Driving forces of technological change: The relation between …moberly1006.weebly.com/uploads/8/9/0/4/890410/technologyandpop… · Driving forces of technological change: The relation

Technological Forecasting & Social Change xxx (2013) xxx–xxx

TFS-17770; No of Pages 14

Contents lists available at SciVerse ScienceDirect

Technological Forecasting & Social Change

Driving forces of technological change:The relation between population growth and technological innovation☆Analysis of the optimal interaction across countries

Mario Coccia ⁎CNR, National Research Council of ItalyBureau d'Economie Théorique et Appliquée, FranceGeorgia Institute of Technology, United States

a r t i c l e i n f o

☆ I thankVittorio Valli (University of Torino, Italy), SecItaly), Patrick Llerena (BETA, France) and John WalshTechnology, USA) for fruitful suggestions and commanonymous referees and Tugrul Daim (Associate Editorfor detailed comments. I also benefited from stimulaseveral colleagues. In addition, I gratefully acknowledgethe CNR – National Research Council of Italy for my vis2011 and at the Georgia Institute of Technology in 20Roemer (Yale University) for his hospitality and theher hospitality at the School of Public Policy (GATprovided excellent research assistance. A preliminaryis included in the Working Paper Series of Ceris-CNRapplies.⁎ CERIS-CNR, Collegio Carlo Alberto, Via Real Collegio, n

(Torino), Italy. Tel.: +39 011 68 24 925; fax: +39 011 68E-mail address: [email protected].

0040-1625/$ – see front matter © 2013 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.techfore.2013.06.001

Please cite this article as: M. Coccia, Drivtechnological innovation, Technol. Forecast

a b s t r a c t

Article history:Received 4 October 2012Received in revised form 30 May 2013Accepted 4 June 2013Available online xxxx

Population growth is one of the major problems facing the world today because it affects thepatterns of sustainable economic growth. The economic phenomenon to be explained in thisstudy is the relation between population growth and technological outputs (patents), focusingon OECD countries, in order to analyze the range of population growth rates favorable tosupport higher technological output, ceteris paribus. The study seems to show the existence ofan inverted-U shaped curve between population growth rate and patents, with an optimalintermediate area in which population growth rates tend to be associated to highertechnological outputs. The public policy consequences of this vital relation are that, onaverage, it is difficult to support an optimal level of technological performance with both alow/negative and a high (higher than 1%) population growth rate-annual % in advancednations. A main consequence is that the estimated relationship of technological outputs vs.population growth is likely to be affected by decreasing returns of technological innovation topopulation growth, because a higher population might decrease research productivity, whichis sensitive to the relation between income per capita and population growth.

© 2013 Elsevier Inc. All rights reserved.

JEL classification:O33J10

Keywords:PopulationPopulation growthTechnological innovationTechnological changeDemographic changePatentsEconomic change

ondo Rolfo (Ceris-CNR,(Georgia Institute ofents, as well as two) of the Journal TF&SCting discussions withfinancial support fromit at Yale University in12. I also thank Johnchair Diana Hicks forECH). Diego Margonversion of this paper

. The usual disclaimer

. 30 - 10024,Moncalieri24 966.

ll rights reserved.

ing forces of technolog. Soc. Change (2013), ht

Societies crystallized in a static demographic distribution,ceteris paribus, tend to survive inadequately and hamper social,technological and economic growth and, as a consequence,their civilization.

1. The problem

Population growth is one of the major problems facing theworld today, associated to resourcemanagement, environmen-tal conservation and restoration (Austin and Brewer [1], p. 47).Since Thomas Robert Malthus’s works in the 1790s, severalscholars have investigated the relation between population andeconomic growth (see Coccia [10] for some historical informa-tion; [5,7,45,61]; etc. Cf. also [43]). Models of endogenousgrowth show a positive association between per-capita incomegrowth and population size [25]. Population growth is also akey element of semi-endogenous economic growth models

ical change: The relation between population growth andtp://dx.doi.org/10.1016/j.techfore.2013.06.001

http://dx.doi.org/10.1016/j.techfore.2013.06.001

mailto:[email protected]

Unlabelled image


http://www.sciencedirect.com/science/journal/00401625

Unlabelled image


2 M. Coccia / Technological Forecasting & Social Change xxx (2013) xxx–xxx

([33]; cf. [24]). 'In its mildest interpretation, semiendogenousgrowth states that economic growth is correlated with thegrowth rate of effort in research and development' (JonesCharles I. as quoted by Strulik [59], p. 131, original emphasis).The common features of thesemodels are decreasing returns ofknowledge in the production of new knowledge and positivepopulation growth [37]. Kremer ([38], p. 688) remarks that, insome models, if research productivity increases with income,technological changewill also increasewith population. Kuznets[40] claims that research productivity increases are associated topopulation growth since a larger population generates moreintensive intellectual contacts: “high population spurs techno-logical change because it increases the number of potentialinventors” (as quoted by Kremer [38], p. 685). In particular,Kuznets ([40], p. 328) states: “Population growth produces anabsolutely larger number of geniuses, talented men, andgenerally gifted contributors to new knowledge whose nativeability would be permitted to mature to effective levels whenthey join the labor force”. As quoted by Strulik ([59], p. 130),Jones Charles I. writes: ‘More people means more Isaac Newtonsand therefore more ideas’. In fact, many inventions and in-novations are demand-driven by larger population, associated toactive demographic change (cf. Boserup [3], p. 5ff). LePoire ([42],p. 1303) points out that: “leadership moves from smaller statesto larger states because larger states have the flexibility todevelop more complex organizational processes and adapt newtechnology”. Instead, Strulik ([59], p. 129) argues that: “long-rungrowth is compatible with a stable population”. In particular,Strulik ([59], p. 131, original emphasis) relaxes the strong tiebetween population growth and economic growth:

First, growth in a general two-sector R&D model is nolonger semiendogenous (driven by exogenous populationgrowth) but fully endogenous (driven by endogenouslyexplained human capital accumulation). Second, growthof an economy is no longer positively tied to populationgrowth. The correlation can be positive or negative; or, asa special intermediate case, economic growth may beindependent of population growth. This result corre-sponds with the empirical findings of a weak, sometimesmixed, and frequently negative correlation betweengrowth rates of population and income per capita.

Some social scientists show that, when population grows,constraints on resources may negatively affect economicgrowth (cf. [63]; [38]). In fact, the contribution of the systemdynamics group “Club di Roma” showed negative scenarios forworldwide economic growth due to high population growthrate and limited natural resources ([44]; cf. [6] for an interestingdiscussion concerning peak oil, human population dynamics,and migration pressures). Despite the existence of severe risksfor the global environment due to population growth, thetypical response to Malthus’s pessimistic argument is thattechnological innovation plays a crucial role, because it cangenerate larger output from the same resources in order tosupport sustainable and continuous economic growth.

Modis ([46], p. 998), on the other hand, notices that:

For decades now, world population increases dramaticallybut this increase comes from the third world and that isnot where most of the innovations are being born. Thereis no evidence that the number of innovations correlates

Please cite this article as: M. Coccia, Driving forces of technologtechnological innovation, Technol. Forecast. Soc. Change (2013), ht

to the number of world population. On the contrary, it canbe seen that Mensch’s bunching of innovations does notcorrelate with any bunching in population number, nordoes the discovery of the stable elements.

Huebner ([28], p. 996) disagrees with this statement byModis [46].

Porter and Stern [49] show a positive relationship betweeninnovative capacity and GDP per capita, which is a maindeterminant of the patterns of technological performance(cf. Kealey [35], p. 240ff).

Besides the wealth of nations, why does technologicalinnovation vary across countries? How does innovationdepend on population growth?

An interesting socio-economic problem ariseswith respect tohow both population and population growth are linked withtechnological innovation [20]. The objective of this paper is toexplain the phenomenon of the interaction between popula-tion growth and technological output in Organisation forEconomic Co-operation and Development (OECD) countries,and to discuss some evidence. In particular, the purpose is toverify whether the relation between population growth andtechnological output exhibits an inverted-U shaped form,which is important to determine the optimal area where therates of population growth are likely to be associated with ahigh level of technological outputs, measured by patentapplications by residents. This form can be similar to one ofthe curves utilized by Kuznets in his analysis of the relationbetween income inequality and per-capita income [61].

