The asset that can leave, and what keeping it costs

Picture a quarter in which three initiatives are live at once. An attribution project, to finally see who contributes what. A retention push, because two seniors resigned in the spring. And a quiet argument about capping what the best seller earns. Each is sensible on its own page. Together they are a machine working against itself: because the asset all three touch is the one kind that can resign, and its risks are connected in ways none of the three programmes admits.

This site has argued that growth compounds when work accumulates into stock, and that every budget quietly pays to maintain that stock. Under both essays sits a premise neither examined: much of the stock is held in people. Russell Coff put the consequence plainly in 1997, in an Academy of Management Review paper written for theory rather than practice: “The most obvious problem is that the firm’s assets walk out the door each day, leaving some question about whether they will return.” Firms cannot own employees. Kidder Peabody, his running example, was devastated when its key brokers moved to competitors.

Which three organizational mechanisms cause human-capital competitive advantage to fail?

Coff’s framework names the failure modes, and it is worth having all three on the table before reaching for any cure:

1. It leaves. Where a person’s value is a client network, some of the network goes with them: “transferable (in varying degrees) when employees move to competing firms,” as the paper puts it. How large this risk really is gets argued in both directions; more on that below.
2. It cannot be verified. Hiring runs on claims no one can check: applicants “may misrepresent themselves by taking credit for the success of their former employers,” and “Causal ambiguity will thwart efforts to verify such claims.” The problem continues after the hire: moral hazard, in Coff’s terms, is what becomes possible “when individual contributions are difficult to observe.” One of his interviewed firms sometimes kept mediocre editors for extended periods, because it took up to three years to know whether a replacement was any better.
3. It captures its value in pay. The oldest fear on the list. The economist Henry Simons put it on record in 1944: “I can see no reason why strongly organized workers, in an industry where huge investment is already sunk in highly durable assets, should ever permit a return on investment sufficient to attract new capital”: quoted here as Card, Devicienti and Maida quote it, because their 2014 study in the Review of Economic Studies set out to test exactly this.

That third fear deserves one careful paragraph, because it is the only one of the three with a famous test attached, and the test does not reach the reader of this essay. Card, Devicienti and Maida matched social-security earnings from Italy’s Veneto region to firm financials and found rent-sharing is real: wages rise with employer profitability, with an elasticity around 4–5% within jobs. But the reassuring half of their result: “little evidence that bargaining lowers the return on investment”: is measured on collectively bargained non-managers, with managers excluded from the sample outright, and the protected return is the cost of booked capital. A client network built out of expensed sales effort is not booked capital, and a star with the book is not a bargaining unit. For the asset this essay is about, the pay fear is neither confirmed nor refuted, and I know of no test that reaches it. That is a different thing from settled, and no ranking of the three risks follows from it.

Why do retention mechanisms impose hidden operational costs beyond compensation?

What makes this a system rather than a list is what the cures cost. The couplings are Coff’s own, scattered through the paper: a footnote, a definition, a late-page conjecture. The assembly is mine.

Cure the verification problem and you feed the other two. Make individual contribution measurable: commissions, rankings, attribution dashboards, and you have produced the market’s pricing sheet for your own team. Coff’s footnote states it exactly: “if individual productivity is observable … the problem may be reversed and the highest performers may be at risk for turnover. This was the case for the Kidder Peabody brokers who were lured away to competitors.” One commercial setting where individual performance is public: star securities analysts, ranked and published: is the one Boris Groysberg, Ashish Nanda and Nitin Nohria reported on in Harvard Business Review: after a star moves, performance tends to fall, the receiving group suffers, and the hiring firm’s market value tends to drop. Legibility invites the bid; the bid, on their reading, overpays. Firms sense this trade, which is why Coff conjectures that some “may actually underinvest in information production to prevent competitors from gaining access to it.”

Cure the leaving problem by contract and the law prices it in cash. In Germany the price is statutory: a post-contractual non-compete binds only if the employer pays, for every year of restraint, “mindestens die Hälfte der von dem Handlungsgehilfen zuletzt bezogenen vertragsmäßigen Leistungen”: at least half of everything the contract last paid them, my rendering: per § 74 Abs. 2 HGB, extended to all employees by § 110 GewO. That is German law, not Austrian and not Swiss. But the principle it prices is general: restraining a departed person is an expense, not a retention.

Cure the leaving problem with money and you have renamed the third risk. Coff is definitionally strict here, and the strictness is useful: retention strategies are “policies that promote retention without allocating the rent. In other words, one simple solution is to pay people enough so they will not quit … Although this promotes retention, it also allocates rent and is therefore considered a rent-sharing strategy.” Paying to stay is not a cure for value capture. It is value capture, arranged voluntarily.

