How Does Data-Driven Decision-Making Affect Firm Productivity and CEO Pay? Erik Brynjolfsson, MIT Sloan Lorin Hitt, Wharton, School, University of Pennsylvania Heekyung Kim, MIT Sloan September 15, 2010 Preliminary and Incomplete Draft For consideration at WISE 2010 Comments welcome. Please do not quote. Abstract While there is a great deal of anecdotal evidence that firms can boost their performance by adopting a data driven decision-making (DDD) approach, there is little data or systematic analysis of these claims themselves. We gather detailed information on the business practices and information technology investments of 165 large publicly traded firms and find that DDD can explain a 3-5% increase in their output and productivity, beyond what can be explained by traditional inputs and IT usage. Furthermore, firms with more DDD tend to have a higher degree of consistency in business practices across business units and geography and stronger linkages between business and technology. In addition, we find that DDD is correlated with a significant increase in CEO pay even after controlling for the average worker’s wage, suggesting that the data-driven decision-making may further widen the gap between top managers and average workers. Intriguingly, the quadratic effect of DDD is positive for CEO compensation, but not for firm productivity, echoing Rosen’s (1982) model of increasing rewards to “superstars”.
29
Embed
How Does Data-Driven Decision-Making Affect Firm ...misrc.umn.edu/wise/papers/1a-1.pdf · How Does Data-Driven Decision-Making Affect Firm Productivity and CEO Pay? ... We seek to
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
How Does Data-Driven Decision-Making
Affect Firm Productivity and CEO Pay?
Erik Brynjolfsson, MIT Sloan
Lorin Hitt, Wharton, School, University of Pennsylvania
Heekyung Kim, MIT Sloan
September 15, 2010
Preliminary and Incomplete Draft
For consideration at WISE 2010
Comments welcome. Please do not quote.
Abstract
While there is a great deal of anecdotal evidence that firms can boost their performance by
adopting a data driven decision-making (DDD) approach, there is little data or systematic
analysis of these claims themselves. We gather detailed information on the business practices
and information technology investments of 165 large publicly traded firms and find that DDD
can explain a 3-5% increase in their output and productivity, beyond what can be explained by
traditional inputs and IT usage. Furthermore, firms with more DDD tend to have a higher degree
of consistency in business practices across business units and geography and stronger linkages
between business and technology. In addition, we find that DDD is correlated with a significant
increase in CEO pay even after controlling for the average worker’s wage, suggesting that the
data-driven decision-making may further widen the gap between top managers and average
workers. Intriguingly, the quadratic effect of DDD is positive for CEO compensation, but not for
firm productivity, echoing Rosen’s (1982) model of increasing rewards to “superstars”.
1. Introduction
Each revolution in science is typically preceded by a revolution in measurement and data
gathering. The invention of microscope, for example, allowed the exploration of an entire new
world that had been invisible, transforming biology and medicine. Today’s data revolution
promises to transform management in an equally fundamental way. Businesses now can
measure consumer behavior, business operations and myriad other aspects of their environment
with unprecedented precision. For instance, clickstream data provides an unprecedented digital
trace of customer actions and ERP systems generate terabytes of data as byproducts of their
normal activities, while email, RFID, mobile phones, and many other information technologies
are all contributing to a doubling of the quantity of business data roughly every 1.2 years.
Data by itself does not change anything. It must be analyzed and used to change decisions to
have any value. The ability to collect and analyze the enormous amount of data is, therefore,
becoming of critical importance for firms to stay competitive and gain market share. Ironically,
although there is a great deal of anecdotal evidence of firms’ using data to gain a competitive
edge in the business press in popular books, there has been virtually no systematic data analysis
of the productivity effects of data-driven decision making using statistical methods. We seek to
address this gap by examining detailed business practices and information technology investment
of publicly traded large 165 firms in the US. We find that approximately 3-5% increase in the
productivity can be explained by data-driven decision-making (DDD) and that the top executives
in data driven firms, particular CEOs, have reaped disproportionately large rewards in terms of
personal compensation. Our results can help explain not only the relatively higher productivity
growth rates that IT-intensive firms have recently experienced, but also the growing gap between
CEO pay and the compensation of ordinary workers.
