Spero Consulting

Pref­ace: A Brief Review of Decentralization

An orga­ni­za­tion is a group of indi­vid­u­als with a com­mon purpose.

A decen­tral­ized orga­ni­za­tion struc­ture is one in which sub­or­di­nates are per­mit­ted a degree of auton­omy to decide or act upon cer­tain mat­ters as rep­re­sen­ta­tives of the organization.

Strate­gic and man­age­r­ial con­trol involve the deter­mi­na­tion of the opti­mal level of decen­tral­iza­tion (or del­e­ga­tion) and choices of related poli­cies and mech­a­nisms like bud­get processes, per­for­mance mea­sures, and reward schemes.

By “del­e­ga­tion” and “auton­omy” we do not mean the absence of con­trols. Instead, we mean the absence of close or con­stant mon­i­tor­ing or super­vi­sion and the pres­ence of devices, like incen­tive pay, that influ­ence behav­ior. Our empha­sis is on the fact that the choices of man­age­r­ial con­trols induce cer­tain behav­iors — both desired and unde­sired — and like Frank Sina­tra sang in a slightly dif­fer­ent con­text, “…you can’t have one with­out the other.”

Incen­tive Problems

With­out close or con­stant mon­i­tor­ing (or their tech­no­log­i­cal equiv­a­lents), when­ever an orga­ni­za­tion per­mits auton­omy, there is a poten­tial for incen­tive prob­lems — the pos­si­bil­ity that sub­or­di­nates will pre­fer out­comes (and there­fore make deci­sions or take actions) that are incon­sis­tent with organization’s goal.

These prob­lems arise because of infor­ma­tion asym­me­tries, which are often cat­e­go­rized as either (1) pri­vate infor­ma­tion prob­lems or (2) hid­den effort prob­lems. ¹ In the for­mer, the pri­vate infor­ma­tion could be about the inter­nal or exter­nal envi­ron­ment or about the per­son him­self, i.e., his type. Often it is nec­es­sary to elicit the pri­vate infor­ma­tion to solve coör­di­na­tion prob­lems among mul­ti­ple indi­vid­u­als or depart­ments. In the lat­ter, the hid­den effort usu­ally involves some task that the orga­ni­za­tion chooses not to observe — usu­ally because of the cost of mon­i­tor­ing — and this leads to moti­va­tion problems.

The organization’s losses due to these prob­lems are know as agency costs because the sub­or­di­nate has the author­ity to act as an agent on behalf of the orga­ni­za­tion. Note that on finan­cial state­ments agency costs are not cat­e­go­rized as such; instead, they appear as addi­tional com­pen­sa­tion expense or reduced out­put and rev­enue or higher pro­duc­tion or invest­ment costs. Our focus in this essay in on the fact that agency costs are nat­ural and pre­dictable byprod­ucts of delegation.

Clearly, no orga­ni­za­tion would per­mit some­one to act on its behalf if the costs of such del­e­ga­tion were greater than the ben­e­fits. So, when we observe sub­or­di­nates with author­ity, we can rea­son­ably infer that the ben­e­fits of del­e­ga­tion exceed the costs, includ­ing the agency costs. ²

In cer­tain envi­ron­ments, there are many good rea­sons to pro­vide a sub­or­di­nate with a high level of auton­omy. These include the cost sav­ings of exten­sive super­vi­sion and mon­i­tor­ing; the oppor­tu­nity for the orga­ni­za­tion to ben­e­fit from faster and more informed deci­sions (using the subordinate’s pri­vate infor­ma­tion and unique exper­tise), and the improved job sat­is­fac­tion for cer­tain types of workers.

In most real set­tings of inter­est, sub­or­di­nates have both pri­vate infor­ma­tion (about them­selves or the envi­ron­ment) and take effort that supe­ri­ors can­not observe. A rea­son­able man­ager — a thought­ful and empa­thetic one — has the fore­sight to pre­dict the unde­sir­able actions that the organization’s struc­ture and poli­cies may induce and there­fore be able to esti­mate the costs asso­ci­ated with them. ³

Thus, if auton­omy is opti­mal, then there is an eco­nomic level of dys­func­tional behav­ior (with an asso­ci­ated level of agency costs). This opti­mal level may vary as envi­ron­men­tal fac­tors vary. First, if the organization’s strat­egy or struc­ture changes as a reac­tion to changes in envi­ron­men­tal fac­tors, the level and cost of dys­func­tional behav­ior may change. Sec­ond, man­age­r­ial or tech­ni­cal inno­va­tions may change dys­func­tional behav­ior or the asso­ci­ated costs.

Regard­less, with all things equal, the orga­ni­za­tion desires the min­i­mum pos­si­ble level of dys­func­tional behav­ior and agency costs with­out los­ing the ben­e­fits of auton­omy. ⁴

A Slight Detour

In a for­mer life as a pro­fes­sor, we lec­tured on deci­sion mak­ing within firms, which we define as orga­ni­za­tions with profit motives. We would use math­e­mat­i­cal prob­lems as mod­els of the firms, and dis­cuss the gen­eral firm prob­lem of profit max­i­miza­tion and the pro­duc­tion prob­lem (or spe­cific action plan) where the firm’s goal was to min­i­mize costs given a cer­tain desired level of activ­ity or pro­duc­tion, e.g., quan­tity. In fact, we would even force our advanced MBA stu­dents to use Math­cad to solve such problems.

