11

Third Workshop on Information Society Measurement in Latin America and the Caribbean, Panama, 23 November 2006

THE IMPORTANCE OF MEASURING BUSINESS USE OF ICT

Martin Schaaper

Economic Analysis and Statistics Division, Directorate for Science, Technology and Industry

22

Why measuring business use of ICT?

Assess take up and diffusion– … and compare with other countries

How does the Internet change the way businesses run their operationsImpacts on overall economic performance (productivity, growth, etc.)– … and compare with other countries

Policymakers need this information to make sound decisions

33

The state of ICT diffusion

44

ICT diffusion among OECD-and non-OECD countries

Diffusion can be measured by activity and financial statistics.OECD has been collecting activity statistics on ICT use and e-commerce from member countries for several years and first published data in 2001 (“OECD Science, Technology and Industry Scoreboard: Towards a Knowledge-based economy”).The most recent version of this publication has been the 2005 edition.

55

Business use of the Internet, 2004, Percentage of businesses with 10 or more employees

(1) All size classes

(2) 16+ employees

0

20

40

60

80

100

Japa

n (20

03)

Denmark

Finlan

dSwed

enGerm

any

Korea

Canad

aAus

tralia

(4)

Mexico

(200

3)

Netherl

ands

Italy

Greece

Norway

Poland

Chines

e Taip

ei (1)

Hong K

ong (

China)

Hunga

ryPort

ugal

Singap

ore (2

003)

(1)

Slovak

Rep

ublic

Thail

and (

2)Chil

e (1)

% Have Internet access Have own Web site Broadband (2)

66

Internet selling and purchasing, 2004, Percentage of businesses with 10 or more

employees

- 40- 30- 20- 10 0 10 20 30 40 50 60

Canada

Hong Kong, China (1)

United Kingdom

Germany

Sw itzerland (2002)

Australia

Singapore (2003) (1)

Japan (2003)

Korea (3)

Chinese Taipei (1)

Mexico (2003)

Chile (1)

Thailand (2)

Purchasing Selling

(1) All size classes(2) 16+ employees(3) Data refer to enterprises with 1 or more employees.

77

Total e-commerce transaction value (including via the Internet), 2002 to 2004.

As a percentage of total enterprise turnover

0

5

10

15

20

25

Irelan

dUnit

ed King

dom

Denmark

German

yFinl

and

Norway

Belgium

Austria

Czech

Rep

ublic

Portug

alSlov

ak Rep

ublic

Italy

Luxe

mbourg

Spain

Poland

Greece

%

2002 2003 2004

88

Business perceptions of the benefits of buying and selling over the Internet

99

Business perceptions of the benefits of buying and selling over the Internet

Australian Bureau of Statistics data show that businesses which buy and sell over the Internet perceive benefits from doing so.The longer a business has been buying or selling over the Internet, the more likely it is to report benefits– first mover advantage?– survival of the fittest (those which are still selling are those

which have succeeded and therefore are more likely to see benefits)?

– or just getting better at it over time?

1010

AUSTRALIA (ABS): BUSINESS PERCEPTIONS OF THE IMPACT OF RECEIVING ORDERS (SELLING) VIA THE INTERNET, 2000–01

Quality

of cu

stomer

servi

ceBus

iness

costs

Sales o

r cus

tomers

Efficien

cy

Compe

titive

ness

Tradin

g hou

rsTra

ding a

rea

% o

f bus

ines

ses

selli

ng o

ver

the

Inte

rnet

Not applicableNo changeIncreasedDecreased

1111

OECD case study work on the impacts of e-business

1212

OECD case study work: Electronic Commerce Business Impact Project (EBIP)

During 2001-2002, the OECD co-ordinated a case study project on the impacts of electronic commerce on business. The study involved nearly 220 firms across 11 countries and a common methodology was used across 14 broad sectors. Participants: Canada, France, Italy, Korea, Mexico, the Netherlands, Norway, Portugal, Spain, Sweden, the United Kingdom. The aim of the project was to improve understanding of impacts of electronic commerce on business. EBIP used common methodology for firm-level case studies to improve cross-country and cross-sector comparability.Previous case study information is anecdotal, fragmented, not comparable across sectors or countries.

1313

What is more important the ‘commerce’factor or the ‘e’ factor?