As information technology has penetrated every step of business practices, business decisions are
increasingly relying on data and experiments. Consumer-facing companies rich in transaction
data are routinely testing innovations and randomized testing is standard procedure in online
firms such as Amazon, eBay, and Google. For example, Amazon.com continually experiments
with their websites for the best appearance. “We have been implementing changes on <online
site> based on opinion, gut feeling or perceived belief. It was clear that this was no way to run a
successful business…Now we can release modifications to the page based on purely on
statistical data” (Kohavi, Longbotham et al., 2009). Kohavi et al. (2009) illustrates how data can
overrule the traditional approach of relying on intuition and “the highest paid person’s opinion”
(HiPPO) for corporate decision-making. “Greg Linden at Amazon created a prototype to show
personalized recommendations based on items in the shopping cart. Linden notes that while the
prototype looked promising, a marketing senior vice-president was dead set against it, claiming it
will distract people from checking out. Greg was forbidden to work on this any further.
Nonetheless, Greg ran a controlled experiment” to prove that the new feature would bring in a lot
more sales at a statistically significant level and the HiPPO was wrong.
This suggests that information technology, once thought of a major driver in widening income
gap between college and high school graduates by replacing workers of routine and simple
clerical jobs (i.e., Autor, Katz et al., 1998) is now poised to replace HiPPOs. On the contrary,
however, we have been witnessing that a further widening income gap in every sector and every
income percentile of the population in recent years (i.e., Piketty and Saez, 2003). In this study
we report that the data-driven decision-making is correlated with the increase in CEO’s pay even
after controlling for the average worker’s wage, suggesting that DDD is further widening the
income gap between the top manager and other workers. Our interpretation of this finding is that
information technology is eroding the high-level managers’ work as it has replaced the workers
of routine jobs (Autor, Katz et al., 1998), shrunken the ranks of middle managers (Pinsonneault
and Kraemer, 1997), and flattened hierarchies (Rajan and Wulf, 2006). At the same time,
however, it is complementing the CEO’s job. A CEO’s vision for the future and steering her
firm to the future cannot be experimented and replaced by data. Rather, talented CEOs are even
more in demand to make sound decisions based on the ever-increasing amount of information, to
shape her firm to be data-driven and encourage her employees to experiment and collect and
analyze data, to exploit the capability of information technology to gain a competitive advantage.
2. Theory
The starting point for economic theories of the value of information typically begins with the
seminal work of Blackwell (1953). Blackwell’s theorem (Blackwell, 1953) states that decision
makers observe signals correlated to the state of nature prior to their choice of action and hence
“update” the probability distribution before the “optimal” action is chosen. In this framework
the value of information (in terms of expected utility) is always positive. Since the advent of
computers in the 1980s, firms have deployed enormous resources to computerize work and
collect information. Many researchers, however, showed that only organizations with certain
characteristics such as high endowments of human capital and decentralized work practices tend
to reap the benefit of their information technology investment (i.e., Bresnahan, Brynjolfsson et
al., 2002). Our study is related to the stream of literature, uncovering another organizational
characteristic that may be associated with this increase in productivity. Controlling for firms’
human capital and information technology investments, we show that data-driven decision-
making practices (DDD) significantly affect firm productivity, and are also highly correlated
with the consistency of business practices across business units.
The increase in the amount of available information may not be always valuable when acquiring
and processing the information is too costly. “The real issue is not the lack of data, but our
ability to properly process and distribute the information quickly. If we cannot properly process
and distribute the information quickly, we run the risk of destroying the value of the information
we are striving so hard to attain.” (Collins, 2010). CIOs have expressed concerns that after
investing in data collection, they lack the capability and personnel to analyze these data.1 Our
study explores the question by empirically examining whether there is an optimal level of DDD.
While the answer to the question is inconclusive due to our limited sample size, a striking result
emerged – the reward to CEOs shows increasing returns to investments in DDD. For instance,
we find that firms at two standard deviations above the mean of our DDD scale have CEO
compensation four times larger than the average firm.
CEO compensation might be related to DDD because it affects the ability of organizations to
transfer and process knowledge. In particular, information technology makes some types of
information alienable from workers or mid-level managers (see e.g. Brynjolfsson, 1993) but not
necessarily the most senior managers. This increases the marginal value of decisions made by
CEOs.
