ICT, the Internet, and Economic Performance:Empirical Evidence and Key Policy Issues

Donald Siegel

Professor of Economics

Rensselaer Polytechnic Institute

Prepared for UNCTAD and UNECE Conference

Geneva

October 20, 2003

Outline

Empirical Studies of the Impact of ICT on

Economic Performance Empirical Studies of the Impact of ICT on

Wages, Labor Composition (“Skill-Biased”

Technological Change-SBTC), and the Work

Environment Computers and the Internet as a General

Purpose Technology (GPT) Conclusions and Policy Recommendations

Investment in ICT and Economic Performance

Levels of AggregationPlant LevelFirm Level Industry LevelNational Level

Indicators of Economic Performance ProfitabilityMarket ShareStock Prices (Short and Long-Term Measures)Productivity

Labor Productivity (Partial Productivity) Total Factor Productivity

Estimating The Contribution of Computers to Economic Growth Augmented Production Function Approach

Qi = f (Ki, Li Mi, Ci)where: Q=output

K=stock of physical capital L=labor input

C=the stock of computer capital i=the unit of analysis (e.g., firm, industry, nation)

Two Approaches:Parametric-Estimate the Marginal Product or Output Elasticity of ComputersNon-Parametric- “Growth Accounting”-Analyze the Sources of Output Growth-(Jorgenson and Stiroh (2000)

Estimating The Contribution of Computers to Economic Growth(cont.)

Typical Parametric Studies (Brynjolfsson and Hitt (1996)):Cobb-Douglas Production Function-Constant Returns to Scale: Q = A eλt K L(1- )C

where: λ is a disembodied rate of growth parameter and and are output elasticities of capital and labor, respectively

Key parameter: the output elasticity of computer capital = (∂Q/ ∂C)(C/Q)

Typical Non-Parametric Studies (Siegel 1997)):Differentiating the Production Function w.r.t. time (t) and re-arranging terms leads to: PRODGROWTH = λ + ρ(C/Q)

Econometric Studies of the Impact of ICT on Productivity-U.S. Based Studies

Author(s) Level of Aggregation Results

Dunne, Foster, Haltiwanger, and

Troske (2000)

Plant-Level Positive Association Between Computers and Labor

Productivity, Which Appears to be Growing Over Time

Brynjolfsson and Hitt (1996)

Firm-Level “Excess” Returns to Computer Capital and Labor

Siegel and Griliches (1992)

Industry-Level Positive Correlation Between Investment in Computers and

TFP Growth

Jorgenson and Stiroh (2000)

Aggregate-Level Growth Contribution of ICT Increased Substantially in the

mid-late 1990s

Econometric Studies of the Impact of ICT on Productivity-Non-US Studies

Author(s) Level of Aggregation Results

Licht and Moch (1999)-Germany

Firm-Level Terminals Have a Positive Impact on Productivity in

Goods Industries, But Not in Services. Strong Positive

Relationship Between PCs and Productivity in Manufacturing

and Services

Gera, Gu, and Lee (1999)-

Canada

Industry-Level Positive Correlation Between Investment in Computers and Labor Productivity Growth

Greenan and Mairesse (1996)-

France

Firm-Level Impact of Computers is Positive and at Least as Large as For

Other Types of Capital. Returns Appear to be Higher In

Services Than in Manufacturing

Econometric Studies of the Impact of ICT on Productivity-”External” Effects

Author(s) Level of Aggregation

Results

Morrison and Siegel (1997)

Industry-Level “External” Investments in Computers by Related

Industries (Suppliers and Customers) Enhance

Productivity in a Given Industry

Key Stylized Facts Regarding ICT and SBTC

Strong Evidence in Favor of SBTC Positive Correlation Between Some Proxy for Technological Change and Shifts in Wages and Workforce Composition in Favor of Highly Educated Workers

Wage Premium of Slightly Less Than 20% in Most OECD CountriesInvestment in ICT Explains Approximately 33%-50% of the Increase in Returns to EducationIn Some Countries, the Demand for Computer-Literate Workers is Increasing More Rapidly Than SupplyThe Gap in Earnings Between Workers Who Use a Computer and Those That Do Not Appears to Have Widened

Key Stylized Facts Regarding ICT and SBTC-International Evidence (cont.)

