WimbleThe Efficient Use of IT: Manufacturing vs. Services
The Efficient Use of Information Technology:
An Industry-Level Comparison of Manufacturing and Services
Matt Wimble
MichiganStateUniversity
Extended Abstract
Introduction
Despite the salience of information technology (IT) investment in many firms, industry and firm-specific variations exist in the efficient use of information technology. This research examines the factors that influence variations in the efficient use of IT resources across industries. Using data from Compustat, the Bureau of Economic Analysis (BEA), and the Bureau of Labor Statistics (BLS), we apply Data Envelopment Analysis (DEA) to trace the industry-level efficiency of IT capital.The results reveal that 1) IT capital is used more efficiently in service industries, 2) industry concentration significantly impacts IT capital efficiency, and 3) the relationship between competitive intensity and IT capital use efficiency differs across the services and manufacturing sectors. After an extensive literature review, it was found that investigations into the economic impact of information technology spending were at the level of analysis of the firm, generally used central tendency measures, and rarely looked at industry-level effects. This paper was developed to 1) provide insights into the efficient use of IT and 2) examine industry-level factors that impact the efficient use of IT. Following the motivation of the paper, the following was done:
Theoretic Development
As the impacts of IT spending have been better understood there has been a greater need to understand the conditions which have impacts on the economic outcomes resulting from expenditure. Service industries have higher skilled workers than manufacturing industries, and Hypothesis 1 follows from Griliches’ (1969) observation that capital and worker skill are complimentary:
Hypothesis1: IT capital efficiency is greater in services than in manufacturing.
Chang and Gurbaxani (2005)found that increased industry concentration is positively associated with IT capital efficiency at the firm-level. The finding that increased concentration leads to increased efficiency is consistent (Baumol, et. al., 1982; Färe, 1986)with industrial organization theory. Given that this paper uses multiple measures we attempt to confirm the existing research on IT efficiency applies to both the industry-level and when multiple output measures are used. This leads to the following:
Hypothesis 2: Increasing industry concentration is positively associated with IT capital efficiency.
Next we explore what explore the differential effect of increased concentration on manufacturing and service industries. The issue are most relevant to discussions of efficiency is industry growth. According to Ghemawat and Nalebuff (1984) increased efficiency from increase concentration is premised on the industry expanding, but when industries are contracting decreased concentration lead to greater efficiency. Most industries expand, so the assumption of an expanding industry normally is not relevant to discussions of industry concentration and efficiency. In more recent years manufacturing has decreased(Filardo, 1997) in size in real terms. Note that this does not contradict H2 because overall growth has been positive because services went up faster than manufacturing went down.. This leads to the following:
Hypothesis 3: Increasing industry concentration is positively associated with IT capital efficiency in Services.
Hypothesis4: Increasing industry concentration is negatively associated with IT capital efficiency in manufacturing.
Methodology
Data
The Data for this study was collected from several sources. All data is at the 3-digit NAICS level, with 61 industries used over seven years. Labor expenditure data was obtained from the BEA industry input-output accounts. IT capital and non-IT capital measures come from the BEA fixed asset tables. Employment information used for IT intensity measures comes from the Bureau of Labor Statistics. Industry concentration and financial variables were aggregated to the industry level from COMPUSTAT. ITK represents IT capital stock which is an aggregation of all the hardware, software, and communications capital listed in the BEA fixed asset tables for a given industry in a given year. K represents non-IT capital stocks, which is the aggregation of all other capital stock categories. ITL represented IT labor, which is a derived measure from NAICS industries that involve IT outsourcing. ITKE and ITLE are represent the IT capital and labor intensities. IT intensities of this form have been used before (Stiroh, 2002) in both ratio form and with a binary transformation to indicated if the industry is above or below the mean. HHI represents the Herfindahl-Hirschman Index (Hirschman, 1964) a commonly used measure of market concentration. The HHI used was the commonly used 4-firm HHI measure.