1 CIO forum, MIT center for digital business, 2010.
Numerous researchers have noted that some information is not easily transferable because it
“tacit” (i.e., Polanyi, 1958; Rosenberg, 1982) or “specific” (i.e., Jensen and Meckling, 1976) or
“sticky”(i.e., Von Hippel, 1994). Advancement in information technology has, however, brought
in a wave of codification of knowledge. Balconi, 2002) presented an illustrative example of the
changes in the progressive codification of technological knowledge occurred over the last 30
years formerly embodied in skilled workers in the steel industry. Until the end of the 1960s
instruments to measure the temperature and chemical composition of liquid steel were not widely
available so measurements were carried out based on some physical characteristics observable by
sight. For example, “in order to know the temperature of liquid steel, a sample was taken out of
the furnace, poured upon an iron plate and the temperature was deduced by observing the
forming of the spot, its shape and the way it solidified and attached itself to the plate.” The
ability to recognize the temperature by sight was acquired with 5 years of experience or more.
By the 1970s the content and the temperature of liquid steel was routinely conducted by
automated measurement tools but the process was still controlled by workers and line managers.
By the end of 1980s the whole processing cycle was automated. Since the mid-90s, workers are
principally involved in monitoring automated data collection and processing equipment, and
approving process changes recommended by the processing algorithms. Over time, the
knowledge once embodied in experienced workers has been supplanted by machine monitoring
and largely automated decision making. A consequence is that the specialized knowledge only
possessed by the most skilled individuals can now be embedded in computer instruments and
available to distant operators in the operating rooms throughout the entire company. More
broadly, firms now gather and propogate knowledge not only from production workers, but from
their consumers, suppliers, alliance partners, and competitors much faster and more cheaply with
the aid of information technology, make information available for distant decision makers.
Knowledge generation does not stop at passive analysis of data. New and broadly available
software has enabled managers to conduct active experiments with their new business ideas and
base their decisions on scientifically valid data. From banks such as PNC, Toronto-Dominion,
and Wells Fargo to retailers such as CKE Restaurants, Famous Footwear, Food Lion, Sears, and
Subway to online firms such as Amazon, eBay, and Google, firms test many business ideas
through a randomized test before launch, called as “information-based strategy” (Davenport,
2009). This information-based strategy alienates the high-level manager’s tacit knowledge – the
knowledge of what was or was not likely to be a successful business innovation. Just as the
process of controlling steel processes was transformed by improved measurement of temperature,
the innovative process in an online business is now being transformed by the improved
measurement of customer preferences. However, this process automation cannot substitute for
strategic decision making, although it does allow strategic decisions to be moved from managers
at the level of a task or a business unit, to a manager with responsibility for the entire firm.
Consequently, a small difference in the talent of top manager who has an enormous amount of
information to process may make a huge difference to the fortune of the firm.
Our study is also related to the effects of IT on command, control, coordination, and organization
of firms (see Leavitt and Whisler, 1958;Rule and Attewell, 1989;Gurbaxani and Whang,
1991;Malone, Yates et al., 1987;Brynjolfsson and Mendelson, 1993;Brynjolfsson, Malone et al.,
1994;Brynjolfsson and Hitt, 2000 and the literature reviews cited therein). In theory, IT could
shift power either toward the center or away from it, leading to centralization or decentralization
of a firm. If IT can transfer knowledge to central managers, a firm can become more centralized
which would suggest higher CEO relative pay. On the other hand, if IT facilitates lateral
communication of local knowledge between employees, employees can coordinate tasks among
themselves more easily with the need for senior management involvement, suggesting lower
CEO relative pay. Researchers have found evidence of decentralization at task level, but the
increase in the scope of CEO as well (Rajan and Wulf, 2006). Our finding is consistent with the
increase in the scope of CEO, but not as a substitute for decentralized work practices. Data
driven firms generally invest in both decentralized work practices, but with greater overall
centralized coordination of these activities, making them consistent across the firm.
Our study is also related to the literature on IT and wage inequality (i.e., Autor, Katz et al., 1998).