The philosophy of science of my research, and as a con-sequence of this study, is based on scientific realism: “scienceaims for and achieves at least approximate truths” (Thagard[60], p. 145; cf. [39]).

The results can provide main findings to detect thecomplex interaction between population growth and tech-nological change, to support adequate public policies ofsustainable economic growth for advanced countries.

The paper is laid out as follows: Section 2 describes thetheoretical framework of the study and presents the workinghypothesis and research design; Section 3 shows the em-pirical evidence by statistical analysis; Section 4 discussesvital theoretical relationships among the observed facts,some critical consequences, and public policy implications.Then, concluding remarks are drawn.

2. Background and methodology

2.1. Theoretical framework and related works

Scholars [2,3,41] argue that population growth, supportedby existing technology, induces people to adopt new techno-logical innovations (cf. also Kremer [38], p. 682ff). Kuznets [40]and Simon [55] claim that higher populations have a higherprobability to create potential inventors: larger populationshave proportionally more individuals with new ideas. Inparticular, Kuznets [40] states that: ‘research productivity percapita increases with population since higher populationallows more intensive intellectual contact and greater special-ization’ (as quoted by Kremer [38], p. 690). Some socialscholars analyze the interaction between demographic andtechnological change, focusing on the role of technological



3M. Coccia / Technological Forecasting & Social Change xxx (2013) xxx–xxx

innovations in the decline of mortality and fertility acrosssocieties (e.g. Boserup [3], p. 184ff; cf. also [52]). On the onehand, the decline of mortality is due to continuous advances inmedical techniques, drugs, and healthcare. On the other hand,the diffusion of innovations for controlling fertility, such as theoral contraceptive pill (cf. “technological contraceptive”, Smart[57], p. 990), has played a critical role in decreasing populationgrowth in advanced countries (see Boserup [3], passim).

The models by Grossman and Helpman [25] – concerningthe theory of endogenous technological change and based onthe assumption of nonrivalry of technology – imply that highpopulation spurs technological change (Kremer [38], p. 681)and that: “total research output increases faster thanproportionally with population due to increases in the sizeof the market” (Kremer [38], p. 690; cf. Young [63], p. 448 andp. 465). Moreover, Kremer ([38], pp. 684–685) notices that:“among technologically separate societies, those with higherpopulation had faster growth rates of technology andpopulation”. Diamond [22] argues that the low technologicallevel of Tasmania is due to its low population. Conversely,Kremer ([38], p. 686) points out that Belgium has a smallerpopulation than Zaire, though the former is richer than thelatter not because of a larger number of inventions generatedwithin the country, but because Belgian human capital andsocio-economic institutions have the capacity to absorb thewidespread technological innovation of the European geo-economic area by a diffusion-orientation policy à la Ergas. Asa matter of fact, Jones ([33], p. 759) claims that growth isgenerated endogenously through R&D and long-run growthrate depends only on parameters that are usually exogenous,including the population growth rate. In particular, Jones ([33],p. 760ff) claims that if the level of resources for R&D is doubled(e.g. number of scientists), then the per-capita growth rate ofoutput should double: “the economy with more researchersshould growth faster” (Jones [33], p.778). Generally speaking,the economic literature highlights a positive correlationbetween per-capita income growth and population size;furthermore, research productivity and innovative capacitycan increase with national income (driven by higher R&Dinvestments; see [11,17,18,35,49]). In addition, when pop-ulation grows, technological change can intensify becausethere are larger intellectual networks and greater speciali-zation, which increase the probability that new ideas,inventions and new innovations can be generated andadopted (demand-induced innovations). In this context,according to Young, some socio-economic mechanismssupport the diffusion of innovations. Among them: conta-gion, “innovations spread much like epidemics”; socialinfluence, “people adopt when enough other people in thegroup have adopted”; and social learning, “people adopt oncethey see enough empirical evidence to convince them thatinnovation is worth adopting” (Young [64], p. 1900).

Boserup [2] and other social scientists argue that peopletend to adopt new technology when population grows.However, this argument is difficult to reconcile with thesimultaneous rise in income and rates of technology change,since increases in income can reduce the effort to invent newtechnologies (Kremer [38], p. 682–683). In fact, according toYoung [63], high population can reduce per-capita incomeand, if research productivity is sensitive to income (cf. Kealey[35], p. 106ff), this may reduce total research output (as


observed by Kremer [38], p. 687ff). Moreover, Modis ([46], p.988) points out that: “Scientific and technological break-throughs generally abound in societies of high income andelevated standard of living. In such societies, the populationsnot only do not grow along exponential patterns but oftenshrink” (cf. Porter and Stern [49], p. 102-104).

The recent economic literature also focuses on the relationbetween immigration (a main determinant of demographicchange) and innovation. Young ([64], p. 1899) claims that:“new ideas, products, and practices take time to diffuse, a factthat is often attributed to some form of heterogeneity amongpotential adopters”. In fact, immigration has positive effects onthe inventive activities of the native population in the US, asshown by Hunt and Gauthier-Loiselle ([32], p. 31): “a 1percentage point increase in immigrant college graduates'population share increases patents per capita by 9–18 percent”.Instead, Kerr and Lincoln ([36], p. 473ff) underline the positivecorrelation between increases in H-1B visas by Indian andChinese immigrants and higher US patenting activities,suggesting the helpful effects of higher caps for immigrantvisas to foster US innovation (cf. also Hunt [30], p. 421ff; [31]).

To sum up, the economic literature remarks that therelation between population growth and technologicalchange can provide interesting and different results, thoughKremer [38] shows that, historically, societies with largerinitial populations and without technological contacts un-dergo faster technological change and population growth.

In short, despite controversial results, several economicstudies confirm a positive long-run association among popula-tion, population growth, and technological change in advancedeconomies. Yet, a main issue of geopolitics concerning patternsof technological innovation is the following:

Which is the range of population growth rates favorable tosupport higher technological output, ceteris paribus?

The next section presents a researchmethodology to answerthis vital economic question, in order to support sustainablepatterns of economic growth in advanced countries.

2.2. Hypothesis and research method

The abduction à la Peirce of this study is the possibility ofquadratic effects concerning the relation between populationgrowth and technological outputs. The hypothetical-deductiveapproach, in line with the philosophy of Hempel [26], is basedon the following hypothesis (HP), which I intend to test:

HP. Technological outputs of countries are negativelyaffected byboth low/negative and highpopulation growth rate.

In particular, to test this HP it is important to determinewhether there exists a range of population growth ratesoptimally supporting the level of technological outputswithin modern advanced countries. The main consequencesof this optimal range of population growth rates are socio-demographic-economic conditions that trigger and amplifythe fruitful interaction between demographic and techno-logical change in order to support general economic growth.The purpose of the present study is to ascertain whetherstatistical evidence, based on empirical observations, supportsthe hypothesis (HP).




Data, explanatory model, and procedures for estimation

□ After a preliminary study [20], the sample selected for thestatistical analysis is represented by 34 countries of theOrganisation for Economic Co-operation and Develop-ment (OECD), with administrations that foster prosperitythrough public policies focused on financial stability andeconomic growth.

□ Non-OECD members are not considered because theydisplay wide socio-economic heterogeneity and in generallower GDP per capita. In addition, the innovation ofcountries with low GDP per capita may not be technolog-ical and may not be patented.

□ The study considers data over a period of twenty years,from 1985 to 2005, concerning about five business cycles(see [48]).

□ The structural indicators of the research (“World Develop-ment Indicators”: [62]) are:– Patent Applications of Residents per million people:

acronym PAR. Innovations are protected by patents,which can indicate the current innovations of coun-tries and also commercially promising inventions(cf. [14,19]). According to Hunt and Gauthier-Loiselle([32], p. 32): “the purpose of studying patents is to gaininsight into technological progress, a driver of produc-tivity growth, and ultimately economic growth”. Inparticular, this study considers patents of residents thatare applications filed through the patent cooperationtreaty procedure or with a national patent office forexclusive rights to an invention — a product or processthat provides a new way of doing something or offers anew technical solution to a problem. Patents as metricsof innovation might have some limits; for instance,transaction costs and disclosure rules vary amongcountries (cf. Coccia [14], pp. 252-253). In addition,Smart [57] suggests finding alternative metrics tomeasure and assess the impact of current innovations,which are more and more virtual and abstract. Howev-er, patents have a positive influence on patterns oftechnological innovation (cf. [47]) and they are themostcommon metrics of innovative output to analyzetechnological performance in modern advanced coun-tries (cf. Steil [58], pp. 3–22).

