Prasanna Tambe

The Productivity of Information Technology Investments: New Evidence from IT Labor Data

This paper uses newly collected panel data that allow for significant improvements in the measurement and modeling of information technology (IT) productivity to address some longstanding empirical limitations in the IT business value literature. First, we show that using generalized method of moments--based estimators to account for the endogeneity of IT spending produces coefficient estimates that are only about 10% lower than unadjusted estimates, suggesting that the effects of endogeneity on IT productivity estimates may be relatively small. Second, analysis of the expanded panel suggests that (a) IT returns are substantially lower in midsize firms than in Fortune 500 firms; (b) they materialize more slowly in large firms---in midsize firms, unlike in larger firms, the short-run contribution of IT to output is similar to the long-run output contribution; and (c) the measured marginal product of IT spending is higher from 2000 to 2006 than in any previous period, suggesting that firms, and especially large firms, have been continuing to develop new, valuable IT-enabled business process innovations. Furthermore, we show that the productivity of IT investments is higher in manufacturing sectors and that our productivity results are robust to controls for IT labor quality and outsourcing levels.

Big Data Investment, Skills, and Firm Value

This paper analyzes how labor market factors have shaped early returns on big data investment using a new data source---the LinkedIn skills database. The data source enables firm-level measurement of the employment of workers with technical skills such as Hadoop, MapReduce, and Apache Pig. From 2006 to 2011, Hadoop investments were associated with 3% faster productivity growth, but only for firms a with significant data assets and b in labor markets where similar investments by other firms helped to facilitate the development of a cadre of workers with complementary technical skills. The benefits of labor market concentration decline for investments in mature data technologies, such as Structured Query Language-based databases, for which the complementary skills can be acquired by workers through universities or other channels. These findings underscore the importance of geography, corporate investment, and skill acquisition channels for explaining productivity growth differences during the spread of new information technology innovations. This paper was accepted by Alok Gupta, special issue on business analytics.

Broadband Adoption and Content Consumption

We explore how broadband access drives changes in the quantity and diversity of consumption of online content by using panel data that describes household Internet usage before and after broadband adoption. Our data suggests that on average, broadband adoption increases usage by over 1300 minutes per month. We also find that information consumption becomes more evenly distributed within the population, driven in part by post-adoption usage gains of almost 1800 minutes per month among individuals who were in the lowest usage quintile before adopting broadband. After adopting broadband, this pre-adoption lowest-usage quintile consumes content in greater quantities than users in neighboring quintiles, passing both the second and third quintiles in terms of absolute usage. This suggests that these users may have had strong preferences for high-bandwidth content that was too costly to consume in a narrowband environment. We also show that broadband adoption increases the variety of content that users consume although many of these gains appear to be associated with an increase in the variety of sites visited within previously visited content categories rather than an expansion in the types of content consumed.

How offshoring affects IT workers

IT jobs requiring interpersonal interaction or physical presence in fixed locations are less likely to be sent out of the country.

Now IT's Personal: Offshoring and the Shifting Skill Composition of the U.S. Information Technology Workforce

We combine new information technology (IT) offshoring and IT workforce microdata to investigate how the use of IT offshore captive centers is affecting the skill composition of the U.S. onshore IT workforce. The analysis is based on the theory that occupations involving tasks that are “tradable,” such as tasks that require little personal communication or hands-on interaction with U.S.-based objects, are vulnerable to being moved offshore. Consistent with this theory, we find that firms that have offshore IT captive centers have 8% less of their onshore IT workforce involved in tradable occupations; those without offshore captive centers have increased the proportion of onshore employment in these same occupations by 3%. In addition, we find that hourly IT workers (e.g., IT contractors) are disproportionately employed in tradable jobs, and their onshore employment is 2%--3% lower in firms with offshore captive centers. These findings persist after considering different measures of employment composition, including controls for human capital, firm performance, domestic outsourcing, and whether firms choose to build or buy software. Instrumental variables and corroborating regressions suggest that our estimates are conservative---the magnitude of the effect generally rises after accounting for reverse causality and measurement error. This paper was accepted by by Chris Forman, guest department editor.