Cure the leaving problem without money, and the instruments are firm-specific by construction. Relationships with specific coworkers, a working environment a competitor cannot copy, participation in real decisions: Coff’s point is that these retain precisely because “other firms cannot duplicate” them, and he reads the industry’s signing bonuses as the market pricing that loss: they “might be viewed as an attempt to reimburse brokers for the loss in firm-specific compensation.” The catch is stated in the same paper, with his hedge intact: competitors “might be able to” imitate these strategies, and if so, they “would shift the focus from a retention problem to an acquisition problem.”

Which three organizational structures manage relational dependency and risk?

Coff illustrated his framework with four firms, which he offered explicitly as illustrations, “not presented to test hypotheses.” Three of them, read side by side, show what a deliberate allocation looks like. Treat the numbers as what they are: managers describing their own firms.

A securities brokerage holds the pay risk on purpose. The asset sits in the brokers and is fully legible: commissions measure everyone. Its high-producing brokers “can bring 95% of their business with them,” yet the firm reported turnover under 10%, held by what money cannot copy: support services, coworkers, influence over real decisions. The design accepts legibility, shares rent deliberately through commissions, and retains through firm-specific ties.

A training company moved the asset out of the people. It “consciously develops proprietary training materials so that presenters cannot leave the firm and compete directly against them. As a result, when presenters exit, they typically do not continue conducting seminars.” The person leaves; the seminar stays. The bill is content production, and the knowledge that a rival could build materials of its own.

A consumer magazine holds the verification risk itself. It sometimes kept mediocre editors for extended periods, because replacing an editor meant up to three years of not knowing. Labor hoarding is not a failure to manage: it is a chosen exposure, taken because the alternative (churning through unverifiable candidates) was judged worse.

The location choice: knowledge in people or knowledge in systems: has a name in the management literature: Morten Hansen, Nitin Nohria and Thomas Tierney called the poles personalization and codification in Harvard Business Review in 1999, and treated which pole to occupy as a strategy question in its own right. The column their frame does not carry is the one above: which of the three risks each pole leaves you holding.

How should commercial leaders document and evaluate human capital dependencies?

The rule falls out of the asset itself. Value that is codifiable and repeated belongs in product and process: a method, a playbook, a template library, priced in the work of building them and the knowledge that rivals can too. Value that is inherently personal: client trust, individual craft: stays in people, and then the honest design questions are only two: how legible do we make individual contribution, knowing legibility is also the market’s pricing sheet; and which retention currency do we pay in: firm-specific ties that take years, or rent, shared knowingly. A commission-run sales team has already answered both questions, whether or not anyone answered them out loud.

Which is the point of the exercise below. For each asset the first essay’s stock-or-flow test surfaced, write down where it sits, which of the three exposures the current design is holding, and what you are paying on purpose. One caution belongs on the sheet itself: a completed copy names where your advantage sits and who holds it: valuable to a competitor, and to the person named. It is worth having for the same reason it is worth protecting. It stays internal.

How should executives evaluate counterarguments against formalizing tacit talent assets?

There is a serious objection: the first and third risks are one risk wearing two labels. What a person can leave with is their bargaining position, so “engineer where the asset sits” collapses into “decide what to pay”, and the wealth-management industry seems to prove it, holding client relationships through deferred compensation and contractual non-solicits, all instruments written in money.

Half of that objection stands, and the essay is better for conceding it: pay structure is the surface several of these cures are executed on. Commissions are what make the brokerage’s book legible; deferred compensation is how a flight risk gets converted into a payroll line; § 74 turns a restraint into a salary-shaped expense. But structure and level are different questions. The fear that dominates the discourse, that talent captures the returns: is about level. The configurations above differ in where the asset sits and how legible it is, which are choices no pay rise makes for you, and one of them keeps its asset without needing to keep the people: the training company’s presenters can leave; the seminars stay. Andris Zoltners and colleagues asked in Harvard Business Review in 2011 who owns your customer relationships, your salespeople or your company; the answer this essay adds is that ownership is built, it is built with more currencies than money, and several of those currencies, as the table above prices them, are another risk in different clothes.

Coff’s own summary is the right closing thought, and it survives translation out of theory: the assets and the systems that manage them: “Each element may be necessary, but neither by itself may be sufficient to bring about a sustainable advantage.” A commercial machine that compounds in people compounds on a foundation that can give notice. The teams that last are not defined by having closed all three exposures: the table above is the list of reasons why that is not for sale. They are the ones that could say, before the resignation letter arrived, which exposure they had chosen, and what they were already paying for the privilege. Which parts of a commercial asset are portable at all, and which were local all along, is the question under my doctoral research, and the reason this essay stops at the border of its evidence.

Where are the analytical boundaries of human capital asset theory?

Boundary. The literature identifies portability risk, not a universal probability that a customer will leave with a representative. Measure concentration, multi-threading and transition outcomes in the specific relationship base.

Evidence base. The analytical frame also draws on these additional sources: Card et al. 2014; Groysberg et al. 2004; Hansen et al. 1999; Zoltners et al. 2011. The links identify the exact works; they support the mechanisms and boundary conditions discussed here, not every claim in isolation.