– Population growth (% annual) at year t: acronymPOPGRW (the rate of growth of midyear populationfromyear t-1 to t, expressed as annual percentage of thepopulation).

– Population density (people per sq.km): acronym POPDENS.The study considers the population as a group of peopleresiding in a country. This population is open andaffected by demographic change, also due to migrationinflows and outflows.

– Fertility rate of the population (%):acronym FER– Household final consumption expenditure (current US$/

1,000,000): acronym HOUSECONS. It is a main proxy ofpopulation growth.

– Electric power consumption (kWh): acronym ELECTRICONS.Same as previous indicator.

Remark. HOUSECONS and ELECTRICONS are plausibleexogenous sources of population variation.


Remark. The linkages among the structural indicators arerepresented in the following chart, which shows criticalpositive feedbacks (+).

The data in the sample have been subjected to horizontaland vertical cleaning, excluding some years for countries withmissing values (e.g. Poland, Italy, etc.) and countries withoutliers. The normal distribution of variables is checked byCurtosi and Skewness coefficients, aswell as by the normal Q-Qplot. As initial variables do not have normal distributions,a logarithmic transformation is carried out to adjust thesedistributions in order to apply parametric estimates; thereby,the specification is based on log-models. Some countries havenegative population growth and it is not possible to apply thelogarithmic transformation; however, this exclusion doesnot affect the reliability of the empirical evidence, as verifiedby further statistical analysis. In addition, as the direction ofcausality between innovative outputs and population growth-annual % can be bidirectional, the estimation of parameters iscarried out by a two-stage least-squares (2SLS) method toremove possible problems of endogeneity. The statisticalanalysis is carried out by using the SPSS statistics software(Statistical Package for the Social Sciences, Version 15 forWindows). The explanatory model based on working equa-tions is represented in Stages I and II of the 2SLS method. Inparticular:

2.3. Stage I of the 2SLS method

In general, population growth tends to be affected by socio-economic and cultural factors, such as consumption (Sheffield[54], pp. 55–56), history of society [51], total population,fertility rate, etc. [20]. An explanatory model considersPopulation growth - annual % (POPGRW) as a function ofPopulation density (POPDENS), Fertility Rate (FER), Household



Table 1Descriptive statistics of OECD member countries.

Arithmetic mean Std. deviation

Patent applications of residents per million people: PAR 312.42 509.83Population growth-annual %: POPGRW 0.64 0.71Population total: POP 34,381,750 53,118,416Population density (people per sq.km): POPDENS 131.94 126.07Fertility rate (%): FER 1.75 0.48Household final consumption expenditure (current US$/1,000,000): HOUSECONS 433,628.87 1,056,590.4Electric power consumption (kWh): ELECTRICONS 243,192,746,031.7 581,352,485,055.4GDP per capita, PPP (constant 2005 international $): GDPPC 24,091.38 8183.78


final consumption expenditure (HOUSECONS), and Electricpower consumption (ELECTRICONS).

The equation is:

LNPOPGRWi;t ¼ β0 þ β1 LNPOPDENSi;t þ β2 LNFERi;t

þβ3LNHOUSECONSi;t þþβ4LNELECTRICONSi;t þ εi;t

ð1Þ

As the underlying data have a time-series structure, theworking equation model is estimated by the Prais-Winstenmethod, which removes autocorrelation. The demographicvariable LNPOP (Table 1) is not inserted into Eq. (1) becauseit can lead to multicollinearity with LNPOPDENS and to anendogenous bias in relation to population growth.

2.4. Stage II of the 2SLS method

The dependent variable is the Ln of patent applications ofresidents (per million people), whereas the fitted value of theEq. (1) is the explicative variable.

The specification is a quadratic model:

LNPARi;t ¼ λ0 þ λ1FIT LNPOPGRWi;t þ λ2FIT LNPOPGRW2i;t þ ui;t

ð2Þ

The square of the annual population growth rate (POPGRW)is introduced to take into account the possibility of non-linear effect, as shown by some economic studies concerningthe relation between population and economic growth (Valliand Saccone [61], pp. 7–9). In addition, the second orderpolynomial Eq. (2) suitably fits scatter data and it isestimated by Ordinary Least Squares (OLS) method.

The estimated relationship Eq. (2) is an objective functionof one (real) variable represented by a polynomial function ofan order higher than the first order. The estimated Eq. (2) is a

Table 2Parametric estimates, results of LNPopulation growth (annual %) on predictors — S

Regression analysis

Predictors Unstandardizedcoefficients

B Std. Error

OECD countries LNHOUSECONS 0.508 0.095LNELECTRICONS −0.580 0.114LNPOPDENS −0.250 0.045LNFER 2.532 0.250(Constant) 7.627 1.967

Note: The Prais–Winsten estimation method is used. Dependent variable: LNPopula


continuous and infinitely differentiable function; it can beanalyzed by differential calculus to find the optimal areas ofpopulation growth rates (around the max value) favorable tosupport higher technological outputs for the sustainableeconomic growth of advanced countries in the long run.

3. Statistical evidence

The descriptive statistics of the sample show a highvariance of some variables (Table 1).

A logarithmic transformation of the variables adjusts thedistributions to achieve normality in order to perform thestatistical analysis and parametric estimates with accuracy.

The bivariate correlation shows that Patent applications ofresidents per million people (PAR) have a negative associationwith population growth rate (r = −0.16, significant at the 0.01level, 2-tailed). The partial correlation between PAR andPOPGRW is also negative: r PAR,POPGRW| GDPPCt-3 = −0.14 (thecontrol variable is GDPPC t-3: GDP per capita, PPP-constant 2005international $ with t-3 time lag, because the past wealth ofnations may affect current technological and demographicchange). If GDPPC t-3 and POPDENS are controlled, the coef-ficient of the partial correlation is r PAR,POPGRW| GDPPCt-3, POPDENS =−0.10: in this case, the association between variables is notinfluenced by the economic wealth of nations and population,ceteris paribus (other things being equal). Instead, the partialcorrelation is positive between PAR and GDP per capita,controlling for POPGRW and POPDENS: rPAR, GDPPC| POPGRW,

POPDENS = +0.59 (cf. [14]).Tables 2 and 3 show the estimated relationships by the

2SLS method. In short, the parametric estimates of theworking equation models are unbiased and the significanceof the coefficients is excellent.

Fig. 1, based on the estimated relationship of Table 3, showsan inverted-U curve fitted to OECD data (concave downwards).

tage I of the 2SLS method.

Model fit

Sig. Adjusted R Square(Std. Error of the Estimate)

Durbin-Watson

0.0 0.223 1.9930.0 (0.557)0.00.00.0

tion growth (annual %).



Table 3Parametric estimates, OLS results of LNPatents of residents (million people) on LNFitPOPGRW – Quadratic Log model – Stage II of the 2SLS method.

Regression analysis Model fit ANOVA

Predictors UnstandardizedCoefficients

T Sig. Adjusted R Square(Std. Error of the Estimate)

Fisher test (Sig.)

B Std. Error

OECD Countries Fit LNPOPGRW −0.904 0.105 −8.573 0.0 0.115 (1.355) 36.891 (0.000)Fit LNPOPGRW2 −0.324 0.045 −7.148 0.0(Constant) 4.829 0.072 67.256 0.0

Note: Dependent variable: LNPAR; The independent variable is FIT for LNPOPGRW from stage I.


This empirical evidence validates the hypothesis (HP). Inparticular, there seems to exist an inverted-U curverepresenting the relation between population and techno-logical growth. In other words, the fitted curve (Fig. 1 basedon the estimated relationship of Table 3) suggests that, onaverage, several OECD countries with low (roughly lowerthan Ln −2 = 0.13%) and high (roughly higher than Ln0 = 1%) population growth rates are characterized by aslower pace of technological outputs. The fitted relationalso seems to show some decreasing returns of technolog-ical outputs to population growth (annual %) above apopulation growth rate of around 1% (annual). In fact,Smart ([57], p. 990ff) claims that “human-initiated inno-vation”, like energy consumption and population growth,is a process that naturally saturates with higher levels ofglobal income and technological intelligence.

Optimal Interme

hLower = 1.549

h*=Ln 1.395 = Max r*= 0.248

h = Fit for LnPOPGRW (Population Gr

y= LnPAR (Patents

per million people)

Curve fitted to OECD data

Fig. 1. Inverted-U curve fitted to OECD member countries data (concave downwavariable) on h = Fit for LNPOPGRW fromautoregression by Prais–Winsten estimationrate and rectangle with broken line indicate the area of rates of the population growth


Of course, the fitted inverted-U curve of this study differsfrom various versions of Kuznets’s curve in relation to boththe variables applied and the positions on the axes.

In order to determine the range of population growth ratesthat optimally supports technological outputs (PAR), themaximum of the estimated relationship-function Eq. (4) iscalculated (note that ui,t is the error term).

Let:

LNPARi;t ¼ λ0 þ λ1FIT LNPOPGRWi;t þ λ2FIT LNPOPGRW2i;t þ ui;t

ð3Þ

LNPARi;t ¼ 4:829−0:904FIT LNPOPGRWi;t

−0:324FIT LNPOPGRW2i;t þ ui;t

ð4Þ

diate Area

hGreater =0.351

owth - % Annual) from stage I

rds): y = LNPatent applications by residents — million people (dependentmethod in Stage I (The fitted relation tends to be non-linear). Note: r* = optimalfavorable to support technological outputs.


image of Fig.�1



If y = LNPAR and h = Fitted LN Population growth-annual %,the necessary condition to maximize Eq. (4) is:

dydh

¼ y’ hð Þ ¼ –0:904–0:648h ¼ 0 ð5Þ

The first derivative equal to 0 gives:

y’ hð Þ ¼ 0 h� ¼ –1:39501 ¼ Max ð6Þ

then Max = Exp (h*) = Population growth (% annual) = r* =0.248 Population growth % annual (see Fig. 1), which tends tomaximize the throughput of technological outputs of advancedcountries. The distribution of r = population growth rate(annual %) is divided in:

− the values below the 15th percentile (i.e. r b 0.2123%) toindicate countries with low average annual populationgrowth rate;

− from the 15th to the 85th percentile (r ≥ 0.2123% to r =1.4208%);

− above the 85th percentile (r > 1.4208%) to indicatecountrieswith high average annual population growth rate.

The results show that, in the intermediate 15th-85thpercentile area [r = 0.2123% to r = 1.4208%], which in-cludes the optimal population growth rate of r* = 0.248

Table 4Technological performance and socio-institutional-demographic-economic conditionsbased on three ranges of percentiles.

P

<15th

r = LOW

Variables (arithmetic mean) r<0.2123

Patent applications of residents per mil-

lion people: PAR 243.1

Population growth (annual %): POPGRW −0.1

Population total: POP 23,678,498.1

Population density (people per sq.km):

POPDENS 132.3

Household final consumption expendi-

ture (current US$/1000000):

HOUSECONS 271,562.0

Electric power consumption (kWh):

ELECTRICONS 136,302,603,896.1

Fertility rate (%): FER 1.4

Standardized freedom house index(1)

(proxy of social innovation ) 93.3

GDP per capita, PPP (constant 2005 in-

ternational $): GDPPC 18,908.9

*In grey: intermediate area (15th–85th percentile) of r = rates of population growthcountries within the rectangle with broken line in Figs. 1–2).(1)The Freedom House Index of liberal democracy measures the level of democratizapolitical pluralism, and the functioning of the government” (Coccia [14], p. 252). Higgovernance (social innovation).


(annual %), r = population growth (annual %) should befavorable to support technological outputs (cf. Fig. 1). In fact,Table 4 shows that PAR per million people in the intermediatearea (i.e. about 373 units) is higher than in countries with low/high population growth. In other words, low (lower than r =0.2123%) and high (higher than r = 1.4208%) POPGRW(annual %) could hamper technological outputs (cf. Table 4).

Remark: y (h) b 0, the second derivative of Eq. (5), isnegative, confirming the decline of PAR as population growsin OECD countries (decreasing returns of technologicalinnovation to population growth).

To sum up, this empirical evidence confirms the epistemo-logical stance through the fitted inverted-U curve betweenpopulation growth rate and technological outputs (patentapplications of residents — million people). In addition, thereexists an intermediate area of population growth rates from0.2123% to 1.4208% (annual) favorable to support technolog-ical output growth (Fig. 1). This theory can satisfy the criteria ofconsilience and simplicity (and analogy with the phenomenadescribed by Kuznets, cf. [61]).

4. Phenomena explained

Long-run technological innovation increase is compatible withmoderate and stable population growth, ceteris paribus.

of OECD countries according to different rates of population growth (annual %)

ercentiles of the distribution of population growth

15th−85th* >85th

r = Rates of population growth (annual%)

r = INTERMEDIATE* r = HIGH

r = 0.2123% to r = 1.4208%* r>1.4208

372.8 158.8

0.7 1.9

38,644,353.8 34,030,795.8

136.7 108.4

561,966.3 130,871.7

339,259,353,277.8 82,694,361,445.8

1.7 2.6

96.0 80.9

26,619.7 18,796.4

favorable to support technological outputs of OECD countries (this includes the

tion: “monitors the existence of political rights in terms of electoral processes,her values indicate more democracy, efficient institutions, and good economic


Unlabelled image


OECD countries without Japan

OPTIMAL

INTERMEDIATE AREA*--Moderate population

growth (Annual%)--Higher population density--Higher GDP per capita--Higher democratizationetc., are

Favorable to highertechnological performance

max r*=0.248%

r*

rHIGH=1.42%rLOW=0.21%

y =PATENTS

OF

RESIDENTSMILLION

PEOPLE

r = RATE OF POPULATION GROWTH ANNUAL %

FITTED CURVEFITTED CURVE

Fig. 2. Fitted inverted-U relation by OLS method (concave downwards) between population growth rate and patents (based on arithmetic mean of data over 1995–2005) and intermediate area of population growth favorable to higher technological outputs (rectangle with broken line). Note 1: r* = optimal rate and rectanglewith broken line indicate the area of rates of population growth favorable to support technological outputs. Note 2: Fig. 2 also includes negative rates ofpopulation growth in comparison to Fig. 1 (that is in double log scale). Some countries with negative growth may have high technological outputs, but theserepresent a temporary condition because, on average, technological outputs tend to be supported in the long run by rates of population growth within theintermediate area [0.21%–1.42%], which is equal to [Ln−1.549; Ln0.351] of Fig. 1 in Logarithmic scale. In fact, only two countries with negative r of populationgrowth display high technological outputs in this period. Note 3: The extremities of the fitted inverted-U curve in Fig. 2 have fewer countries because the analysisis based on OECD countries. If the statistical analysis considers all the countries in the world, several economies with low technological outputs are located in thelower left and lower right areas of the fitted relation, as verified by a preliminary study.

= Population growth near the optimal intermediate area tends to be associated to higher technological outputs

= Population growth distant from the optimal intermedi-ate area tends to be associ-ated to low technological outputs

y=PATENTS

OF

RESIDENTS

MILLION PEOPLE

OPTIMAL INTERMEDIATE AREA

(*AREA)

LOW 0.21% (*AREA) HIGH 1.4208%r = RATE OF POPULATION GROWTH (ANNUAL %)

Area of negative population growth and low technological outputs: Countries with excessive ageing and mi-gration outflows

Area of high population growth and low technological outputs: Countries with high fertility rate, low industrialization and tertiary activities

Moderate population growth rate Rates of population growth within this area tend to support technological out-puts of countries by active and stable demographic change

Area of countries with stagnant technological change

Uncertainty Area

0 r*= 0.248%

Max r*= 0.248%

UncertaintyArea

Fig. 3. Types of countries according to the effects of population growth on technological outputs. Note: r* = optimal rate and rectangle with broken line indicatethe Optimal Area* of population growth rates favorable to support technological outputs. Uncertainty Area: This area includes countries that may generate hightechnological performance but this is a temporary condition because population growth rates lower/higher than the optimal intermediate area do not generatesustainable high levels of technological output in the long run.


Please cite this article as: M. Coccia, Driving forces of technological change: The relation between population growth andtechnological innovation, Technol. Forecast. Soc. Change (2013), http://dx.doi.org/10.1016/j.techfore.2013.06.001

image of Fig.�2

image of Fig.�3



The relation between population growth and technolog-ical outputs is therefore likely to be non-linear.

I can postulate, through evidence, the following socio-demographic laws of technological output:

■ 1. a moderate and steady population growth rate is capableof maintaining technological output increases in advancedcountries (maximizing the throughput of technologicaloutputs), ceteris paribus.

■ 2. high (or low/negative) population growth rates (annual %),ceteris paribus, tend to be negatively associated to patterns oftechnological throughput (measured by patent filing behaviorof countries over time).

■ 3. A main consequence of the socio-demographic laws isdecreasing returns: high rates of population growth canhamper the production of technological outputs due todecreasing returns of technological innovation to populationgrowth (annual %).

Fig. 2 shows the inverted-U curve fitted to OECD data basedon the arithmetic mean of variables; the countries within theintermediate area represented by a rectanglewith a broken lineare characterized by a fruitful interaction between populationgrowth and technological outputs.

Fig. 3 displays a theoretical inverted-U curve with differenttypes of economic behavior by countries in relation topopulation growth rates and associated technologicalperformance.

Evidence shows that there exists an inverted-U curve,similar in shape to Kuznets’s curve, between the populationgrowth rate and technological outputs. The theoreticalmodelsable to explain the main sources of these nonlinearities of theestimated relationship Eq. (4) can be described by Kremer[38] regarding research productivity as a function of techno-logical level (pp. 689–690) and research productivity as afunction of population (pp. 690–692).

An explanation of the inverted-U curve between thepopulation growth rate and technological outputs, and associ-ated socio-demographic laws of technological output, canprovide twomain inductive theoretical implications concerningthe relationship between the observed facts.

□ On the one hand, when the population grows, higherpopulations have a higher probability to create potentialinventors since there are proportionally more individualswith new ideas. In addition, higher population spursdemand-induced inventions and innovations by largerintellectual networks and greater specialization, whichincrease the probability that new ideas and new innova-tions can be introduced and adopted. More innovationssupport income growth, which further favors the

A higher population withmoderate growth (annual %),associated to other factors,can be a vital determinant ofdemand-induced innovativeactivities.

Higher innoverate income which supporlogical (and echange.

Phase 1. Fruitful linkages from demographic factors


development of inventions and innovations. This argumentcan be rooted in the theory of population-push ([2]; cf.[56]). The basic linkages can be schematically representedby the following inductive chart (Phase 1):Phase 1, based on moderate and stable population growth,creates a virtuous circle for technological and demographicchange. The United States of America are a major historicalexample (Steil [58], Chps. 2–3). In particular, a steady-stategrowth, in which the rate of growth of all relevant variablesremains constant over time, can support a Golden Age (JoanRobinson in The Accumulation of Capital as quoted by Coccia[13], p. 71ff).

□ On the other hand, research productivity in a growingphase of the social systemmay increasewith income, but ahigh population can reduce per-capita income; hence, ifresearch productivity is sensitive to income, this canreduce total research output. In addition, the incentives tointroduce new innovations may become smaller whenincome grows as a result of previous innovations. As amatter of fact, the convergence approach by Barro andSala-i-Martin shows that, on average, the economies ofrich countries grow at a slower pace (cf. [13]).In other words, high population might decrease researchproductivity; as a consequence, the evolution of the socialsystem may degenerate into functional inconsistenciesconcerning technological change, as summarized by theinductive chart (Phase 2):This pattern of high population and higher populationgrowth may generate a vicious circle with negative effects(impact) on technological and economic change of thesocio-economic system (decreasing returns).

The explanation and consequences of socio-demographiclaws of technological output are:

o Firstly, the economic behavior of countries with moderatepopulation growth rates (annual %) tends to be favorable tosupport higher technological outputs. In fact, innovativecountries are also likely to be driven by a stable and moderaterate of population growth roughly equal to r =0.248% annual(within the intermediate area [r = 0.2123% to 1.4208%]:Upper central part of the fitted inverted-U curve in Figs. 1 and2). This economic behavior is a specificity of richer countries(cf. [49]), which have a higher average number of patents incomparison to nations with either higher or lower averagepopulation growth (see Table 4). In addition to moderatepopulation growth, other determinants (driving forces) ofhigher technological performance are: an efficient nationalsystem of innovation, driven by high public and privateinvestments in R&D [12,18], a developed industrial

ations gen-growth, ts techno- conomic)

Technological (and eco-nomic) change furthersupports demand-induced innovations,demographic changeand economic growth.

to technological and economic performances.


image of Phase 1


Beyond the optimal area of population growth (roughly 1.42% annual), innovative output and in general scien-tific activity can slow down because higher populationgrowth can reduce incomeper capita and generatestagnant economic change.

Stagnant economic change, associated to other socio-economic factors, is a coefficient of friction for technological outputs andtechnological change.

Low technological (and economic) change hampers economic growth and can also lead to stagnation of technological and eco-nomic change.

Phase 2. Frictional effects in the linkages between population growth and technological performances.


structure, a modern tertiary sector, higher democratizationthat is the basis for good institutions and economicgovernance [14] and higher financial stability of the eco-nomic system [11,18]. In this type of economic behavior bycountries, innovative outputs are demand-driven throughthe growth of income and living standards (cf. [35,46]),supported by fruitful social innovations and good economicgovernance (cf. [14,17,18]). This club of countries, such asthe US and the UK, holds the leadership for what concernstechnological output, based on similar socio-economicsystems.

o Secondly, a main public policy implication of the study isthat, in the long run, it is difficult to support a high level oftechnological outputs with either a low/negative (lower thanr = 0.2123%) or high (higher than r = 1.4208%) averagerate of population growth-annual.In fact, a low/negative population growth (lower than0.2123% annual) can generate excessively ageing countries.This factor leads to less dynamism in starting entrepre-neurial and innovating activities. In addition, the govern-ments of countries with ageing population devote greatereconomic resources to pensions and healthcare servicesrather than to R&D investments, which can supporttechnological outputs [18]. Moreover, older people be-have differently from younger people: they tend toconsume more services and prefer to invest their savingsin low-risk financial assets rather than in high-riskproductive investments stimulating technological inno-vations. The main effect of low/negative populationgrowth can be a stagnant demographic, technological, andeconomic change (e.g. in Italy, Spain, Portugal, etc.).Vice versa, countries with high population growth (higherthan 1.42% annual) – driven by high fertility rate (1.97%annual; cf. [53]) but not supported by a solid industrialstructure – tend to have a lower GDP per capita andstagnant economic change. Educated people have difficultyin finding adequate employment, and this hampers thedevelopment of human capital, which may support tech-nological innovation. High unemployment rates can alsospur intensivemigratory outflows towards richer countries,and this represents an economic, human, and scientific lossfor these nations. In general, these countries devote limitedfinancial resources to R&D intensity [12,17,18]. Said limitedresources, associated to imperfect capital markets, slowdown real investments, entrepreneurship, patenting activ-ities and, as a consequence, patterns of technologicalinnovation and economic growth (e.g. in Mexico, Turkey,etc.).


These explanations show that the socio-demographiclaws of technological output appear to be true and representmain driving forces of technological innovation patterns inmodern advanced countries.

