Just to be clear, the theory discussed in this post comes from joint work with David Hoopes; we have a draft of a paper on “An Engineering Science of Capabilities.” The modeling paper to which you refer is my attempt to formalize some of the ideas in the joint paper. Let me respond a bit to your comments.

On 1) and 2) You have it exactly right that this is a normative theory of what managers might be able to do to make for more effective capabilities. (BTW, they do other things besides this, too, such as figure out ways to deploy capabilities.) Our analogy is to engineering problems, such as designing engines. The laws of thermodynamics give a theroetical maximum efficiency for an engine. Actual engine designs fall far short of that theoretical maximum because of waste heat radiation, friction within the engine, material flexing, and all the other sources of energy losses.

The idea is to develop some science about each of these kinds of losses. Such engineering science will never be able to replace the art and craft of design itself, which encompasses complex contextual factors and endless possibilites for interaction among the parts of whatever is designed. But it provides very useful guidance to anyone trying to practice the craft of design.

I’m perfectly comfortable with this kind of theory. If managers are really good at understanding this engineering science of capabilities already, and they are motivated to apply it, then we have a good positive description of the world. To the extent that they do not, we have a chance to improve the world, or at least consulting opportunities.

In terms of the concrete actions managers can take to deal with the identified impediments to capability, these are pretty well known. (In keeping with the family theme from Pam #1, my father was wont to answer questions about what he did at work with “I talked on the telephone and wrote with a pencil.”) Without getting too granular, we have some basic actions: Informing some agents about the actions of other agents, setting down schedules and milestones, translating information from one domain to another, publicly articulating goals, monitoring agents for shirking or effort distortion, rewarding or punishing agents, resolving disputes, etc.

As for why they don’t always do a good job of this, maybe it’s our field’s fault, because no one has yet developed a theory that makes it easy enough to figure out when to intervene and which way to intervene. Or maybe it’s the fallen nature of humanity, or the fact that the managers themselves are subject to the same individual and collective foibles as those they manage (“Who will guard those selfsame guardsmen?”)

3) and 4) This is a theory of management, not a theory of the firm, if you mean an explanation of firm boundaries as laid out in the Coase-Williamson-Grossman&Hart-Hart&Moore-etc. tradition. Boeing engages in heavy-duty management of its component vendors, who are separate firms; Toyota does the same with its suppliers. Conversely, units within a single firm may interact in a much more arms-length way without significant management integration effort.

There is definitely some overlap with older work in all this. If I go back and read Barnard, for example, I’m struck by similarities to some of these ideas. Trans-specialist understanding, though, need not be relationship-specific (that’s a contingency that might be of great importance in setting up one’s knowledge accumulation path, though).

I think the novelty of our approach is in a) focusing sharply on the integrating role of management, b) looking at integration failure and finding regularities there instead of looking at integration success, c) trying to be very precise and specific about the symptoms of integration failure, and d) trying to focus on a small number of proximate causal mechanisms for those failures. Those features make it possible to reason carefully about, and even model, how integration does or doesn’t make things better.

I will look at Gibbons’s aritcle again. It’s a good idea to see if there’s some kind of mapping between us and him or not.

Thanks for commenting!

Apologies for the late response — I’ve been so deep into my own research for the last month that I’ve had no chance to log on and read the blog.

I’ve been a very big fan of your paper on this topic ever since I saw you present it at ACAC last year. It’s really a great and highly insightful paper, and I’m sure it will have huge impact. I strongly recommend it to all of the blog readers. (I’ve also been saying for years — over Steve’s protests — that he’s the smartest person in the strategy field.)

So, here’s a few scattered comments and questions…

1.) Prior to reading this blog posting, I had simply thought of this paper as a general theory of inter-departmental screw-ups. I had not thought of it as a theory of how management creates value. That’s an interesting, and very ambitious, re-framing. Actually, I think it might be more accurate to say that it’s a theory about how management CAN create value, not about how management actually DOES create value. (After all, those Dilbert cartoons about the futility and counter-productivity of management still do ring true to most folks…) In other words, this paper delimits the upper limit of what management can hope to accomplish — the upper limit of potential value-creation by management. In this sense, it can provide an absolute benchmark by which to measure the performance of management, independent of all other inputs that might contribute to overall firm performance. That’s a very nice contribution, and very much needed. After all, it’s been 70+ years since the field of management began with Hawthorne and Taylor and Barnard and all that, and it’s a bit surprising that we’ve gotten along for all this time with no such theory of the potential value of management, and with no such benchmark for evaluating management’s contribution to firm performance. And perhaps it’s even more surprising that nobody has noticed that it’s been missing for all this time.

