Internet and Growth in Developing Countries Eric Bartelsman 1 1 Vrije Universiteit Amsterdam, Tinbergen Institute, IZA BMZ WDR Berlin Nov. 6, 2014
Internet and Sustainable Growth Historical Impact of Business use of ICT Drivers of Adoption of ICT by Business Unleashing future potential of ICT
Evidence from Advanced Economies Multi-country (EU), Firm-level, linked panel production and ICT Exploration of ICT adoption, use, impact at firm and aggregate level Summary of methods and findings in Bartelsman et al. (2014)
Productivity impact of Broadband
ICT and Productivity For the average firm, ICT usage does not seem to increase productivity significantly However, industry aggregate productivity is correlated significantly with Broadband usage This is not a direct causal relationship. ICT use may increase the possibility of a high productivity outcome, similar to paying a fixed cost in the Hopenhayn model, or the Bartelsman et al. (2014) version. The next part of the causal chain is a shift of resources to the successful firm ICT usage is indeed seen to be correlated with an increase in the variance of productivity (both in cross section and in expectation over time.
Growth 1996-2008, ICT-intensive vs non-intensive Firms
Output and Employment: ICT intensive vs non-intensive firms The chart shows that in most country/industry observations: Relative to non-ict intensive firms, ICT intensive firms have higher i) productivity growth, ii) employment growth, and iii) output growth. These data are from the long panel, output and employment by country, industry, year, and ICT. Datapoint gives difference in estimated y and e growth coefficients in regression of y or e on time trend, comparing ICT and non-ict firms.
ICT use and dispersion: Industry-level evidence Table : Std. Dev. of firm-level productivity distribution regressed on Broadband intensity Levels γ 0.47 (5.02) First-differences.28 (2.59) R 2 0.52 0.03 D.F. 1180 1021 Fixed effects ctry, ind, time ctry, ind, time σ c,i,t = α + γbbi c,i,t + FE + ε c,i,t FE: country, industry, time fixed effects Source: ESSNet, Micro-moments database, v4.0
ICT Use and Dispersion: Firm-level evidence Table : Output Growth Dispersion by ICT intensity Time Series Cross Section Country ICT=0 ICT=1 ALL ICT=1 DK.24.26.29.32 FI.21.31.30.33 FR.22.18.21.19 NL.11.13.20.21 NO.21.29.33.35 SE.32.38.49.52 Source: ESSnet, Micro-moments database v3.4
Wage polarization: ICT intensive vs non-intensive firms log(w ICT =1 /w ICT =0 ) c,i,t = α + γ time-trend + FE + ε c,i,t γ = 1.4(3.95): each year 1.4% larger wage difference Instead of time-trend, time dummies: coefs shown below
Wage polarization Similar wage divergence seen from long panel with fixed firm classification (with γ =.8) The data do not show a widening of wage differences over time for firms split by foreign ownership, MNC, exporter, innovator, size, or age. Even when correcting for observable employee characteristics (education, for a subsample of countries), wedge of ICT-intensive firms widens Evidence other from datasources with employee characteristics show widening of distribution, in many countries. Direct link with ICT not shown (although smoking guns)
ICT Impact in low and medium income countries Lower transactions costs and lower MC: increase geographic scale of sales Increased utilization/efficiency of capital (equipment, structures, knowledge) Changes in market structure: some worry about concentration and inequality What about boundary formal/informal business?
ICT Adoption: Lessons from advanced economies Strong country effect. Heterogeneous adoption related to policy environment Carrot: Ability to scale up succesful business model Stick: Rigorous market selection At firm-level: adoption of ICT related to other good firm characteristics: human capital, exporting, innovativeness IT is no free lunch, but it is better to have lunch than be lunch (Zvi Griliches)
ICT Adoption in low and medium income countries Infrastructure and Human Capital bottlenecks require policy attention Carrot: local network effects still small; global scale requires local resources; level of trust (and legal back-up) between network participants Stick: bancruptcy laws, entry barriers, political favoritism At firm-level: governance and ownership structure Adotpion in informal sector
Economic Analysis of Future Impact Changes in structure of production technology: ICT is non-rival in production. So, industry structure as in Hopenhayn (1992): high fixed costs of entry, low marginal costs, ex-post quasi rents. ICT brings substitution of some tasks and complementarity with other tasks (Acemoglu and Autor) Share of income to flexible factors (capital, labor) declines as fixed cost rises. Distribution of quasi-rents to fixed costs become winner take all It is very difficult to forecast which tasks will be substitutes, given rapid change of possibilities from combinations of technologies. A different organisation of work (time/place) becomes possible (and desirable)
Sustainable Policy Adoption is prerequisite for future potential Human Capital: not just for the young Legal and physical infrastructure Level playing field, reduce distortions Vision of future direction Employment vs self-employment; formal vs informal-in-network ICT in public sector Opening up of public service markets