Our view on the role of IT in increasing CEO pay is close to those of Garicano (Garicano and
Rossi-Hansberg, 2006;Garicano, 2000). Their model suggests a hierarchy of knowledge. Some
knowledge resides in lower level employees and is used to solve routine problems, while other
knowledge resides in higher-level employees that is primary directed at solving non-routine
problems. As the cost of communication among agents decreases, the complicated problems that
lower-ranked employees cannot solve can be easily passed to their superiors. Once solved, the
solution can be disseminated throughout the firm and become part of routine work practice. This
can lead to the dependency of the problem-solving on a few “superstars” (Rosen, 1981) and thus
a higher wage for the superstars. Their explanation is consistent with our view that the
availability of data allows a distant few problem solvers to provide a solution to non-trivial
problems and increase their marginal productivity disproportionately. Therefore, data driven
firms become dependent on a small number of superstars which should be associated with
greater pay. CEOs may or may not be one of these problem solvers, but they are nonetheless
responsible for coordinating, managing, and rewarding their activities, which also increases the
marginal value of their effort.
Finally, our study is related to the literature on CEO compensation, especially studies that
document or explain the observed increase in CEO pay. Following Gabaix and Landier’s (2008)
summary of the literature on the topic, we recapitulate the literature in four views. The agency
view (Jensen and Murphy, 1990;Dow and Raposo, 2005;Holmstrom and Kaplan, 2001) suggests
that increased expected compensation is necessary to provide incentives in an increasingly
volatile business environment and to reward difficult to measure activities like innovation. The
market failure view (Yermack, 1997;Bertrand and Mullainathan, 2001;Bebchuk and Fried,
2005;Bebchuk, Fried et al., 2002;Hall and Murphy, 2003) suggests that the rise in CEO pay is a
consequence of failure of corporate governance mechanisms. A third possibility is that the CEO
job has changed (Frydman, 2005) to become more general which has increased their outside
options and put upward pressure on pay. Finally, the market equilibrium view (Gabaix and
Landier, 2008;Tervio, 2008) suggests that CEO pay increases simply reflect an increase in firms
size. In equilibrium, the most talented CEO is correctly matched to the largest firm and paid the
highest because the marginal value of their decisions is greatest there.
Our paper is most closely linked to the market equilibrium view; the increased marginal
productivity of CEOs resulted from the increase of firm size is a major determinant of the recent
rise of CEO pay. We articulate that what is relevant to the marginal productivity of CEO is an
effective size, of which data-driven firm is greater than others of the same nominal size. The
more data-driven a firm is, the easier the flow and access of knowledge and thus the larger the
effective size of the firm that distant managers can control. As a result, the marginal productivity
of CEO increases and this drives an increase in CEO pay.
Our argument is related to the study by Baker and Hall (Baker and Hall, 2004), who attributed to
a specific mechanism of increased marginal productivity. In particular, they argue that when
CEOs perform tasks that affect the entire firm, optimal compensation should be connected to the
proportional change in value rather than the absolute change in value. As firms become larger,
this implies larger CEO pay. Since modern investments in DDD tend to lead to firm wide
changes or standardization of processes across the firm, the Baker and Hall argument would
suggest that DDD leads to greater CEO compensation.
One challenge to investigating these theories is that different firms, in different vary in the
marginal productivity and therefore wages of all workers (Katz and Summers, 1989;Krueger and
Summers, 1987;Krueger and Summers, 1988;Gibbons and Katz, 1992). For example, the
median workers’ wage in the industry of computer systems and related services, $65,000, was
over 3-times higher than that in the industry of food services, $18,000, in 2005. This variance is
at least partly due to the fact that DDD can raise the productivity of all workers. However, our
theory predicts not just increases in wages, but disproportionate increases in the wages of top
executives. Therefore, to isolate this effect, we will control for wages of other workers in our
analysis.
Data
Our business practice and information system measures are estimated from a survey
administered to senior human resource (HR) managers and chief information officers (CIO) from
large publicly traded firms in 2008. We received matched responses (both HR and CIO) from
127 firms, HR only responses from 122 firms, and only CIOs-only responses from 81 firms. The
survey asks over 80 questions on business practices as well as organization and usage of
information systems. The questions extend a previous wave of surveys on IT usage and
workplace organization administered in 1995-1996 and 2001 (Hitt and Brynjolfsson, 1997
Tambe, Hitt et al., ), but adds additional questions on organizational structure, the usage of
information for decision making, and the consistency of their organizational practices. To test
our hypothesis, we used the survey response to construct measures of firms’ organizational
practices. We combine these measures with publicly financial data and CEO compensation.