Successful e-commerce strategies led by commercial considerations.

E-commerce part of larger business and economic transformations. Successful application and use are embedded in broader business strategies with major emphasis on both e-commerce and ICT skills.

But e-commerce a major business innovation that most firms will have to adopt.

1414

Overall the Internet is having large impacts on how firms conduct business

Expected/actual e-commerce impacts by business function

0%10%20%30%40%50%60%70%

Adverti

sing

Ca ta logues and st

ock list

s

Informatio

n servi

cesNegotia

tion

Orderin

g

Billing &

Paym

entFinanceDe liv

eryInform

ation C

ap ture

Inform

ation M

anagement

Market Ana lysis

Market Development

E x pec ted E ffec ts A c tua l Im pac ts

T RA NSA C T IO NPREPA RA T ION

T RA NSA C T IONC O M PL ET IO N

PRO DUC T IONSUPPO RT

1515

.. Reshaping many business processesExpected/actual e-com process impacts by business process

0 %

5 %

1 0 %

1 5 %

2 0 %

2 5 %3 0 %

3 5 %

4 0 %

4 5 %

5 0 %

Diversi

ficati

onDiffe

rentia

tion

Custom

isatio

nBun

dling

Design

Logis

tics

Produc

tion

Co-ordi

natio

n

Expan

sion

Segmen

tation

Trust

Loya

lty

E xp e c te d E ffe c ts A c tu a l Im p a c ts

P R O D U C TIN N O V A T IO N

P R O C ES SIN N O V A T IO N

R EL A T IO N A LIN N O V A T IO N

1616

.. With generally positive benefits:one-third of firms had positive impacts on

turnover or profitability

0% 20% 40% 60% 80% 100%

Employment

Profitability

Turnover

Increase

Decrease

No Change

NA

1717

Analysing the impacts of ICT

1818

OECD Growth ProjectA two year study on the reasons for differences in underlying economic growth of OECD member nations during the 1990s.Final report released in 2001 - “The New Economy: Beyond the Hype”.General findings were that:– ICT is a key technology with the potential to transform economic

and social activity.– But it is not the only factor, others are:

• quality of human capital• providing more scope for entrepreneurs to explore business

opportunities• support for innovation• getting the fundamentals right – sound macro-economic

management, openness to trade and investment, efficient markets, well functioning economic and social institutions.

1919

OECD Growth Project (continued)

ICT plays three roles– Through capital deepening, as ICT is an important asset

in overall business investment.– Through multi-factor productivity growth in the

production of ICT goods and services (e.g. technological progress in semi-conductors).

– Through MFP growth thanks to the use of ICT, either through efficiency gains in individual firms, or through network/spillover effects from its use.

2020

OECD Growth Project (continued)

Findings in respect of ICT considered both ICT production and consumption and were as follows:– Productivity in the ICT sector can improve economic

productivity overall, but– … successful economies were more likely to have rapid

diffusion of ICT, particularly in service industries.– Deregulation of ICT industries encourages competition and

greater investment in ICT.– Policies which build confidence – an appropriate regulatory

and legal environment and government leading by example –are important.

2121

OECD Growth Project (continued)

In respect of ICT particularly, the policy recommendation from the final report was that:– “ICT is an enabling technology, that is transforming

economic activity. – Governments should:

• focus policy efforts on increasing the use of new technology

• increase competition and continue with regulatory reform in the telecommunications industry to enhance the uptake of ICT

• ensure sufficient competition in hardware and software to lower costs

• build confidence in the use of ICT for business and consumers; and, make e-government a priority.”

2222

More recent analytical work on ICT by the OECD and others

2323

Much better evidence on the economic impacts of ICT at three levels of analysisthrough macro-economic evidence on the role of ICT investment in capital deepeningthrough sectoral analysis showing the contribution of ICT-producing sectors and ICT-using services to productivity growthand through detailed firm-level analysis that has demonstrated the wide-ranging impacts of ICT in the economy, even in sectors where sectoral data suggested that little was happening – this result is thanks to work in 13 OECD countries.

2424

2525

The data on ICT have improved

Early 1990s: mainly private data, sometimes with poor coverage, sample bias, and unknown quality.