Berman, Bound, and Machin (1998): Strong Evidence of Pervasive SBTC in OECD Countries-Patterns Are Quite

Similar Across Countries

Berman and Machin (2000): Comparison of SBTC in Developed and Developing CountriesAnalysis of “Technology Transfer” from Developed (OECD) to Developing

Countries Evidence of “Transfer” of Skill-Biased Technologies from High Income to Middle

Income Countries, But Not From High Income to Low Income CountriesLess Evidence of SBTC in Developing Countries

Evidence on SBTC in Developing Countries(From Country-Specific Studies)

Korea: Workers in Industries With Rapid Technological Change Are Paid More

Mexico, Taiwan, Colombia, Malaysia: Wages and Training Tend to Be Higher in Industries Experiencing Rapid Technological Change

Singapore: The Rise of the “Knowledge Economy” Appears to Explain The Increase in the Returns to Education

Vietnam: If Work Requires Computer Skills, Wages Are 10-14% Higher

ICT and Organization Change Siegel (1999)-Firm-Level Study

Implementation of new ICT Associated With Organizational Change Implementation of new ICT Associated With Increases in Training and Skill Upgrading Implementation of new ICT Associated With Enhanced Employee Empowerment Major Obstacles to Additional Investments in New Technologies:

High Cost of Customizing Software to Fit Company Needs Difficulties in Quantifying Benefits Some Firms Also Report Shortages of Skilled Workers

ICT and Organization Change Brynjolfsson and Hitt (2000)

ICT Reduces the Cost of Coordination, Communication, and Information Processing ICT is Associated with a “Cluster” of Complementary Organizational Changes/Practices

Transition from mass production to “flexible” manufacturing technologies Changing Interaction With Suppliers Transforming the Firm

Replacing Back-Office Jobs and Increasing the Importance of Front-office Skills and Leadership

Changing Interaction With Suppliers Decentralized Decision Making and Enhanced Communication

Computers/Internet as a General Purpose Technology (GPT)

GPTs-Technologies That Cause Dramatic Economic Changes by Stimulating New Applications and Industries and Rejuvenating Existing Sectors Strengthens the Rationale for Government Intervention

Examples of GPTs: Steam Engine, Electric Dynamo, Lasers, Computers

Computers/Internet as a GPTCreation of New Industries:

Internet Service Providers Network Communications Equipment Consulting Services

Conclusions and Policy Recommendations

ICT Appears to Be Generating High Social Returns

(Although Not Quite As High As Those Associated With

R&D, But Still Quite High)-Not Showing Up As Much In

Aggregate Statistics Because Productivity Gains Are In

Services ICT Has Transformed the Workplace in Many Industries

and Increased the Returns to Education in

Developed and Developing Countries (To a Lesser Extent) Developing Countries Appear to Be Falling Further

Behind in the Digital Divide, as Well as in the “Learning"

Divide

Conclusions and Policy Recommendations (cont.)

Increased Demand for ICT Workers May Not be Met in Many Developing Countries (Role of Outsourcing)

Important For Policymakers in Developing Countries To

Invest in Infrastructure That Complements/Supports Use

of ICT (Physical and Educational): Low-Cost, High-Bandwith Internet Connections Internet Security Ensure That Educational System Addresses Skill

Deficiencies Important For Developing Countries To Foster Institutions That

Complement/Support Use of ICT

Conclusions and Policy Recommendations (cont.)

Developed Countries-Public-Private Partnerships Have Been Useful in Alleviating Innovation Market Failures

Public-Private Partnerships in Developing Countries Can Be

Useful for: Better Access to Financial CapitalEnhancement of Human CapitalStimulating the Development and Extension of Networks

That Increase the Returns to ICT and E-Business