This yielded 164 firms with complete data for an analysis of firm productivity, and 165 firms
during the same period with the variables needed to analyze CEO compensation. Our sample
spans manufacturing, retail/wholesale trade, information, and finance/insurance industries over
the period from 2005 to 2009.
Business Practices
We constructed our key independent variable, data-driven decision-making (DDD), from three
questions of the survey: 1) the usage of data for the creation of a new product or service, 2) the
usage of data for business decision-making in the entire company, and 3) the existence of data
for decision-making in the entire company. Two other measures from the survey were also
constructed to indicate the general employee’s human capital; 1) the importance of typical
employee’s education and 2) the average of % of employees using e-mail and % of employees
using PCs, workstation, or terminal. All measures except the ones based on percentages were
capture in 5-point Likert scales.
We created DDD by first standardizing each factor with mean of zero and standard deviation of 1
and then standardizing the sum of each factor for each composite measure:
DDD = STD(STD(use of data for creation of a new product/service) + STD(use of data for
business decision in the entire company) + STD(existence of data for such a decision))
To examine the factors which may be correlated with DDD, we constructed a number of other
organizational measures listed in Table 1. “Consistency” refers to the consistency of business
practices in the entire company. The consistency measure was constructed from a composite of
responses to six survey questions on consistency of business practices across operating units,
within business units, across functions, and across geographies (4 questions); the effectiveness
of IT for supporting consistent practices; and consistency of prioritization of projects (Table 1).
“BT-link” captures the integration of business strategy with information technology and is
derived from size survey questions focusing on the interaction of employees from both groups,
understanding of each other, and the importance of technology in strategy. “Dissemination” was
based on two questions focusing on embedding processes in technology and the ability of the
firm to disseminate business practices across the firm.. “IT-governance” is a measure on how IT
project is governed in terms of prioritization, involvement with business, and managing demand
and benefit. “Find” is the effectiveness of employees’ finding knowledge or a colleague with a
particular expertise or broadcasting their expertise. “Exploration” is the capability of innovating
and creating a new business capability, product or service, a composite from 7 survey questions.
“Exploitation” is the ability of improving their existing capability, a composite from 15 survey
questions. The number of instances of IT systems on ERP/SCM, CRM, HR, and Financial
Systems and the level of centralization of IT resources on application development and
maintenance was used to construct a composite designated “Centralized IT system”. We
capture employee human capital a question about the importance of typical employee’s
education and the percentage of employees using emails or pc were included.
Financial data
Financial measures were derived from Compustat. Measures of physical assets, employees, sales
and operating income were taken directly from the Industrial annual file from 2005 to 2009.
Materials were estimated by subtracting operating income before tax and labor expense from
sales. In the case that labor expense was not available, the industry-average was used with the
industry average based on average labor expense for all firms that report labor expense and
employees in the same industry at the most detailed industry breakdown available.
CEO compensation
We used Compustat database for Executive pay for the period from 2005 to 2009 which provides
data on compensation of as many as 13 senior executives from each company. The Executives
database is compiled from proxy statements filed by the companies in compliance with
Securities and Exchange Commission (SEC).
The two measures of CEO compensation were used; one is the variable, tdc1, from Compustat
Executives data set and the other is the sum of salary and bonus. The tdc1 includes salary, bonus,
other annual, restricted stock grants, LITP payouts, all other, and value of option grants valued at
the day of grants.
For a variable for firm’s nominal size, we used three variables; total employee number, sales, or
market capitalization. Market capitalization was estimated in the same way as Gabaix and
Landier (2008) as the market value of equity at the end of the fiscal year, plus debt less deferred
taxes.