Now:– Surveys on ICT use by households and businesses in most OECD

countries, with considerable detail on uses and technologies.– Official estimates of ICT investment, including software.– Industry data for many countries.– Growing comparability of data.– Efforts to establish longitudinal (linked) databases in many

countries.

Greater prospects for empirical research on ICT.

2626

Prices have fallen…(producer price index for PCs and workstations, 1998=100)

Source: BLS.

010

20304050

607080

90100

Dec-98

Apr-99

Aug-99

Dec-99

Apr-00

Aug-00

Dec-00

Apr-01

Aug-01

Dec-01

Apr-02

Aug-02

Dec-02

Apr-03

Aug-03

Dec-03

Apr-04

2727

But countries have not invested to the same extent…(ICT investment as % of non-residential investment)

0

5

10

15

20

25

30

35

Spain

Portug

alIre

land

Austria

France

Japa

nIta

lyGerm

any

Netherlan

dsGreec

eAus

tralia

Canada

Denmark

Belgium

United King

domFinl

and

Sweden

United Stat

es

1980

1990

2002*

Source: OECD Database on Capital Services, March 2004.

2828

Why the differences in growth from ICT?Returns (& investment) in ICT depend on other factors:– Skills, innovation, organisational changes.– Not all firms succeed – experimentation and selection play a

role – can successful firms gain market share?– The scope for change depends on the business environment;

impacts in the US seem larger, perhaps because of greater scope for innovation and organisational change.

Competition matters:– Competition forces firms to enhance efficiency.– Firms that invest first/most in ICT are often new or foreign.– Competition reduces the costs of ICT & fosters diffusion.

Structural factors – share of services and of large firms.

2929

Aggregate impacts of ICT investment

Mainly analysed through standard growth accounting approach, based on capital services:– Requires long time series of ICT investment.– For international comparisons – use of hedonic deflators for

all countries to enhance comparability of results.

Official data on ICT investment are now available for 15-20 OECD countries.Still some measurement problems, e.g. as regards software investment.

3030

The US and small EU countries have had a large contribution of ICT investment, France, Germany and Italy a small one

(contribution to GDP growth, in percentage points)

0.0%

0.1%

0.2%0.3%

0.4%

0.5%

0.6%

0.7%0.8%

0.9%

1.0%

France

German

yIta

lySpa

inPort

ugal

Finlan

dGree

ceJa

pan

Canad

aIre

land

Austra

liaNeth

erlan

ds

United

Kingdo

mBelg

iumDen

markUnit

ed Stat

esSwed

en

90-95

95-2002*

Source: OECD Productivity Database, May 2004.

3131

Finding: ICT investment and diffusion mainly differ because of lack of demand

Demand has been held back by a business environment that was not sufficiently geared towards effective use of ICT – e.g. regulatory barriers, lack of skills, difficulty to change organisational set-ups, lack of innovation, etc.Competition has not been equally strong in all countries – costs of ICT and communications differ, as do pressures to improve performance.Demand has been held back by lack of security and trust.Structural factors also matter – ICT is not suited to all markets, sectors, or business models.

3232

Sectoral approach to impacts of ICTFocus on ICT production – what role does it play, does it explain US-EU differences:– Requires definition of ICT production:

• Developed by OECD (both ICT manufacturing and ICT services)• But not easily applicable for constant price series.

– Typically focus on some key components.

Focus on ICT-using industries:– Sectors that use ICT most (or have the greatest potential for

ICT use) might first observe productivity impacts – these are mainly services industries.

– Definitions differ and are not used the same across studies.

3333

ICT manufacturing is only important for some countries(average annual contribution to labour productivity growth, in %)

-0.10.00.10.20.30.40.50.60.70.80.91.01.1

Spain

Italy

Mexico

Netherl

ands

Canad

aGerm

any

Denmark

United

Kingdo

mSwitz

erlan

dBelg

iumFranc

eJa

pan

United

States

Sweden

Finlan

dIre

land

Korea

1990-95

1996-2002*

3434

Productivity growth in ICT services (telecom and software) also plays a role

Annual average contribution to labour productivity growth, in %

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Switzerl

and

Belgium

Norway

Denmark

Austria

France

Spain

United

States

Netherl

ands

Japa

nMex

ico Italy

Canad

aSwed

en

United

Kingdo

mIre

land

Korea

Luxe

mbourg

Finlan

dGerm

any

1990-95

1996-2002*

3535

Countries with less product market regulation have seen a stronger pick-up in productivity in ICT-using services