Labor Economists Get Their Microscope: Big Data and Labor Market Analysis

This article describes how the fine-grained data being collected by Internet labor market intermediaries, such as employment websites, online labor markets, and knowledge discussion boards, are providing new research opportunities and directions for the empirical analysis of labor market activity. After discussing these data sources, we examine some of the research opportunities they have created, highlight some examples of existing work that already use these new data sources, and enumerate the challenges associated with the use of these corporate data sources.

The Effects of Organizational Factors on Healthcare IT Adoption Costs: Evidence from New York Nursing Homes

We investigate how organizational factors influence healthcare IT adoption costs, using new data collected from an EMR adoption experiment conducted in a sample of New York State nursing homes. We find that workplaces characterized by higher levels of employee satisfaction and employee discretion incur lower EMR implementation costs. Our estimates suggest that nursing homes that are one standard deviation above the mean in terms of employee satisfaction and employee discretion measures incur about 10% lower EMR adoption costs than average facilities. Implications for policy makers and managers are discussed.

Health Care Information Technology, Work Organization, and Nursing Home Performance

The authors investigate whether electronic medical record (EMR) systems are associated with higher levels of nursing home performance. Their difference-in-differences analysis is based on a survey of health care information technology (HIT) use in approximately 304 New York State nursing homes, combined with regulatory data from the Center for Medicaid and Medicare Studies (CMS) Nursing Home Compare database and the New York State RHCF-4 financial reports. For nursing home owners, the authors find a positive effect of EMR-system implementation, on the order of 1% higher productivity, 3% greater efficiency, but approximately 2.7% higher cost. They also find that EMR systems amplify the returns to modern workplace organization. Facilities that are one standard deviation higher on a work-organization scale—composed of practices that encourage employee collaboration, decision making, suggestions, and problem solving—have no adverse cost impact of adoption of HIT, and adoption of HIT is associated with a productivity increase of 1.5% or more. They find no evidence of an impact on health care quality.

Technical Report: The Business Case for Healthcare Information Technology in Nursing Homes

We investigate whether the implementation of electronic medical records is associated with higher levels of economic performance in nursing homes in terms of quality, profitability, cost, productivity, and efficiency. Our analysis is based on a survey of Healthcare Information Technology (HIT) usage for approximately 200 New York State Nursing Homes, including 26 that participated in the NY Nursing Home Demonstration project. The survey data are combined with regulatory data from CMS Nursing Home Compare and the NY State RHCF-4 financial reports. The analysis uses a “difference in differences approach” comparing changes in economic performance in a facility before and after HIT implementation to changes over the same time periods in other facilities that are known to not have implemented HIT. Overall, we do not find conclusive evidence that adoption of HIT raises productivity, increases commonly used performance measures, or improves quality scores in the short run. There is some evidence of a positive effect after several years. We do find that there appears to be a slight increase (1-3%) in variable costs, but also slightly greater efficiency when measured by frontier analysis. However, facilities that adopt HIT and either adopt or have in place progressive work practices (which include greater staff autonomy, cooperative labor-management relations, and greater teamwork) consistently show performance gains. Firms that are one standard deviation higher on our progressive work practice score get a 2-3% increase in productivity, a 2-3% reduction in costs and a 1-2% incremental efficiency gain upon adoption of HIT relative to their peers.Our results suggest that adoption of EMR can have significant positive economic benefits when implemented in conjunction with progressive work practices.

Data Assets and Industry Competition: Evidence from 10-K Filings

This paper constructs new, longitudinal measures of corporate data practices by conducting textual analysis of annual 10-K filings from 1996 to 2012. We provide descriptive evidence of a dramatic increase in data-related activities over this period, beginning in IT-producing firms and then moving to other sectors. Next, we demonstrate that these measures are positively and significantly correlated with Tobin’s q and higher profits relative to the mean of a firm’s six-digit NAICS industry. These findings are consistent with the argument that data assets are driving performance dispersion within IT-using industries, in which data-intensive firms are outperforming their competitors. Finally, we use dimension reduction techniques to demonstrate that premiums in market value accrued to data collection activities in the early 2000s, and to data quality initiatives (i.e., standardization and security) after 2007.