IT-employee data
The survey included the questions about IT budgets, outsourcing, change of IT budgets from
2008 to 2009, and full time IT employment. The number of full-time IT employees for the year
2008 was asked in the survey, but for the year 2009 it was estimated from the questions on IT
budget. Using the change of IT budget from 2008 to 2009, the percentage of outsourcing, and IT
FTE for 2008, we were able to estimate the IT FTE for the year 2009. The year from 2005 and
2006, we used data collected in a previous study (Tambe and Hitt, ). For the year 2007, a
moving average from 2005, 2006, 2008 and 2009 was used. The number of non-IT employees is
equal to the number of employees reported on Compustat less our computed IT employment
measure.
3. Methods
Productivity Tests
We use the Cobb-Douglas specification, the most commonly used model in information
technology and productivity literature (e.g., Brynjolfsson and Hitt, 1993, 1996; Dewan and Min,
1997). Our primary regression model can be written as the following:
Partial correlation, controlling industry at 2-digit NAICS. ***p<0.001, **p<0.05, *p<0.1. Test is against the null hypothesis that the correlation is zero.
Table 4. Robust check on the assumption that data-driven decision-making (DDD) is quasi-fixed over the last
5 years. Firms were categorized into 5 groups based on their response to a question on the change in the
consistency of business practices. The question was “To what extent do the following statements describe the
work practices and environment of your entire company; our business processes have become more
consistent over the past 3 years.” The choices were from 1 as “Describes not all” to 5 to “Completely
describes”. The first column is a model controlling for the group and the second not controlling for the
Importance of Typical Employee's Education 0.128*** 0.122**
(0.0490) (0.0486)
Ln(% of Employees using PC/e-mail) -0.0242 -0.0274
(0.1110) (0.1110)
Ln(Average Worker's Wage) 0.0527
(0.0503)
Constant 4.189*** 4.189*** 3.603***
(0.4030) (0.6940) (0.8630)
Observations 683 663 663
R-squared 0.337 0.357 0.358
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
5. References
Autor, D., L. Katz, et al. (1998). "Computing Inequality: Have Computers Changed the Labor Market?*." Quarterly Journal of Economics 113(4): 1169-1213.
Autor, D. H., L. F. Katz, et al. (1998). "Computing Inequality: Have Computers Changed the Labor Market?" Quarterly Journal of Economics 113(4): 1169-1213.
Baker, G. and B. Hall (2004). "CEO incentives and firm size." Journal of Labor Economics 22(4): 767-798.
Baker, G. P. and B. J. Hall (2004). "CEO Incentives and Firm Size." Journal of Labor Economics 22(4): 767-798.
Balconi, M. (2002). "Tacitness, codification of technological knowledge and the organisation of industry* 1." Research Policy 31(3): 357-379.
Bebchuk, L. and J. Fried (2005). "Pay Without Performance: Overview of the Issues." Journal of Applied Corporate Finance 17(4): 8-23.
Bebchuk, L., J. Fried, et al. (2002). "Managerial power and rent extraction in the design of executive compensation." University of Chicago Law Review 69: 751-846.
Bertrand, M. and S. Mullainathan (2001). "Are CEOS Rewarded for Luck? The Ones without Principals Are*." Quarterly Journal of Economics 116(3): 901-932.
Blackwell, D. (1953). "Equivalent comparison of experiments." Ann. Math. Statist. 24: 265-272.
Bresnahan, T., E. Brynjolfsson, et al. (2002). "Information Technology, Workplace Organization and the Demand for Skilled Labor: Firm-level Evidence." Quarterly Journal of Economics 111(1): 339-376.
Brynjolfsson, E. and L. Hitt (2000). "Beyond computation: Information technology, organizational transformation and business performance." The Journal of Economic Perspectives: 23-48.
Brynjolfsson, E., T. Malone, et al. (1994). "Does information technology lead to smaller firms?" Management Science: 1628-1644.
Brynjolfsson, E. and H. Mendelson (1993). "Information systems and the organization of modern enterprise." Journal of Organizational Computing 3(3): 245-255.
Collins, D. (2010). "Destroying the value of information." Journal of Management Excellence: The Value of Information(9).
Davenport, T. (2009). "How to design smart business experiments." Harvard Business Review.
Dow, J. and C. Raposo (2005). "CEO compensation, change, and corporate strategy." Journal of Finance: 2701-2727.
Frydman, C. (2005). Rising through the ranks: The evolution of the market for corporate executives, 1936-2003. Working Paper, Harvard University.