Australia

AustriaBelgium

Canada

Denmark FinlandFranceGermany

Ireland

ItalyJapan

SwedenSpainNetherlands

United Kingdom

United States

-1.0

-0.5

0.0

0.5

1.0

1.5

0.0 0.5 1.0 1.5 2.0 2.5 3.0Index of Product Market Regulation, 1998

Incr

ease

in p

rodu

ctiv

ity g

row

th, I

CT-u

sing

indu

strie

s, 19

90-9

5 to

199

6-20

01 (%

poi

nt) Correlation = -0.71

T-statistics = -3.80

3636

Countries with strict employment protection legislation and product market regulation have invested less in ICT

Australia

Austria

Belgium

Canada

Denmark

Finland

France

Germany GreeceIreland

ItalyJapan

Netherlands

Portugal

Spain

Sweden

United Kingdom

United States

0

5

10

15

20

25

30

0 1 2 3 4Employment protection legislation index

ICT

inve

stm

ent i

n 19

98 (a

s %

of G

FCF)

Correlation = -0.65T-statistics = -3.46

3737

Countries with strong innovation in applications and software have invested more in ICT

Austria

BelgiumDenmark

France

GermanyGreece

Ireland

Italy

Netherlands

Portugal

Spain

Sweden

United Kingdom

AustraliaCanada

Japan

United States

10

12

14

16

18

20

22

24

26

28

30

0 10 20 30 40

Share of ICT patents in all patents, 1998

ICT

as a

% o

f non

-resi

dent

ial i

nves

tmen

t, 19

98

Correlation = 0.59T-statistic = 2.84

3838

Countries with an abundance of high-skilled workers have invested more in ICT

Sweden

France

United States

Greece

Italy

GermanyDenmark

Belgium

Netherlands

United Kingdom

Austria

Spain

Finland

Luxembourg*

0

1

2

3

4

5

6

0 1 2 3 4

Share of high-skilled ICT workers in total occupations

ICT

inve

stm

ent (

%G

DP

)

Correlation = 0.68T-statistic = 3.04

3939

And countries with a rapid increase in investment in ICT have had more rapid MFP growth

Ireland

Finland

Denmark

Sweden

Australia

CanadaUnited States

Netherlands

Austria

Italy

Japan

France

Germany

Spain0

2

4

6

8

10

12

14

-1.5 -1.0 -0.5 0.0 0.5 1.0Change in MFP growth from 1980-90 to 1990-2000

Cha

nge

in IC

T in

vest

men

t as

% o

f GFC

F, 1

990-

2000

4040

Problems with the industry approachBreakdown in ICT-producing and ICT-using industries is not always very satisfactory.Industry data are not always comparable, e.g. for ICT production and certain services.The impact of ICT on productivity is not formally analysed, ideally:– Regression analysis of impacts ICT use/ICT capital at the

industry level.– Growth accounts at the industry level.

Unfortunately, industry data on ICT investment are still lacking for many countries (no time series).

4141

A firm-level perspective on ICTOECD project involving researchers and

statisticians in 13 countries

4242

Large variety of methods and dataData based on combination (linking) of sources:– Economic performance often from production surveys;– ICT from investment surveys, specialised ICT and e-

commerce surveys, innovation surveys (US from computer network use survey).

– Additional data, e.g. on organisation, skills & innovation.Methods include:– Labour productivity regressions.– Estimates of production functions.

Variety of methods and data can help strengthen evidence.The studies point to many important interactions.

4343

Example 1 Hollenstein: What determines ICT uptake by firms?

Study for Switzerland, shows that ICT is linked to other firm-level factors.Firms that adopt ICT:– Anticipate benefits from improved customer-orientation

and lower costs.– Have overcome problems in financing and skills.– Have a high capacity to absorb new knowledge (linked to

human capital and innovation).– Are faced with strong competition and in markets with a

high level of uptake of ICT.– Also introduce new forms of work organisation.