Table 5 illustrates the main socio-economic structuralcharacteristics of countries according to their location inFigs. 1–3. Table 5 might provide critical information about themain socio-institutional drivers of high technological perfor-mance associated to moderate population growth. In fact, Set 1includes predominantly protestant countries, whereas Set 2 ispredominant catholic; Set 1 countries have been independentfor longer and display a higher index of democratization; hence,they have better institutions and economic governance [14]. Set1 is also characterized by longer administration stability(duration) due a higher number of seats assigned to the largestparty in each legislature. In addition, countries with highertechnological performance (Set 1) have a lower party fraction-alization index, a higher federalism index and fiscal decentral-ization in comparison to low performing countries (Sets 2–3).Another characteristic of high technological performers ishigher religious fractionalization. In short, better performingcountries are richer, with more democratization, longer dura-tion of administration (institutional stability), higher federalismand fiscal decentralization, and predominance of the protestantreligion with greater religious fractionalization. Hence, thesesocio-institutional determinants generate a higher indexof humandevelopment,which better supports the social systemand may spur higher technological and economic performance.

4.1. General remarks on empirical analyses

The econometric model confirms, ceteris paribus, that amoderate population growth rate, within the intermediatearea of the inverted-U shape (about 0.248% annual), associ-ated to efficient institutions and favorable socio-culturalfactors, seems to support technological outputs (patents permillion people). The relationship between the residuals ofEq. (2) and country population growth, measured by thecoefficient of correlation, is roughly r = 0.005. However, theresiduals of the model have a great amount of variance whichought to be explained (Table 3). This strongly suggests thatthe estimated relation between population growth andtechnological outputs is also affected by other hidden factors,influencing both the socio-economic-demographic structureof countries and their patterns of technological innovation. Infact, for instance, Spain and the UK have a low average rate ofpopulation growth-% annual (0.3 vs. 0.6), but Spain has anannual average of about 57 patents per million people,


image of Phase 2


Table 5Structural characteristics of countries considering their rate of population growth (Annual %) and technological performance (cf. Figs. 1–3).

Typology of economic behavior

VARIABLES (Arithmetic mean) Countries withintheoptimal intermediatearea of the fittedinverted-U curve

Countries within thestagnanttechnologicalchange area (belowthe intermediatearea)

Countries close to theextremities of the fittedinverted-U curve (high/lowrates of population growth)

Set 1 Set 2 Set 3

r = POPULATION GROWTH-% ANNUAL Moderate(0.21 b r b 1.42)

Moderate(0.21 b r b 1.42)

Low/High(r b 0.21 or r > 1.42)

TECHNOLOGICALPERFORMANCE (measured by patents of residents per million people: PAR) HIGH (PAR > 350) LOW LOWPredominantly Protestant nations (CIA World Factbook): 0 = low;1 = high 0.62 0.14 0.33Predominantly Catholic nations (CIA World Factbook): 0 = low;1 = high 0.38 0.86 0.33Predominantly Muslim nations (CIA World Factbook): 0 = low;1 = high 0.00 0.00 0.33Period of Independence in years (2013 — year of independence) 430 287 105Freedom House Index Liberal Democracy standardized scale 100 pts 97.8 95.9 73.8Press freedom (Freedom House) 86 83 57Number of years since the last regime transition 78 53 10Age in years of current regime (Polity IV 2003) 91 54 10How long has executive party been in office? 88 42 26Number of Seats, Largest Party in Legislature (Banks) 140 126 124Party Fractionalization Index (Banks) 71.4 70.1 75.0Herfindahl government index (DPI) 0.58 0.72 0.56Federalism index (Gerring-Thacker) 3.77 3.43 3.67Fiscal decentralization 68 nations (Schneider) 0.55 0.47 0.41Ethnic fractionalization (combined linguistic and racial): 0 = low;1 = high 0.2255 0.3086 0.4560Linguistic fractionalization (Alesina): 0 = low;1 = high 0.2028 0.3173 0.2890Religious fractionalization (Alesina): 0 = low;1 = high 0.4579 0.4163 0.2277Human development index 2002 (UNDP 2004) 0.929 0.924 0.802Gross National Product Per Capita -Market Prices-(Banks) - $ 26,918 17,474 3,917Extrajudicial killings (CIRI): close to 1 = practiced infrequently;close to 2 = have not occurred

2.00 1.67 1.67

Note: Set 1: Austria, Denmark, Finland, France, Germany, Hungary, Luxembourg, New Zealand, Norway, Sweden, Switzerland, United Kingdom, United States,Japan; Set 2: Australia, Belgium, Canada, Italy, Portugal, Slovenia, Spain; Set 3: Estonia, Mexico, Turkey; (cf. Figs. 1–3). Other countries are not included because ofmissing data.Sources: See these sources for details on coding, measurement and definitions: CIA World Factbook: www.cia.gov; UNDP Human Development Report: www.UNDP.org; WVS World Values Survey: www.worldvaluessurvey.org; World Bank: www.worldbank.org; Freedom House: www.freedomhouse.org.


whereas the UK has an annual average of roughly 334 patents(Fig. 2). The ratio of the arithmetic mean of observed valuesLn(PAR/POPGRW) is 4.56 for Spain and 7.01 for the UK: thepattern of technological outputs over population growth(annual %) is much bigger in the UK than in Spain. Hence, thedifferent patterns of technological innovation of theseadvanced nations may be associated to other complexsocio-economic forces, such as: efficient national system ofinnovation [18]; fruitful university, industry and governmentlinkages (Triple Helix); effective institutions based on higherlevels of democracy (see Tables 4 and 5; cf. [14]); higher R&Dspending by governments and business enterprises [17];active industrial structure and service sector [18]; fertilizinghigh-skilled immigration inflows; etc.

In brief, institutions, social innovations (democratization),cultural factors and other socio-economic attitudes differacross countries and tend to affect demographic and economic

Higher population, moderate and stable population growth and other socio-institutional-economic factors

Higher demandgoods and serv

Fig. 4. Population growth push for demand-indu


trends. As a consequence, they might generate a great varietyof economic and technological performances, respectively.

5. Concluding remarks

High population and a marked acceleration or decline indemographic growth, ceteris paribus, might hinder increases intechnological output.

In general, the statistical evidence seems to confirm thathigher technological outputs can also be explained bymoderate population growth within the intermediate areaof the inverted-U curve fitted to OECD data. This controlledpopulation growth, over time, is one of the vital conditionsto support technological output, in combination with othersocio-economic drivers such as: higher democratization,good economic governance, stability in the duration of

for ices

Hence, demand-induced inven-tions (patents) and innovations that support economic growth and technological progress

ced technological and economic change.


http://www.cia.gov

http://www.UNDP.org

http://www.UNDP.org

http://www.worldvaluessurvey.org

http://www.worldbank.org

http://www.freedomhouse.org

image of Fig.�4



administration, higher GDP per capita, etc. (see Table 5).Outside the optimal intermediate area between populationgrowth and technological output there is a factor of disturbanceto the increase of technological output, due to a markedacceleration or decline in population growth dynamics, whichmight affect economic structural indicators underlying tech-nological change [8,9,11,19].

As a matter of fact, the main findings of the study are:

■ The estimated relationship shows that a moderatepopulation growth rate (annual %) is likely to be a maindeterminant, along with other socio-economic factors, oftechnological outputs. This result can be due to thefollowing linkages: higher population growth leads togreater demand for goods and services, which leads tomore innovation and economic growth (demand-driven).This theoretical relationship between observed facts canbe schematically summarized in Fig. 4.

■ There exists an inverted-U curve (concave downwards),similar in shape to Kuznets’s curve, between the popula-tion growth rate and technological outputs.

■ There is an intermediate area of population growth rates,within the fitted inverted-U curve [r = 0.2123% to 1.4208%annual], which tends to be associated to higher technolog-ical outputs (necessary but not sufficient condition).

■ Beyond the intermediate area of population growth rates(annual %), there can be decreasing returns of technolog-ical outputs to population growth.

The presence of an inverted fitted U-shaped curve betweenpopulation growth rate and patent applications by resident(million people) seems to show that a moderate populationgrowth rate is capable of maintaining a sustainable increaseof technological outputs (the first socio-demographic lawof technological innovation). In other words, a steady-stategrowth of the population, within the intermediate area of thefitted inverted-U curve, may reduce problems associated to theextremities: i.e. excessive population growth (demographicboost) or inadequate population growth (ageing population).These vital results have several consequences.