Gabaix, X. and A. Landier (2008). "Why has CEO pay increased so much?" Quarterly Journal of Economics: 49-99.
Garicano, L. (2000). "Hierarchies and the Organization of Knowledge in Production." Journal of Political Economy 108(5): 874-904.
Garicano, L. and E. Rossi-Hansberg (2006). "Organization and Inequality in a Knowledge Economy*." The Quarterly Journal of Economics 121(4): 1383-1435.
Gibbons, R. and L. Katz (1992). "Does Unmeasured Ability Explain Inter-Industry Wage Differentials?" Review of Economic Studies 59(July): 515-35.
Gurbaxani, V. and S. Whang (1991). "The impact of information systems on organizations and markets." Communications of the ACM 34(1): 59-73.
Hall, B. J. and K. J. Murphy (2003). "The Trouble with Stock Options." Journal of Economic Perspectives 17(3): 49-70.
Hannan, M. and J. Freeman (1984). "Structural inertia and organizational change." American sociological review 49(2): 149-164.
Hitt, L. and E. Brynjolfsson (1997). "Information technology and internal firm organization: an exploratory analysis." Journal of Management Information Systems 14(2): 101.
Holmstrom, B. and S. Kaplan (2001). "Corporate Governance and Merger Activity in the United States: Making Sense of the 1980s and 1990s (Digest Summary)." Journal of Economic Perspectives 15(2): 121-144.
Jensen, M. and K. Murphy (1990). "Performance pay and top-management incentives." Journal of Political Economy 98(2): 225.
Jensen, M. C. and W. H. Meckling (1976). "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics 3(4): 305-60.
Katz, L. and L. Summers (1989). "Industry Rents: Evidence and Implications." Brookings Papers on Economic Activity: Microeconomics: 209-75.
Kohavi, R., R. Longbotham, et al. (2009). "Controlled experiments on the web: survey and practical guide." Data mining and knowledge discovery 18(1): 140-181.
Krueger, A. and L. Summers (1987). Reflections on the Inter-Industry Wage Structure. Unemployment and the Structure of Labor Markets. K. Lang and J. Leonard. Oxford, Basil Blackwell.
Krueger, A. and L. Summers (1988). "Efficiency Wages and the Inter-Industry Wage Structure." Econometrica 56(March): 259-93.
Leavitt, H. and T. Whisler (1958). "Management in the 1980's." Harvard Business Review November-December: 41-48.
Malone, T., J. Yates, et al. (1987). "Electronic markets and electronic hierarchies." Communications of the ACM 30(6): 455.
Milgrom, P. and J. Roberts (1990). "The economics of modern manufacturing: technology, strategy, and organization." The American Economic Review: 511-528.
Piketty, T. and E. Saez (2003). "Income Inequality in The United States, 1913-1998*." Quarterly Journal of Economics 118(1): 1-39.
Pinsonneault, A. and K. Kraemer (1997). "Middle management downsizing: an empirical investigation of the impact of information technology." Management Science 43(5): 659-679.
Polanyi, M. (1958). Personal knowledge, Routledge London.
Rajan, R. and J. Wulf (2006). "The flattening firm: Evidence from panel data on the changing nature of corporate hierarchies." The Review of Economics and Statistics 88(4): 759-773.
Rosen, S. (1981). "The Economics of Superstars." American Economic Review 71(5): 845-857.
Rosenberg, N. (1982). Inside the black box: technology and economics, Cambridge Univ Pr.
Rule, J. and P. Attewell (1989). "What do computers do?" Social Problems: 225-241.
Tambe, P. and L. Hitt (2008). "Job Hopping, Knowledge Spillovers, and Regional Returns to Information Technology Investments." Working paper.
Tambe, P., L. Hitt, et al. (2009). "The Extroverted Firm: How External Information Practices Affect Productivity." working paper.
Tervio, M. (2008). "The difference that CEOs make: an assignment model approach." American Economic Review 98(3): 642-668.
Von Hippel, E. (1994). "" Sticky information" and the locus of problem solving: Implications for innovation." Management Science 40(4): 429-439.
Yermack, D. (1997). "Good timing: CEO stock option awards and company news announcements." Journal of Finance: 449-476.