4444

Example 2 Gretton, et. al: The impacts of ICT in Australia

The use of ICT raised aggregate multi-factor productivity by about 0.2% points – over and above impacts of ICT capital.Impacts are largest in finance and insurance.Impacts taper off over time – they are largest a few years after adoption.Complementary factors play a large role in explaining firm’s success with ICT – human capital, experience with innovation, use of advanced business practices and intensity of organisational restructuring.Early users of ICT are large firms with skilled managers.Firms with high openness to trade use more ICT.

4545

Estimated contribution of ICT to MFP growth in Australia (% points)

0.18

0.07

0.13

0.17

0.14

0.11

0.06

0.29

0.090.09

0.140.14

0.04

0.14 0.14

0.12

0.28

0.09

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Man

ufac

turin

g

Con

stru

ctio

n

Who

lesa

le tr

ade

Ret

ail t

rade

Acco

m.,

cafe

s &

rest

uara

nts

Tran

spor

t &st

orag

e

Cul

tura

l &re

crea

tiona

lse

rvic

es

Pro

perty

&bu

sine

ssse

rvic

es

Gro

up to

tal

Computer use & Internet aloneComputer use, Internet & computer use interactions

Source: Gretton, et al. 2004.

4646

Example 3 Hempell, et al. : The link between ICT and

innovation

Comparative study with comparable data for Germany and the Netherlands from innovation surveys.Key question: do firms that engage in complementary innovation benefit more from investment in ICT?In both countries, returns to ICT investment are much higher when accompanied by innovation.Continuous innovation leads to higher returns than occasional innovation efforts.This is also the case in the services sector.

4747

Example 4 Maliranta and Rouvinen: The impacts of ICT in

FinlandThe additional productivity of ICT-equipped workers ranges from 8 to 18% (after controlling for many other factors).This effect is higher in young (new) firms and in the ICT-producing sector.Manufacturing firms benefit mostly from internal networks (e.g. local area networks), whereas service firms benefit most from external networks (e.g. the Internet).Organisational factors are important, as suggested by the greater productivity effects in new firms.Selection also plays a role – not all firms succeed – having a dynamic business sector is important.

4848

Example 5 Clayton, et al.: Impacts of e-commerce in the UK

Computer networks for trading have positive impacts on productivity.E-buying has positive impacts on output growth, whereas E-selling has negative impacts.This is likely due to pricing effects – e-buyers can source more competitively, e-sellers may loose out.In some industries, e-commerce seems to have contributed to lower prices, perhaps due to greater price transparency and more rapid reactions of suppliers.

4949

UK: Productivity analysis and ITInvestment + Employees + Transactions + Communications

Purpose of analysis is to understand different effects

Labour Productivity: by Employee Use of ICT in Manufacturing and Services

0

10

20

30

40

50

Les s than 25 percent

25 per cent to 49 percent

50 per cent to 74 percent

75 per cent and over

Pe rce ntage of e mploye e s using compute rs / inte rne t

£thousand

Using computers Using Internet

5050

UK: Productivity analysis and IT

– Hardware and software capital show significant productivity gains, both levels and yearly changes

– Gains attributable to IT capital across service sectors higher than those in manufacturing

– Returns to IT capital across high and low user sectors, and for large and small firms

– IT capital reinforces productivity effects of other capital investment, and of skills

– Returns to IT investment ‘explains’ productivity advantage of US multinationals in UK, concentrated in high IT using sectors

Significant results – IT investment

5151

UK: Productivity analysis and ITSignificant results – IT use

– Employee use adds significantly to productivity, over and above effects of hardware and software capital,

• after controlling for sector, size, ownership etc

– Manufacturing firms achieve an extra 2.2% output for each additional 10% employees using computers

– For employees using internet the extra is 2.9%

– Productivity effect greater for more recently created firms

– Productivity effects associated with communications useUK firm level productivity results can be found in full athttp://www.statistics.gov.uk/cci/article.asp?ID=1235 (Web link to ONS summary article)and are broadly consistent with firm level results for Chinese manufacturing athttp//www.mo.rcast.u-tokyo.ac.jp/index-eng.html (analysis by Kazu Motohashi)

5252

UK: Productivity and broadband

No clear link from broadband availability to productivity, but:

– Investment in IT higher in regions / sectors with broadband available, especially for UK firms in non-IT intensive industries.

– Early adopters of broadband experienced high initial telecommunication costs but these have declined over time

– Broadband users more likely to have multiple e-business links.