Although the behavior of countries in relation to populationgrowth and technological outputs varies greatly, World Bankand CIA Factbook 2011 data show that the worldwide averagepopulation growth rate was 1.19% in the 2002–2008 periodand 1.0% in 2011. Hence, it seems that, on average, the worldpopulation has gone beyond the optimal condition (cf. Fig. 1)and has entered a phase of increasing population growth rateswith decreasing returns as far as technological innovation (i.e.patent outputs) is concerned. According toModis ([46], p. 988)the innovation-producing individuals as percentage of theworld population are decreasing and, when developingcountries eventually reach high standards of living with highincome, they might support world innovation.

This study can provide a further interpretation of Huebner’sapproach ([27], p. 985; [29]) concerning the declining trend ofinnovation rate worldwide. In particular, this may also be due toa high worldwide population growth which triggers resourcemanagement problems, affects the dynamics of socio-economicsystems and, as a consequence, the patterns of technologicalinnovations (cf. [46,57]). This resultmay be valuable informationtowards designing adequate public policies to support economicgrowth.


More specifically, a main policy implication of this study isthat it is difficult to support high technological outputs witheither a low/negative (lower than 0.2123% annual) or high(higher than 1.4208% annual) population growth rate. A fruitfulpolitical economy of growth for advanced countries should alsofocus on demographic change, based on a controlled andmoderate population growth rate, within the optimal interme-diate area of the fitted inverted-U relation, which is a necessarybut not sufficient condition to spur technological outputs andeconomic growth. In fact, this necessary condition should bereinforced by sufficient conditions (drivers) to support techno-logical outputs, innovations and new firms, such as: higher rateof R&D intensity [12,17,18]; higher level of democracy (socialinnovation, cf. Tables 4 and 5) supporting good economicgovernance and institutions [14]; efficient national systems ofinnovation and effective linkages in the Triple Helix mechanism[18]; and highly skilled immigration inflows (“Brain Gain”, cf.[4,23,34]). As a matter of fact, the demographic prosperity ofpopulations is also driven, inter alia, by highly skilled immigra-tion, which can be a cross-fertilization factor for supportingtechnological outputs (cf. Hunt and Gauthier-Loiselle [32],p. 31ff; [30,31,36]). In particular, higher technological outputs(patents) tend to be associated to stable population growthsupported by capital accumulation and highly skilled humancapital accumulation, which generate the intellectual dividend.These factors fertilize the socio-economic setting and have defacto fruitful effects on the patterns of technological innovation(cf. Coccia [14], pp. 260–261). In short, these criticalsocio-economic factors are also driving forces for sustainablepatterns of technological innovation and economic growth inthe long run (cf. [15,16]; see [58]).

Hunt and Gauthier-Loiselle [32], Hunt [30], and Kerr andLincoln [36] argue that policies supporting the immigrationof scientists and engineers (highly skilled immigrationinflows) should be pursued in order to support innovationand total factor productivity, as happens in the US.

In general, societies crystallized in a static demographicdistribution, ceteris paribus-other things being equal, tend tosurvive inadequately and hamper social, technological andeconomic growth and, as a consequence, their civilization.

These conclusions are, of course, tentative. Nevertheless,the relationship between population growth and techno-logical innovation generates intertwined links. Boserup([3], p. 5ff) claims that the causality between technologicaland demographic change can be bidirectional, whereasKremer [38] argues that the relation between populationand technological change, as well as the effects of policiesbased on population growth to support technologicalinnovation, can be ambiguous. In fact, social systems arecurrently becoming more and more complex and it mightprove difficult to identify cause and effect of severalrelations. As a matter of fact, patterns of technologicalinnovation interact, in complex ways, with biological,social, and cultural factors [21]. This study has simplifiedthe issue by analyzing technological output in relation topopulation growth, though it is affected by severalsocio-economic factors. In addition, the country trends areanalyzed over 1985-2005, which is a rather limited period.Therefore, caution is needed to generalize predictionsbecause the results may not be universal but restricted tospatial and temporal dynamics of advanced countries.




Moreover, patents as metrics of technological outputsshould be integrated with other new technometric indicators,which are difficult to identify because modern innovations arebecoming of “higher-order, virtual, abstract” (Huebner [28],p. 998; Smart [57], p. 992 and p. 994).

In general, socio-economic laws are problematic when weknow that other things are often not equal. The results of thisstudy have tried to provide, through empirical evidence, averisimilitude or degree of closeness to true economic facts(realism). However, socio-economic laws have an infinite setof spatial-temporal determinants and consequences and, inparticular, no socio-economic rule will be true in all situations[50]. Although other socio-demographic-economic factors areimportant for a systematic analysis of thismain relation betweenpopulation growth and technological change across countries,the partial model discussed here, focusing on two criticalvariables – population growth (annual %) and technologicaloutputs (patents) – provides an interesting theory (representedby socio-demographic laws of technological outputs) based onan important empirical relation, which is likely to be non-linear.We should be aware of the provisional nature of socio-economicscience associated to the perpetual susceptibility to error, whichspurs continuous revisions that are the foundations for break-throughs (advances) in scientific knowledge. Hence, muchmoredetailed research is needed, especially into the relations betweenpopulation growth and technological innovation. Future re-search should focus on more comprehensive models and casestudies on critical countries from a historical perspective toexplore how the association between demographic and techno-logical variables can change over time.

References

[1] A.L. Austin, J.W. Brewer, World population growth and relatedtechnical problems, Technol. Forecast. Soc. Change 3 (1) (1971) 23–49.

[2] E. Boserup, The Conditions of Agricultural Growth: The Economics ofAgrarian Change under Population Pressure, Aldine: Chicago, Allen &Unwin, London, 1965.

[3] E. Boserup, Population and Technological change: A study of long-termtrends, University of Chicago Press, Chicago, 1981.

[4] T. Boeri, H. Brücker, F. Docquier, H. Rapoport, Brain drain and braingain, The Global Competition to Attract High-Skilled Migrants, OxfordUniversity Press, Oxford, 2012.

[5] A. Bucci, Population growth in a model of economic growth withhuman capital accumulation and horizontal R&D, J. Macroecon. 30 (3)(2008) 1124–1147.

[6] C.J. Campbell, Petroleum and people, Popul. Environ. 24 (2) (2002)193–207.

[7] M. Chamon, M. Kremer, Economic transformation, population growth andthe long-run world income distribution, J. Int. Econ. 79 (1) (2009) 20–30.

[8] M. Coccia, Technometrics: origins, historical evolution and newdirection, Technol. Forecast. Soc. Change 72 (8) (2005) 944–979.

[9] M. Coccia, Measuring intensity of technological change: the seismicapproach, Technol. Forecast. Soc. Change 72 (2) (2005) 117–144.

[10] M. Coccia, Macchine, lavoro e accrescimento della ricchezza: Riflessionisul progresso tecnico, occupazione e sviluppo economico nel pensieroeconomico del Settecento e Ottocento, Working Paper Series Ceris-Cnr,vol. IX, 2007, (n. 1, ISSN (Print): 1591-0709).

[11] M. Coccia, A new taxonomy of country performance and risk based oneconomic and technological indicators, J. Appl. Econ. 10 (1) (2007) 29–42.

[12] M. Coccia, What is the optimal rate of R&D investment to maximizeproductivity growth? Technol. Forecast. Soc. Change 76 (3) (2009)433–446.

[13] M. Coccia, A new approach for measuring and analyzing patterns ofregional economic growth: empirical analysis in Italy, Ital. J. Reg. Sci. 8(2) (2009) 71–95.

[14] M. Coccia, Democratization is the driving force for technological andeconomic change, Technol. Forecast. Soc. Change 77 (2) (2010) 248–264.


[15] M. Coccia, The asymmetric path of economic long waves, Technol.Forecast. Soc. Change 77 (5) (2010) 730–738.

[16] M. Coccia, Foresight of technological determinants and primary energyresources of future economic long waves, Int. J. Foresight Innov. Policy6 (4) (2010) 225–232.

[17] M. Coccia, The interaction between public and private R&D expendi-ture and national productivity, Prometheus Crit. Stud. Innov. 29 (2)(2011) 121–130.

[18] M. Coccia, Political economy of R&D to support the modern compet-itiveness of nations, Technovation 32 (6) (2012) 329–414.

