Monday, February 17, 2014

Exploring GM's Decline: Trust vs. Legal Contracts

US Dept. of Commerce Chief Economist Susan Helper and HBS Professor Rebecca Henderson have written a new working paper examining the decline of General Motors.  They review many of the traditional explanations for the firm's demise, and they try to dig deeper to understand why GM failed to change sooner and more effectively in response to external threats.  Helper and Henderson base their explanation on the concept of "relational contracting" as explained here:

Here we make the case that GM struggled for so long because Toyota’s practices were rooted in the widespread deployment of effective relational contracts agreements based on subjective measures of performance that could neither be fully specified beforehand or verified after the fact and that were thus enforced by the shadow of the future and that GM’s history, organizational structure and managerial practices made it very difficult to maintain these kinds of agreements either within the firm or between the firm and its suppliers.

To step back, the scholars first argue that it was very difficult for GM to understand precisely what were the secrets to Toyota's remarkable success.   Strategy scholars refer to this barrier to imitation as "causal ambiguity."  In other words, the precise drivers of competitive advantage are not well understood by those outside the firm.  In hindsight, of course, we can explain Toyota's success with relative ease. At the time, though, the details of how the Toyota Production System worked, and how it could be imitated, were very difficult to ascertain.   Observation of the system at work, perhaps by touring the factories at Toyota, would do you no good.  Much deeper research was required.  

Beyond that, the scholars argue that emulating Toyota, even after an understanding had developed, was very challenging.  Here they base their argument on the relational contracts concept.  They describe this problem both with regard to how GM related to its employees, as well as to its suppliers. Here's an excerpt from the working paper, focusing on the workers in GM's factories:

It was, for example, very difficult to specify under exactly what circumstances a worker should pull the andon cord, or what behaviors constituted being an effective team member. Shutting down the line for a popular model could cost $10,000 in lost profits per minute (Helper 2011), so management setting up this system needed to be confident that a worker deciding to pull the andon cord would have both the knowledge and the incentive to exercise sophisticated judgment. Conversely, workers would only pull the cord if were confident that an appropriate relational contract wasin place (Gibbons and Henderson 2013). Similarly MacDuffie’s(1997) detailed description of the practices underlying shop-floor problem solving in the industry suggests that successful process quality improvement depended on processes that allowed for the inclusion of multiple perspectives on any single problem, the use of problem categories that were “fuzzy,” and the development of a common language for discussing problems. It seems implausible that employees could be motivated to participate in these kinds of activities through the use of formal contracts that specified in advance every kind of quality problem and its appropriate response.

Why did GM have a hard time building relational contracts?  Helper and Henderson offer several reasons.   In my view, the most compelling explanation is that GM lacked the credibility to work in this very different manner with both the employee unions and the external suppliers.  The lack of trust precluded working in this manner.  Toyota management, on other hand, had developed a deep reservoir of trust from which it could work much more flexibly with workers and suppliers.  Put simply, there are two ways to make a relationship work: trust vs. traditional legal contracts.  GM relied on the latter, but emulating Toyota required the former.  It could not make the switch.

No comments: