capital investment-led models of Romer (1990) and Lucas
(1988), respectively, thereby endogenizing two types of knowledge, ideas and
human capital. These models (Funke and Strulik, 2000; Dalgaard and
Kreiner, 2001; Bucci, 2003; Strulik, 2005; Boonprakaikawe and Tournemaine,
2006; Papageorgiou and Perez-Sebastian, 2006) predict non-scale endogenous
growth, meaning long-run economic growth does not increase with population
size and is generated within the model by R&D and human capital accumula-
In these models, investment in human capital fuels R&D, the engine of
technological progress that sustains economic growth. The modelling is often
It is in fact a slight misrepresentation of Romer (1990) and Aghion and Howitt (1992) to
say that economic growth is proportional to population. Both models predict that economic
growth is proportional to human capital employed in R&D.
Jones (1995a) observes a constant growth path in US per capita output over the past
century, despite an exponential increase in population. However, Kremer (1993) observes
that a region’s technological rank reflects its initial population size and there is evidence that
economic growth is increasing in the size of the market which reflects initial population size
(Sokoloff, 1988; Ciccone and Hall, 1996; Alesina et al., 2000).
These models exhibit weak scale effects, whereby an increase in population size perma-
nently raises the level of per capita output.
An alternative branch of the literature consolidates the variety and quality ladder R&D
models of Romer (1990) and Aghion and Howitt (1992). The properties of these models are
analysed by Li (2000).
Scottish Journal of Political Economy
©2016 Scottish Economic Society

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT