Beyond Incentive Pay: Insiders' Estimates of the Value of Complementary Human Resource Management Practices
GAICIL Insights
Ask a CEO whether incentive pay works and the honest answer is: it depends on what else is in the system.1 Beyond Incentive Pay: Insiders' Estimates of the Value of Complementary Human Resource Management Practices, by Casey Ichniowski and Kathryn Shaw, builds the evidence behind that answer. The authors ask whether incentive pay, teamwork, training and careful screening raise productivity, and under what conditions, using an approach they call insider econometrics: going deep inside businesses to obtain data and insights into how practices affect specific production processes.
The conclusion is nuanced and powerful. Sets of complementary practices appear to raise performance. Some firms benefit more than others: those making complex products, and those starting up brand new facilities, gain more. The authors stress that designing organizations to get the best performance from workers is a complex managerial issue, and convincing analysis must acknowledge that complexity. An individual worker with given education and experience can be high- or low-quality depending on the work environment.
That last sentence is a quiet revolution in how to think about talent. Worker quality is partly a property of the system the worker inhabits. Productivity depends on the match among workers, practices and production processes.
The application to AI is direct. AI enters as a powerful new component of that system. The same worker, with the same skills, may become far more productive when AI tools are paired with the right practices, or less productive when paired with the wrong ones. Asking whether AI raises productivity is like asking whether incentive pay does: the answer depends on everything around it.
Complementarity is the central lesson. HR practices work as bundles. Incentive pay is more effective with teamwork, training and careful selection; one practice alone may deliver little. AI follows the same rule. It is likely to raise productivity most alongside complementary changes in training, job design, incentives and team organization. Firms installing AI without changing practices may be disappointed. Firms redesigning their organizations around the technology may see large gains. The technology is the easy part to buy.
The finding about complex products and new facilities sharpens the point. Complex production requires coordination, problem-solving and adaptation, which teamwork and training support. Where AI makes production more complex, complementary HR practices become more valuable. New facilities, built from scratch, adopt integrated systems more easily than established ones with entrenched habits. That hints at a competitive dynamic: new firms and new plants may be best positioned to capture AI's gains combined with modern practices. The greenfield site may beat the incumbent factory at absorbing new technology.
Insider econometrics is a methodological gift for studying AI. Aggregate data struggle to capture how a technology affects specific production processes. Going inside firms, observing production and measuring how practices and technology interact can reveal the mechanisms behind gains. As firms experiment with AI, detailed process-level studies will show which combinations work.
Policy implications follow. Productivity growth depends on organizational change as well as technology. Encouraging experimentation with management practices, supporting training and avoiding rigid rules on compensation and work organization let firms find effective combinations. Regulations locking in particular practices, for example by constraining incentive pay or team structures, could impede the reorganization AI requires.
One-size-fits-all prescriptions fail here too. Because practice benefits vary across firms, mandating particular practices would help some and harm others. Flexibility lets each firm adopt the combination suited to its production process, and as AI changes processes across the economy, that flexibility becomes more valuable.
The broader message: the value of workers depends on the systems in which they work. Complementary practices raise performance, especially in complex environments. AI is a powerful new element in those systems, and its contribution will depend on whether firms build the complements that let people and technology perform at their best, and on whether policy gives them freedom to do so.
Boards might start here: budget for the reorganization as seriously as for the software. On this evidence, the reorganization is where the return lives.
Introduction
Do human resource management (HRM) practices, such as incentive pay, teamwork, training, and careful screening practices, raise productivity, and if so, under what conditions does productivity rise? Recently, this question has been a central focus in organizational and personnel economics. We emphasize the value of a new research approach – an approach we label “insider econometrics”–that is aimed going deep inside businesses to obtain data and insights into the ways in which HRM practices affect specific production processes. We conclude that sets of complementary HRM practices appear to raise performance, but that some firms, such as those that make complex products or those that are starting up brand new facilities, benefit more from these practices.
Takeaway
The adoption and performance effects of innovative human resource management practices offer a number of intriguing empirical research questions. We have emphasized in this paper the potential value of applied insider-econometric productivity analysis as a method that provides rich and accurate insights into the economic causes and consequences of human resource management practices. New insider econometric studies in other industry settings would add to currently available evidence on the adoption of innovative human resource management practices and the effects of these practices on performance.1
However, the narrow focus on single production processes required for applied insider-econometric analysis is both its chief advantage and its chief disadvantage. It provides perhaps the most persuasive estimates of the effects of human resource management practices, but is not generalizable. Thus, future empirical research in this area must continue to broaden the existing evidence, and the Alfred P. Sloan Foundation continues to provide primary support for these studies.2 In addition to new insider-econometric studies, richer and more comprehensive data from broader surveys of establishments will help produce more general results, as has been done in the recent work of Black and Lynch (2000), Bresnahan, Brynjolffson and Hitt (2002) and Brynjolffson, Hitt and Yang (2002).
We also suspect there is a fertile middle ground for interplay between insider-econometric studies and broader empirical approaches. For example, broader surveys can focus on specific industries, and these surveys can be tailored to those industries after analysts gain insights from field research in a few businesses in those industries. Alternatively, insights from insider-econometric studies should be used to direct the points of emphasis in broader surveys, and several studies along these lines are being sponsored by the Russell Sage Foundation and the Rockefeller Foundation. Another promising approach can be found in studies that link employee and employer data.3 These studies, we believe, also hold great potential for analyzing such workplace-level outcomes as productivity, skill demand, human resource management practices and information technology investments.
The problem of designing organizations to get the best performance from their workers is a complex managerial issue. Convincing economic analysis of this problem must acknowledge these complexities. It is not simply a matter of finding the optimal level of human capital investment for individual workers. An individual worker with a given amount of education and work experience can be a high- or low-quality worker depending on the nature of the work environment. The research reviewed here suggests that this problem of creating a productive work environment is a multidimensional decision that forces managers to craft a web of human resource management practices that motivate workers to contribute effort and ideas to the goals of their firms.
Notes
- Casey Ichniowski & Kathryn Shaw, Beyond Incentive Pay: Insiders' Estimates of the Value of Complementary Human Resource Management Practices, 17 J. Econ. Persp. 155 (2003), https://doi.org/10.1257/089533003321164994. Takeaway source: Directions for Empirical Research on Human Resource Management Practices, reproduced from the free-to-read version from the publisher. ↩
- The Sloan Foundation has provided the key support since 1990 through their Industry Centers and their pin-factory project at the National Bureau of Economic Research. See Helper (2000) for a description of these projects and a discussion of methodological approaches., 13. ↩
- For the United States, see Abowd, Haltiwanger and Lane (2002), and for studies of European firms, see Chennells and Van Reenan (2002), Caroli and Van Reenan (1999), Borghans and Weel (2001), Mairesse and Greenan (1999), Greenan and Guellec (1997) and Abowd, Kramarz and Margolis (1999)., 14. ↩
Cite as: Casey Ichniowski & Kathryn Shaw, Beyond Incentive Pay: Insiders' Estimates of the Value of Complementary Human Resource Management Practices, 17 J. Econ. Persp. 155 (2003), https://doi.org/10.1257/089533003321164994
Header image: Assembly MBTA Orange Line Station, July 2026.jpg by IliketrainsR211T, CC BY 4.0, via Wikimedia Commons.