DEA efficiency
The first step in our analysis was to obtain efficiency scores for each industry-year using data envelopment analysis (DEA). We calculated the efficiency scores over 61 industries for seven years available across years, for a total of 727 industry/years. DEA measures offer several advantages. First DEA measures are inherently (Charnes, et al., 1978) prescriptive, as opposed to the descriptive nature of central-tendency measures such as ordinary least squares regression. Second, DEA provides a means to include multiple output measures and requires no statistical assumptions be made about the data. DEA was implemented using constant (CCR) returns to scale. We implemented an output-oriented formulation of DEA where the objective is to maximize output with a given set of inputs. The output-oriented formulation is expressed as:
Subject to:
The outputs in out DEA formulation were revenue per employee and value added per employee. Inputs were capital, labor, IT capital and IT labor.
Results
DEA efficiency
The efficiency results reported for both services and manufacturing represent those for which matching data was available for all covariates.Mean IT labor efficiency in manufacturing was higher than both IT labor efficiency and IT capital efficiency in services.
Figure 1. Comparative cumulative density functions of distributions of efficiency scores
This graph shows that the relative frequency of efficiency scores remains similar for both IT labor and IT capital in the service sector, but the distribution of efficiency scores in manufacturing differs significantly when capital and labor efficiency scores are compared within manufacturing. In manufacturing industries there are a nearly 80% of industry-years that failed to make obtain even 10% efficiency for IT labor, but 55% of the industries were at least 50% efficient when it came to IT capital efficiency.
Covariate analysis
After obtaining efficiency scores for all industry/years,the scoreswere into to sets.One set for service industries and one for manufacturing industries. Next we regressed the efficiency score against the market concentration ratio, IT capital-intensity measure, and the IT labor-intensity measure over 3 equations:
I )
II )
III )
Where ITKEFF is the IT capital efficiency, HHI is industry concentration, SERV is a binary indicating a service industry, ITLE is IT labor-intensity, ITKE is IT capital intensity, and Year is a 1-7 number indication the year. Regression I includes both service and manufacturing industries, regression II is on services, and regression III is manufacturing.Due to the cross-sectional nature of the data we checked for heteroskedasticity using the White Heteroskedasticity Test with cross-terms on all regressions and corrected for it using White Heteroskedasticity-Consistent Standard Errors (WHCSE) where indicated. Regression results are shown in tables 1 & 2.
Table 1. Regression results
Equation I: Manufacturing and ServicesVar / Coeff. / Std. Error / t-stat. / Prob.
C / -0.0963 / 0.0464 / -2.0774 / 0.0386
HHI / 0.2991 / 0.0817 / 3.6598 / 0.0003
SERV / 0.0759 / 0.0227 / 3.3482 / 0.0009
ITLF / 0.0751 / 0.0605 / 1.2414 / 0.2154
ITKE / 0.5510 / 0.0561 / 9.8128 / 0.0000
YEAR / -0.0021 / 0.0076 / -0.2741 / 0.7842
R-squared / 0.5100
Adj. R-sq. / 0.5016
Hypothesis 1 is supported by α2 or SERV in equation 1. Hypothesis 2 is supported by α1 or HHI in equation 1. Overall model fit is good. The year variable is insignificant indicating that calculating the efficiency over multiple years does not cause significant issues. IT capital intensity was significant, but it was not considered in this study.
Table 2. Regression results
Eq II: Manufacturing / Eq III: ServicesVar / Coeff. / Std. Err. / t-stat / Prob. / Coeff. / Std. Err. / t-stat / Prob.
C / 0.0493 / 0.0076 / 6.4654 / 0.0000 / -0.0577 / 0.0592 / -0.9748 / 0.3310
HHI / -0.0497 / 0.0118 / -4.2130 / 0.0001 / 0.3581 / 0.0935 / 3.8286 / 0.0002
ITKE / 0.2129 / 0.0208 / 10.2439 / 0.0000 / 0.6289 / 0.0671 / 9.3698 / 0.0000
ITLF / 0.0866 / 0.0081 / 10.7299 / 0.0000 / 0.0825 / 0.0980 / 0.8417 / 0.4011
YEAR / 0.0037 / 0.0018 / 2.0191 / 0.0458 / -0.0043 / 0.0121 / -0.3566 / 0.7218
R-squared / 0.9162 / R-squared / 0.5190
Adj. R-sq. / 0.9133 / Adj. R-sq. / 0.5080
Hypothesis 3 is supported by γ1 or HHI in equation 2. Hypothesis 4 is supported by β1 or HHI in equation 3. Overall model fit is good. The year variable was insignificant for services, but significant in manufacturing. This might indicate that calculating the efficiency over multiple years could cause significant issues and needs further exploration.
Table 3. Summary of findings
Hypothesis / Findings / Supported?H1: IT capital efficiency is greater in services than in manufacturing. / Significant at 1% / Yes
H2: Increasing industry concentration is positively associated with IT capital efficiency. / Significant at 1% / Yes
H3: Increasing industry concentration is positively associated with IT capital efficiency in Services. / Significant at 1% / Yes
H4: Increasing industry concentration is negatively associated with IT capital efficiency in manufacturing. / Significant at 1% / Yes
Discussion
This paper explored several under investigated aspects of the impacts of information technology spending. First, this paper examined the efficient use of IT, most prior studies use central-tendency measure to examine the impacts of IT on average. Second, this paper disaggregated industries and showed that industry concentration has different effects on services compared with manufacturing. The paper represents a significant step forward on those fronts. The economic impacts of information technology (IT) been investigated primarily through a firm-level lens, this paper looks at the impacts of information technology through the underused industry-level lens. Additionally this study results in relative-to-optimum view through the use of DEA analysis, rather than the on-the-average results the come from regression analysis. Using efficiency scores has the benefit of being more prescriptive to managers. As far as future research industry-level data is becoming available in increasing granularity and would seem to warrant future use. While this study looked at industry concentration, other factors such as price uncertainty have shown to be relevant factors in other industry-level analysis and should be pursued.This paper makes a novel contribution on three fronts: 1)IT capital efficiency is greater in service industries than manufacturing industries, as well ascompetitive intensity at first appears to increase IT capital efficiency, but actually2) competitive intensity increases IT capital efficiency in services, and 3) competitive intensity decreases IT capital efficiency in manufacturing. The results of this study represent a significant step toward understanding, not if IT expenditures have positive impacts, but under what conditions IT expenditure has positive effects.
Works Cited
1)Baumol, Panzar, and Willig, 1982, Contestable Markets and the Theory of Industry, Harcourt Brace, San Diego.
2)Chang and Gurbaxani, 2005, “An Empirical Investigation of IT Returns: the Role of IT and Competition as Determinants of Efficiency”, working paper
3)Charnes, A., Cooper, W., and Rhodes, E., 1978, “Measuring the efficiency of decision making units”, European Journal of Operational Research, v.43(2), pp.429-44.
4)Dedrick, Gurbaxani, and Kraemer, 2003, “Information Technology and Economic Performance: A Critical Review of the Empirical Evidence”, ACM Computing Surveys, V35, no 1, pp1-28
5)Färe, R., 1986, “Addition and Efficiency”, Quarterly Journal of Economics
6)Filardo, A., 1997,“Cyclical Implications of the Declining Manufacturing Employment Share.", Economic Review Federal Reserve Bank of Kansas City, 82(2), pp. 63-87.
7)Ghemawat, P. and Nalebuff, B., 1990, “Declining Markets: The Devolution of Declining Industries”, Quarterly Journal of Economics, v. 105, pp167-86.
8)Griliches, Z., 1969, “Capital-skill Complementarity”, Review of Economics and Statistics, 51(4), 465–468.
9)Grover, V., Cheon, M., and Teng, J., 1996, “The Effect of Service Quality and Partnership on the Outsourcing of Information Systems Functions”, Journal of Management Information Systems, 12(4), pp. 89-116
10)Hirschman, A., 1964, “The Paternity of an Index”, The American Economic Review, Vol. 54, No. 5, p 761.
11)Morrison, C., 1997, “Assessing the Productivity of Information Technology Equipment in U.S. Manufacturing Industries”, The Review of Economics and Statistics, v79, n3, pp. 471-481
12)Stiroh, K., 1998, “Computers, Productivity, and Input Substitution”, Economic Inquiry, pp175-191
13)Stiroh, K., 2002, “Information Technology and the Productivity Revival: What Do the Industry-Level Data Say?”, American Economic Review , v92 no5, pp1559-1576
1
Conference on Information Systems and Technology, Pittsburg, Pennsylvania November 04th&05th 2006