– Multiple links plus broadband improve labour productivity.

– Firms with a high broadband equipped labour share have significantly higher productivity

Full results in OECD paper DSTI ICCP IIS(2006)9 Shikeb Farooqui and Raffaella Sadun

Significant results – broadband use

5353

Example 6 Atrostic et al. : Impacts of computer networks in the

United States, Japan and Denmark

Computer networks have positive impacts on productivity in all three countries.In the US manufacturing sector, plants using computer networks have 5-11% higher productivity than plants not using networks.In Japan, both interfirm and intrafirm networks are linked to strong firm performance.

5454

The firm-level studies all show that ICT use can contribute to improved firm performance

Positive impacts of ICT on labour productivity (and MFP, where measured) or market shares in all countries:– But these are conditional on other factors and firm

characteristics (skills, innovation, organisational change).– Not all firms succeed – selection plays a role.– ICT is only part of a broader strategy to improve firm

performance.Networking technologies are particularly important.The impacts of ICT are also found in the service sector (despite lack of evidence at the industry level).

5555

Why the difference between firm-level and aggregate evidence on ICT’s impacts?

Aggregate impacts ICT in some countries may still be disguised by other factors.Lags – the US invested earlier and more – aggregate impacts in other countries might still come.US firm-level impacts may be larger – because of scope for reorganisation and complementary innovation.Successful US firms may be able to gain more market share (e.g. Walmart) – more re-allocation.Spill-over effects? - Perhaps already some of this in the US (and Australia), not much in other OECD countries.

5656

Conclusions

Better data have enabled a wide range of empirical research on ICT and its economic impacts.This research has had important policy implications:– It has shown that ICT plays an important role in economic

growth.– ICT is no panacea, and linked to many other factors – only

focusing on ICT diffusion is not sufficient.– Countries have not equally benefited from ICT – there is scope

for learning and improvements in the policy environment.

There is scope for much more analysis, in particular with firm-level data, and in the context of international comparisons.

5757

Policy implications from work on ICT and growth

Fostering a business environment for effective use of ICT– Competition matters for ICT uptake and effective use.– Regulatory burdens may limit investment and

experimentation.– Enable firms to make the necessary organisational

changes.– New skills are needed, requiring changes to education

and training systems.– Need for coherent innovation policies.– Management matters – ICT is no panacea.

5858

Policy implications (continued)

Competition in ICT goods and services– Evidence of the benefits of competition is convincing and

liberalisation should be pursued.– Crucial for broadband development.– Regulators must be vigilant for anti-competitive

behaviour.– Technology continues to develop – policy frameworks

and regulations should be technology neutral.

5959

Policy implications (continued)

Boost Security and Trust– Uncertainty remains regarding the security of electronic

commerce.– Attention is required to security of information systems

and networks, authentication, protection of privacy and consumer protection.

– Cross-border issues and enforcement.– E-government can help build confidence – and is

important to improve government efficiency and the delivery of public services.

6060

Policy implications (continued)

Unleash growth in the services sector– particularly important because of the potential

benefits– policies to reduce regulation and reduce barriers

to entryHarness the potential of innovation and technology diffusion– ICT linked to, and assists, innovation– ICT fosters networking and knowledge transfer.

6161

Measurement issuesBetter measurement has clarified the picture– Quality-adjusted price indices have helped

demonstrate the impact of ICT on growth.– Software was included in 1993 SNA as an asset.– Measures of R&D are being improved:

• New (2002) edition of OECD Frascati Manual.• R&D may be included in new SNA.

– Measures of innovation are being improved.– There is more attention at statistical offices for

analysis at the firm level -linking existing data sources can help.

– Data on ICT demand are improving with adoption of OECD standards.

6262

However, some measurement problems remain

Skills data (apart from formal education statistics) are poor.Organisational factors are difficult to capture.Software investment is not yet fully picked up in national accounts.Deflators for ICT production differ considerably.Measurement of services output is poorly in some areas e.g. finance, health, education. Cross-country firm-level studies are still scarce, although much work is underway. OECD is trying to act as forum for exchange of best practices.

6363

martin.schaaper@oecd.org

THANK YOU!THANK YOU!www.oecd.org/sti/measuring-infoeconomy