2.) That said, it is still (at least at this stage) ONLY a theory of the POTENTIAL value created by management, not the ACTUAL value created by management. Undoubtedly, this is a necessary first step toward the latter goal — it would probably be impossible to develop a theory of actual value creation without first knowing what the potential for value creation is. But it is still just a first step. (I guess this is perhaps a kinder, gentler way of making the same point that Pam made above in her comment #1.) In order to get from here to a theory of the actual value created by management, we would next need to postulate: (a) what concrete actions do managers actually take to prevent or alleviating these coordination problems and inter-departmental screw-ups, (b) what might prevent, impede, or discourage managers from taking those actions, and (c) if those actions are taken, what makes them effective or ineffective.

3.) Now, to say that this is a theory of the value created by management sounds awfully close to saying that this is a theory of the firm. After all, if there is a situation where management could potentially create value by preventing or alleviating the sort of coordination problems that you identify, then that value certainly is not going to be created without a firm — or, to be more precise, without the two parties submitting themselves to the authority of some third party (this is the point of Joe Mahone’s 2001 Journal of Management article — what I’ve been calling the “Mahoney conjecture”). And conversely, nature abhors a vacuum: When there is such a potential for management to create value, there must also be a potential gain to trade (i.e., the potential for someone who fulfills the role of management to appropriate some portion of the value he/she would create), which someone should sooner or later step in to exploit. So, is this also a theory of the firm?

4.) If this is a theory of the firm, then that doesn’t necessarily mean that it’s a completely new and independent theory of the firm. So, is it? Or is it merely old wine in new bottles — i.e., new labels applied to ideas that we’ve seen before? For example, trans-specialist understanding (TSU) sounds to me like it might be a relationship-specific asset. Does this theory “map onto” previous theories of the firm in some way? See Bob Gibbons’s recent paper “Four Formal(izable) Theories of the Firm” (JEBO, 2005) for an inventory of possibilities to map onto.

Anyway, kudos again on a great paper. I look forward to seeing it continue to develop and evolve. Please let me know if you think there’s anything I might be able to do in order to help that process along.

Cheers,
Rich Makadok

Well, the single biggest state intervention was probably the “internal improvements” of the 19th century (especially the railroad), that first led to industrial firms serving a national market. On this, I’m following Chandler. The process was continued by the government-subsidized civil aviation and highway systems of the 20th century. Subsidized transportation infrastructure makes distribution costs artificially low, and market areas artificially large. The same might be said of the government role in creating centralized communications infrastructure (the Bell patent system and AT&T, the global satellite telecom system, DARPA and the internet, etc.).

Patents, especially the pooling of patents, played a large role in the cartelization of industry.

Tax policy and direct subsidies both encourage more capital-intensive forms of production (the depreciation allowance, and subsidies to R&D and technical education), and the predominance of a more capital-intensive mode of production serves as a market entry barrier.

The state also subsidizes the operations by which mergers take place: the interest deduction for corporate debt, and the capital gains exemption of security transactions involved in mergers.

But transportation subsidies and IP are probably the two biggies. I suspect the economy would be decentralized beyond recognition if patents were abolished and transportation were funded with cost-based user fees.

Of course, all this involves a large helping of counter-factual speculation (or pulling it out of my hiney, in less charitable terms), so I can understand your skepticism.

In some cases, regulations act to make firms smaller, as when labor regulations are only imposed on firms with more than ten employees or family workers are exempted. My understanding is that many Italian firms are limited in size for precise this reason.

Firm sizes probably have a lot more to do with fully exploiting assets such as brand names, reputations, technologies, and capital stocks without having these assets excessively depreciated by users who are not under the control of the asset owner. If I own a brand name and try to rent it out to lots of small players, each one has an incentive to free-ride by not fulfilling the brand promise. I can get much better control over such behavior by only letting the brand be exploited by employees whose incentives I control and whom I can easily get rid of. (Franchises are interesting examples of getting around this, but not everything can be controlled adequately by franchise contracts.)

Capital equipment is another good example. I don’t see semiconductor fabs working real well if each worker and engineer were a free agent with a separate contract. And that’s ignoring the transactions costs of arranging such an array of contracts.

