Monday, March 1, 2010

Incentives for Employees in Just-in-Time Settings

In some manufacturing environments, having workers engage in just-in-time production—maintaining production quotas without any inventory stockpiling or project overhang to the next day—can actually cause motivational problems and increase costs. The answer is to make sure employees' pay is tied to their actual productivity—and that means allowing for bad days and, consequently, some inventory build-up.

STANFORD GRADUATE SCHOOL OF BUSINESS—Manufacturing philosophy has undergone a revolution in the past few decades emphasizing continuous improvement, low-inventory techniques, and increased collaboration, cross-training, and problem-solving by workers. In the 1990s, the American economy steadily trended toward lower inventory levels and higher firm and labor productivity.

But what happens in the real world does not always correspond tidily to theories about how growth in manufacturing may best be achieved. Using these practices, some companies show mixed results or no improvement while others are finding that inventory reductions actually lower productivity. New research emerging from Stanford provides a clue as to why.

A recent mathematical analysis of company microdata shows that in some manufacturing environments, having workers engage in "just-in-time" (JIT) production—maintaining production quotas without any inventory stockpiling or project overhang to the next day—can actually end up causing motivational problems and increasing costs. The answer is to make sure employees’ pay is tied to their actual productivity—and that means allowing for bad days and, consequently, some inventory build-up.

The challenge with new, leaner manufacturing practices, say the investigators, is to rely on accurate, up-to-the-minute information about production conditions on any given day. Is equipment functioning properly? Have all supplies arrived? Much of that information is reported by workers on the factory floor—meaning it is in the hands of workers rather than managers. Proper control systems, the study shows, are crucial for motivating workers to use their localized information to further the firm's best interests as opposed to their own.

"Say a worker is required to assemble a certain number of printers in a given day," says Richard Saouma, PhD '06, an assistant professor of accounting at UCLA and coauthor of the study. "To make the just-in-time work environment function well, you need to know how productive that person is going to be so you don't end up with unused inventory lying around at the conclusion of the day. But if you ask him, 'Are you going to be productive or not today?' he may be tempted to just say 'no,' earning the same pay for less work. That results in less efficiency for the firm. You have to make sure it’s in his best interest to be honest."

The paper, coauthored by Madhav Rajan, the Gregor G. Peterson Professor of Accounting in the Stanford Graduate School of Business, and Professor Venky Nagar of the University of Michigan, shows that aligning worker and corporate best interests can be managed through communications-based salary contracts. That is, workers' pay should be based on how much inventory they request at the start of the day, as well as how much they ultimately complete. Work schedules are designed around what an employee requests, perhaps via a morning email, even though his actual productivity is not determined until the work begins.

"If the employee says, 'Today all systems are go, and I’ll be very productive,' then the manager sends additional work to the employee," explains Saouma. “If it turns out the employee is not productive that day, a flexible control system would allow him not to complete the work. Allowing such flexibility and paying employees based on their initial requests combined with their actual completion rates empowers them to 'communicate' honestly how productive they think they'll be at the start of the day, as well as how much they've accomplished once the work lands on their desk." In order to use all this information, the employee must be allowed to incur occasional inventory buildups, which means not producing exactly as much as hoped on a particular day.

Surveying employees for information, however, is not without cost. Managers have to make sure that employees are rewarded for honesty, especially when telling the truth involves more work. By asking both an employee and those nearby for productivity forecasts, the researchers find an additional benefit to cross-training. "It's another set of eyes, which keeps employees more honest," says Saouma. "It prevents them from claiming, say, that that they're unable to work at the regular pace because the computers are down."

With or without cross-trained employees, the researchers suggest that left-over inventory is not always a bad thing. On the contrary, surplus inventory reflects the fact that things can go wrong––equipment might break down, and the worker may not be able to complete predicted tasks. Without having a buffer—knowing that something can wait until tomorrow if absolutely need be—an employee might be forced to use another, inefficient piece of equipment, or take other costly measures such as working overtime, outsourcing the job, or even taking shortcuts.

Quality-oriented firms have traditionally avoided JIT practices. For example, Ferrari production workers test every shipment of incoming aluminum prior to casting motors. If the aluminum is found to have excessive impurities, then the metal is used in less critical components, potentially delaying the casting process. If Ferrari used strict JIT goals, then employees would be compelled to use the substandard metal, and the resulting warranty claims could be catastrophic. Thus, many just-in-time environments would function better if they could encourage the use of timely employee productivity information, which requires both communication-based contracts and a good dose of flexibility.

"The more critical employee information is for production, the more often we’re going to observe left-over inventories. It's not exactly in line with what the bestselling books are saying," Saouma concludes.