[19] M. Coccia, Driving forces of technological change in medicine: Radicalinnovations induced by side effects and their impact on society andhealthcare, Technol. Soc. 34 (4) (2012) 271–283.

[20] M. Coccia, Population and technological innovation: the optimal interac-tion across modern countries, Working Paper Ceris del ConsiglioNazionale delle Ricerche, AnnoXV, 2013, (n. 7 - ISSN (Print): 1591-0709).

[21] T.C. Devezas, J.T. Corredine, The biological determinants of long wavebehavior in socioeconomic growth and development, Technol. Fore-cast. Soc. Change 68 (1) (2001) 1–57.

[22] J. Diamond, Ten thousand years of solitude, Discover 14 (3) (1993)48–57, (Mar).

[23] F. Docquier, H. Rapoport, Globalization, brain drain, and development,J. Econ. Lit. 50 (3) (2012) 681–730.

[24] I. Ehrlich, F. Lui, The problem of population and growth: a review of theliterature from Malthus to contemporary models of endogenouspopulation and endogenous growth, J. Econ. Dyn. Control. 21 (1) (1997)205–242.

[25] G. Grossman, E. Helpman, Innovation and Growth in the GlobalEconomy, MIT Press, Cambridge, 1991.

[26] C.G. Hempel, Aspects of Scientific Explanation, The Free Press, NY, 1965.[27] J. Huebner, A possible declining trend for worldwide innovation,

Technol. Forecast. Soc. Change 72 (8) (2005) 980–986.[28] J. Huebner, Response by Jonathan Huebner, Technol. Forecast. Soc.

Change 72 (8) (2005) 995–1000.[29] J. Huebner, Discussion of Coates' commentary, Technol. Forecast. Soc.

Change 73 (7) (2006) 906–912.[30] J. Hunt, Which immigrants are most innovative and entrepreneurial?

Distinctions by entry visa, J. Labor Econ. 29 (3) (2011) 417–457.[31] J. Hunt, Are immigrants the best and brightest U.S. engineers? NBER

Working Paper, 2013, (n. 18696 Issued in January 2013).[32] J. Hunt, M. Gauthier-Loiselle, How much does immigration boost

innovation? Am. Econ. J. Macroecon. 2 (2) (2010) 31–56.[33] C.I. Jones, R&D-based models of economic growth, J. Polit. Econ. 103 (4)

(1995) 759–784.[34] C.I. Jones, P.M. Romer, TheNewKaldor facts ideas, institutions, population,

and human capital, Am. Econ. J. Macroecon. 2 (1) (2010) 224–245.[35] T. Kealey, The Economic Laws of Scientific Research, Macmillan, 1996.[36] W. Kerr, W. Lincoln, The supply side of innovation: H-1B visa reforms

and U.S. ethnic invention, J. Labor Econ. 28 (3) (2010) 473–508, (NBERWorking Paper 15768, HBS Working Paper 09-005).

[37] S.S. Kortum, Research, patenting, and technological change, Econometrica65 (6) (1997) 1389–1419.

[38] M. Kremer, Population growth and technological change: one millionB.C. to 1990, Q. J. Econ. 108 (3) (1993) 681–716.

[39] A. Kukla, Studies in Scientific Realism, Oxford University Press, NY,1998.

[40] S. Kuznets, Population change and aggregate output, Demographic andEconomic Change in Developed Countries, Special conference series n. 11Universities-National Bureau for Economic Research, Princeton Uni-versity Press, Princeton, 1960, pp. 324–340, (Reprinted in EconomicGrowth and Structure: Selected Essays. New York: W.W. Norton & Co.,1965, pp. 123-141).

[41] R. Lee, Induced population growth and induced technological progress:their interaction in the accelerating phase, Math. Popul. Stud. 1 (3)(1988) 265–288.

[42] D.J. LePoire, Long-term population, productivity, and energy use trendsin the sequence of leading capitalist nations, Technol. Forecast. Soc.Change 77 (8) (2010) 1303–1310.

[43] C. Marchetti, P.S. Meyer, J.H. Ausubel, Human population dynamicsrevisited with the logistic model: how much can be modelled andpredicted? Technol. Forecast. Soc. Change 52 (1) (1996) 1–30.

[44] D.H. Meadows, D.L. Meadows, J. Randers, W.W. Behrens III, The Limitsto Growth, Universe Books, New York, 1972.

[45] L.C.M. Miranda, C.A.S. Lima, On the logistic modeling and forecasting ofevolutionary processes: application to human population dynamics,Technol. Forecast. Soc. Change 77 (5) (2010) 699–711.

[46] T. Modis, Discussion of Huebner article, Technol. Forecast. Soc. Change72 (8) (2005) 987–1000.

[47] P. Moser, Patents and innovation: evidence from economic history,J. Econ. Perspect. 27 (1) (2013) 23–44.


http://refhub.elsevier.com/S0040-1625(13)00126-1/rf0005





















































































































[48] NBER, National Bureau of Economic Research, http://www.nber.org/cycles/cyclesmain.html 2013, (accessed 4th March, 2013).

[49] M.E. Porter, S. Stern, National innovative capacity, The Global Competi-tiveness Report 2001–2002, Oxford University Press, New York, 2001,pp. 102–120, (Ch. 2.2).

[50] K. Popper, The Logic of Scientific Discovery, Hutchinson, London, 1959.[51] G.G. Ross, Nonlinear perturbations of difference equations in popu-

lation dynamics, Technol. Forecast. Soc. Change 15 (4) (1979)247–257.

[52] W.W. Rostow, Population in the twenty-first century: the limitedhorizon of public policy, Technol. Forecast. Soc. Change 67 (1) (2001)19–34.

[53] Y. Sato, K. Tabata, K. Yamamoto, Technological progress, incomeinequality, and fertility, J. Popul. Econ. 21 (1) (2008) 135–157.

[54] J. Sheffield, World population growth and the role of annual energy useper capita, Technol. Forecast. Soc. Change 59 (1) (1998) 55–87.

[55] J.L. Simon, The Economics of Population Growth, Princeton UniversityPress, Princeton, 1977.

[56] J.L. Simon, The Ultimate Resource, Princeton University Press, Princeton,1981.

[57] J. Smart, Comments by John Smart, Technol. Forecast. Soc. Change 72(8) (2005) 988–995.

[58] B. Steil, D.G. Victor, R.R. Nelson, Technological Innovation andEconomic Performance, Princeton University Press, Princeton, 2002.

[59] H. Strulik, The role of human capital and population growth in R&D-basedmodels of economic growth, Rev. Int. Econ. 13 (1) (2005) 129–145.


[60] P. Thagard, Computational Philosophy of Science, The MIT Press,Cambridge, MA (USA), 1988.

[61] V. Valli, D. Saccone, Economic development and population growth: aninverted-U shaped curve? Working Paper Series, Departments ofEconomics Torino, Working Paper n. 5/2011, 2039-4004, 2011.

[62] World Bank, World Development Indicators on CD-ROM, The WorldBank, Washington D.C, 2008.

[63] A. Young, Invention and bounded learning by doing, J. Polit. Econ. 101(3) (1993) 443–472.

[64] H.P. Young, Innovation diffusion in heterogeneous populations:contagion, social influence, and social learning, Am. Econ. Rev. 99 (5)(2009) 1899–1924.

Mario Coccia is a economist at the National Research Council of Italy andvisiting professor of Industrial Organization at the University of PiemonteOrientale (Italy). He has been research fellow at the Max Planck Institute ofEconomics (Germany), visiting professor at the Polytechnics of Torino(Italy), visiting researcher at the Georgia Institute of Technology (USA), YaleUniversity (USA), University of Maryland (College Park, USA), BETA (Bureaud’Économie Théorique et Appliquée, Strasbourg-France) and Institute forScience and Technology Studies (University of Bielefeld, Germany). He haswritten extensively on Economics of Innovation and Science; his researchpublications include more than two hundred papers in several socio-economic disciplines.


http://www.nber.org/cycles/cyclesmain.html

http://www.nber.org/cycles/cyclesmain.html






































Documents

Driving forces of technological change: The relation between …moberly1006.weebly.com/uploads/8/9/0/4/890410/technologyandpop… · Driving forces of technological change: The relation