IT and Firm Performance: Roles of Internal and External IT specialists* Mika Maliranta, ETLA & Univ. of Jyväskylä Petri Rouvinen, ETLA/Etlatieto Aarno Airaksinen, Statistics Finland Final conference of the ESSLimit project: ICT and Innovation: Accessing and linking Microdata for Firm-level Economic Analysis Sevilla, 9-10 October 2012 * A part of the ESSlimit project. Rouvinen has contributed to this paper as a part of BRIE-ETLA collaborative research.
Data show that Conclusions computer use is almost saturated (72%=>75%, 2004=>2010) share of IT specialists is growing in businesses (4% => 6%, 1995=>2009) IT support from external IT specialists is growing (42%=>53%, 2004=>2010) Attention to the role of IT specialists & IT outsourcing (i.e external IT specialists) Analyses give indication Significant productivity effects of computer use Effects bigger when supported by external IT specialists (vs. by internal IT specialists) Very significant with OLS (and careful control for other factors) Weakly significant (but supportive) with control for unobservable time invariant factors (management etc.)
Outsourcing Large and rapidly growing literature on outsourcing/offshoring (see Helpman 2006). Effects on employment (e.g. Falk-Wolfmayr 2008, Hsu 2011, Ohnemus 2011) productivitity (e.g. Olsen 2006, McCann 2011, Falk 2012) subjective well-being (e.g. Geishecker 2008, Böckerman-Maliranta 2012) Much less is known about IT outsourcing (e.g. Maliranta-Rouvinen-Airaksinen 2008, Ohnemus 2011, Kite 2012)
What IT Outsourcing Is in Real Life?
IT Outsourcing A process to contract-out or to sell the organization s IT assets, staff and/or activities to a service provider (Kern, Willcocks & Heck 2002) New, Big & Growing? Around since 1990s Commoditified in the 2000s Global IT outsorcing $246 bn in 2011 (http://www.gartner.com/it/page.jsp?id=2021215) Major Suppliers (# of employees) Massive R&D on how to IBM 400,000 provide services Accenture 200,000 Tata Consultancy Ser. 174,000 Solutions databases: Solve Infosys & its subsidiaries 122,468 issues once (globally)
Evidence from Finnish National Accounts Labor input share of IT services (NACE.2, 62&63) in Business sector Source: National Accounts
Why outsource? The client firms point of view Scale/scope economies/efficiencies Lower quality-adj. costs Business vs employment contracts Higher flexibility Less hardware/software investment Lighter capital structure Why not? Long-run effects on firm s own innovativeness? Loss of control Transaction/coordination costs Induces a major organizational change Theoretical perspectives on IT outsourcing (Mahnke, Overby & Vang, 2005): (1) transaction cost economics (2) capacity-based (3) relational views.
Some empirical literature Few studies with mixed evidence, e.g. Ohnemus 2011 / German service & manuf. sectors -Cross-section (2004) and panel (2000-2007) data -Consideration of endogeneity -Evidence on: * significant positive effects of IT * complementarity between IT use and IT outsourcing Kite 2012 / Indian service & manuf. sectors -Panel data 2005-2008 -Consideration of endogeneity -Evidence on: * significant positive effects of IT * outsourced IT has a larger effect than in-house IT
Our paper
External support Theory IT production users IT non-users IT specialists Supporters Developers
Definitions 1/2 Internal: services provided by own IT specialists (supporters and developers) External: services provided by external (outsourced) IT specialists Internal IT specialists Supporters. Following occupations: -Computer assistants, computer equipment operators and related associate professionals (3120 according to ISCO-88) Developers. Following occupations: -Computing services managers (1236) -Computing professionals (213) -Electronics and telecommunications technicians (3114) -Electronics mechanics and servicers (7242) -Telecommunications installers and mechanics (7244)
Definitions 2/2 External support -In the 2005 survey section 7.1 and in the 2011 survey section 9.1 asks - To what extent are the following IT functions performed by your firm s own (hired) labor / outside labor? a. Design/development of Internet homepages, b. Maintenance of Internet homepages, c. Internet marketplace for private/retail customers, d. Internet or extranet marketplace for businesses, e. Other business to business commerce application (for example EDI), f. Development and maintenance of applications, g. Development of other information technology systems, h. Operation/maintenance of servers, i. Operation/maintenance of a PC environment, and j. User support, For each of the above, one of the following answers: Completely performed by external labor 100% Mostly performed by external labor 75% Equally performed by external and own labor 50% Mostly performed by own labor 25% Completely performed by own labor 0% I am unable to say / Irrelevant Outsourced:
Data Finnish Longitudinal Employer Employee Data (FLEED) Proportion of IT specialist occupations Characteristics of employees (education, age etc.); for control variables E Commerce Survey Share of computer users Share of outsourcing (years 2004 & 2010) Use of own or external IT specialists (year 2006), for international comparisons Structural Business Statistics Production information for measuring firm performance Business Register Industry and other basic information
Some summary statistics IT users (sh.): 72% (2004) 75% (2010) of use a computer Internal supporters (sh.): 2.0% 2.2% Internal developers (sh.): 5.1% 5.3% External specialists (sh. of all specialist) 42% 53%
Source: Finnish Longitudinal Employer- Employee data Shares of internal IT specialists, business sector
Share of external IT specialists of all IT specialists, Business sector Source: E-Commerce surveys
Basics of the econometric model A Cobb-Douglas labor productivity specification derived at the individual level but employing grouped data The specification assumes that workers are similar after controlling their observable qualities extensive controls The specification uncovers an independent productivity effect of computer use Productivity effect of -Supporters & developers & -External vs. internal IT specialists on productivity effect of computer use Labor weighted pooled OLS, long difference & fixed effects
POOLED WEIGHTED OLS RESULTS LHS: ln(value added / labor) (1) (2) (3) (4) (5) ICT: External specialists 0.140 ** ICT: Users 0.184 *** ICT: Users with internal specialists 0.086 + ICT: Users with external specialists 0.291 *** ICT: Production users 0.193 *** ICT: Internal supporters 0.297 ** 0.301 ** 0.152 ICT: Internal developers 0.536 *** 0.504 *** 0.309 ** ICT: Product. users with int. special. 0.085 + 0.113 ** ICT: Product. users with ext. special. 0.285 *** 0.260 *** ICT: Internal External specialists 0.943 *** Ed: Associate or eq. degree 1.086 *** 1.086 *** 1.044 *** 1.039 *** 1.022 *** Ed: Bachelor or eq. degree 0.65 *** 0.636 *** 0.575 *** 0.595 *** 0.547 *** Ed: Master or eq. degree 1.215 *** 1.237 *** 1.11 *** 1.201 *** 1.129 *** Ed: PhD or eq. degree 0.527 ** 0.509 * 0.548 ** 0.496 * 0.556 ** Age: Youngest, 16 24 yrs 0.046 0.064 0.068 0.055 0.027 Age: Younger, 25 34 yrs 0.246 0.25 0.223 0.225 0.255 Age: Older, 45 54 yrs 0.036 0.047 0.004 0.035 0.097 Age: Oldest, 55 70 yrs 0.07 0.085 0.045 0.036 0.032 Gender: Women 0.506 *** 0.5 *** 0.49 *** 0.493 *** 0.495 *** CD: ln(capital/labor) 0.107 *** 0.107 *** 0.108 *** 0.106 *** 0.104 *** CD: ln(labor) 0.011 0.011 0.006 0.009 0.008 Industry effects yes yes yes yes yes Year 2010 effect yes yes yes yes yes Obs. 2854 2854 2854 2854 2854
WEIGHTED FIXED EFFECTS RESULTS LHS: ln(value added / labor) (1) (2) (3) (4) (5) ICT: External specialists 0.078 ICT: Users 0.043 ICT: Users with internal specialists 0.031 ICT: Users with external specialists 0.155 + ICT: Production users 0.053 ICT: Internal supporters 0.158 0.159 0.587 ICT: Internal developers 0.166 0.225 0.32 ICT: Product. users with int. special. 0.057 0.074 ICT: Product. users with ext. special. 0.189 * 0.193 * ICT: Internal External specialists 0.643 + Ed: Associate or eq. degree 1.226 ** 1.197 * 1.196 * 1.177 * 1.106 * Ed: Bachelor or eq. degree 2.025 *** 1.984 *** 1.948 *** 1.961 *** 1.981 *** Ed: Master or eq. degree 0.154 0.103 0.174 0.123 0.2 Ed: PhD or eq. degree 0.268 0.273 0.318 0.336 0.446 Age: Youngest, 16 24 yrs 1.132 ** 1.096 ** 1.198 ** 1.092 ** 1.102 ** Age: Younger, 25 34 yrs 1.3 *** 1.312 *** 1.342 *** 1.337 *** 1.33 *** Age: Older, 45 54 yrs 1.366 *** 1.347 *** 1.41 *** 1.367 *** 1.436 *** Age: Oldest, 55 70 yrs 0.173 0.214 0.129 0.245 0.286 Gender: Women 0.256 0.226 0.287 0.212 0.222 CD: ln(capital/labor) 0.039 ** 0.037 ** 0.034 * 0.036 * 0.037 ** CD: ln(labor) 0.189 *** 0.188 *** 0.194 *** 0.192 *** 0.202 *** Industry effects no no no no no Year 2010 effect yes yes yes yes yes Obs. 960 960 960 960 960
International context & comparison
Cross-country evidence Based on the Eurostat s project the ESSnet on Linking on Microdata on ICT Usage (ESSLimit) Make use of harmonized SAS-codes for creating indicators and performing analysis Analyses concern 7 (out of 15) countries Data on internal (own) IT specialists and external ( outsourced ) IT specialists for year 2006 Dependent variable (log of) value added per employee Control for firm size and age, broadband use, IT education and exporting (and capital use in case of TFP)
Cross-country results: distribution of firms
Cross country results: labor productivity differences
Data show that Conclusions computer use is almost saturated (72%=>75%, 2004=>2010) share of IT specialists is growing in businesses (4% => 6%, 1995=>2009) IT support from external IT specialists is growing (42%=>53%, 2004=>2010) Attention to the role of IT specialists & IT outsourcing (i.e external IT specialists) Analyses give indication Significant productivity effects of computer use Effects bigger when supported by external IT specialists Very significant with OLS (and careful control for other factors) Weakly significant (but supportive) with control for unobservable firm factors (management etc.)