It seems to me that a lot of the management functions, regardless of how they’re “embodied,” presuppose a particular organizational form that should not by any means be accepted as simply given. Many of the management functions are ways of coping with agency problems that only exist in the first place because the dominant organizational size has grown far beyond optimum level–i.e., far beyond what it would be in a free market if all the inefficiency costs of large size were internalized, and if large size were not suited for exercising special privileges created by the state. In other words, many of these functions are a way of (to quote Drucker) “doing well what ought not be done at all.” As one of Ursula LeGuin’s characters said of a professional army, it’s the most efficient way of getting a lot of people to do something they have no rational interest in doing, given the substitution of extrinsic for intrinsic motivation.

The dominant organization has grown to the point that the cost of aggregating information from its component parts almost (if not altogether) exceeds the value of the benefit from aggregating the information.

In a free market, technical labor would probably be best performed by self-directed peer groups, and be integrated with other functions by market contract rather than hierarchy. It would resemble a lot of Tom Peters’ more extravagant rhetoric, with the difference that it wouldn’t take place as Peters imagines it as a simulated market within a corporate framework of finance, branding and IP. Rather, it would look a lot more like the P2P capitalism of Eric Raymond (or the P2P libertarian socialism of Michel Bauwens, for that matter).

As far as the knowledge selection problem goes, I have a 2002 paper in Organization Science where I discuss the role of management in avoiding glitches, which includes selecting people with the proper degree of knowledge specialization and trans-specialist understanding. A key issue is that trans-specialist understanding is very expensive, so you want to be sure to invest in it only when necessary (which the paper tries to identify). On the other impediments you discuss, motivation would fall under cooperation problems (private motivation differs from what is optimal for the firm); different time frames and risk preferences, if privately motivated would be cooperation problems; and if due to ambiguity in organization-level marching orders would be goal ambiguity problems.

1) Lack of staff engagement or motivation seems like a possible impediment. Let’s say you already have impediments of a given magnitude; that people have a long list of reasons for not communicating and coordinating better, and they are correct because the organization and its processes are in fact unhelpful to this key task. Management can take action to hunt down the actual impediments and try to remedy them directly. Or not. It might instead focus on getting people to make more effort to coordinate better on their own. They might try to do this by taking various actions meant to increase the level of “engagement” that people feel in their work, and to encourage them to devote their discretionary efforts more heavily in favor of the company. Pay could even be involved. If people are more strongly engaged, their energy and even altruism may simply overcome glitches or goal incongruences or other impediemnts to a greater extent. So, a poor level of employee engagement could be a kind of impediment, perhaps one that is distinct from the list you provided.

2) Regarding goal congruence, people might be incongruent not only in the goals themselves but in their time frames and risk tolerances. Many initiatives require time and commitment, for uncertain results. If you are someone who discounts heavily for time or risk, you’ll discount the payoffs from a given initiative. Maybe this is just another dimension of goal incongruence.

3) Some people are better at communicating and overcoming glitches than others, particularly within technical functions. Some people are more likely than others to make a sow’s ear into a silk purse. If you have a pack of poor communicators in each of the departments between whom coordination is very important, you’re worse off. Again without really reducing organizational impediments, a company might get better results by being better at identifying those technical functions in which coordination and communication are pivotal and then hiring appropriately. So, “wrong people” or “poor selection” could be an impediment distinct from the other classes of glitch.

Overall, I think the fact that people are just differently endowed and differently inclined may be a general place to look for impediment or advantage, as a potentially separate matter from the list of impediments you cited. In any event, It seems like it would affect the chances that trans-specialist understand would go missing.

It continues to strike me that these matters (like trying to reduce impediments and increase goal congruence) are one among about a zillion examples of things within the ordinary purview of management authority. I’m not really following why this subject matter rises to the level of calling into question what it is that management does. On the contrary, it seems like their job descriptions just put them in a position where they can take concrete actions to enable a certain fluidity in communication, a congruence of efforts, adaptability, all kinds of stuff. Or they can mess it up. Being the “deciders,” their roles seem definitionally decisive.

“…we don’t see some firms with too much integration…”

You probably haven’t properly defined over-integrated then. Integration has to do with the flow of information. A firm with “too much integration” would probably be one of two types.

First, where management took too long to make decisions because there were too many factors to be considered, and so on. Analysis-paralysis. Or where nothing could ever change or improve in one area because it would negatively impact a nearby area.

Second, where the level of innovation was low because the level of cross-training was too high. If everyone has to know everything about the methods and thinking of all the other departments, then the speed of adaptation slows to a crawl. Lots of European firms like this.

Notice that both of the above are management failures, and the first one could actually look like its opposite – no integration and no communication. Very much like the left-right scale where far left and far right are nearly indistinguishable.