We would ram­ble on about var­i­ous assump­tions and char­ac­ter­is­tics that were needed to solve the math prob­lems and at the end the dis­cus­sion would ask the fol­low­ing question:

Below are two state­ments. Are both, either, or nei­ther true?

• Profit max­i­miza­tion implies cost minimization.
• Cost min­i­miza­tion implies profit maximization.

While it may be unfair to pose these same ques­tions to the reader with­out the six hours of lec­ture, the Pow­er­Point slides, the graphs, and the math, we ask any­way. What does the reader think? Yes, it is a rather vague set­ting, but please give it a quick try. ⁵

Profit max­i­miza­tion does imply cost min­i­miza­tion. If the firm’s plan of action max­i­mizes profit then it must also min­i­mize costs (given that plan). We will prove it by con­tra­dic­tion, which is a very typ­i­cal method used in eco­nom­ics. So, we will assume that the state­ment is false, and then show that it can­not be false.

Sup­pose the firm could find a cheaper way to act or imple­ment the same plan. In that case, the plan wasn’t profit-​maximizing as we had sup­posed, and that is a con­tra­dic­tion. So, unless the envi­ron­ment changes — say, prices, costs, or tech­nolo­gies — if one is max­i­miz­ing profit, there is not a cheaper way of pro­duc­ing the planned out­put (or pro­vid­ing the planned ser­vice level).

The sec­ond state­ment is false. Cost min­i­miza­tion does not imply profit max­i­miza­tion. The fact that total costs are min­i­mized via a cer­tain plan of action says noth­ing about the ben­e­fits derived from that plan or about the net dif­fer­ence between the ben­e­fits and costs, i.e., the profit. Thus, min­i­miz­ing costs does not guar­an­tee profit max­i­miza­tion. In fact, in sim­ple, stark eco­nomic mod­els, with increas­ing con­cave rev­enue and con­vex cost func­tions, the global cost-​minimizing plan of action is usu­ally… inac­tion. To min­i­mize costs, the firm should not oper­ate and, there­fore, not con­sume any resources. Total cost is zero, and you can’t get lower than that. Unfor­tu­nately, and of course, unless you are a farmer, rev­enue is zero, too.

The Baby and the Bath Water…
Or the unin­tended con­se­quences of elim­i­nat­ing agency costs.

What’s our point? Except in extreme cases, the profit-​maximizing choice gen­er­ates a non-​zero, eco­nomic, level of costs that can­not be elim­i­nated with­out elim­i­nat­ing the ben­e­fits, too.

Now, why are we espe­cially long-​winded about this issue? Partly because it is our nature but mainly to empha­size the point that in decen­tral­ized orga­ni­za­tions, many man­agers ignore this maxim and try to elim­i­nate all agency costs when, in fact, agency costs are nor­mal costs of doing busi­ness in a decen­tral­ized man­ner. That is, these man­agers lack the dis­ci­pline and willpower to imple­ment truly decen­tral­ized controls.

The word “willpower” in the pre­ced­ing sen­tence was pur­pose­fully cho­sen because per­mit­ting employ­ees dis­cre­tion and auton­omy often requires the same level of per­sonal dis­ci­pline as main­tain­ing a diet or for­sak­ing other temp­ta­tions. In fact, due to our own per­sonal weak­ness, we can’t avoid the temp­ta­tion of another aside on Strate­gic Con­sis­tency and Man­age­r­ial Dis­ci­pline, but we will hide it over here.

So, what can a poor man­ager do? Eco­nom­ics pro­vides a lit­tle guidance.

The Rev­e­la­tion Principle

In eco­nom­ics the Rev­e­la­tion Prin­ci­ple states that in cer­tain envi­ron­ments — under cer­tain assump­tions — one can learn the truth from oth­ers. Per our side essay on man­age­r­ial dis­ci­pline, this usu­ally requires some type of com­mit­ment by the supe­rior to elicit the pri­vate infor­ma­tion. That means that while one can learn the truth, one can­not learn the truth for free. (Obvi­ously, if one could learn the truth for free, there wouldn’t have been a prob­lem in the first place.)

Unfor­tu­nately, in most real cor­po­rate settings, all the con­di­tions needed to get truth­ful report­ing are rarely present, and even if one could get it, it isn’t nec­es­sar­ily worthwhile. So, the supe­rior must trade-​off the value of the infor­ma­tion ver­sus the costs of its acqui­si­tion, where the major cost is the infor­ma­tional rent that the sub­or­di­nate demands (for know­ing some­thing the supe­rior doesn’t). The cost may be a pay­ment sim­i­lar to a brib or it may involve the oppor­tu­nity cost of allow­ing the worker to shirk or con­sume perquisites in exchange for the information.

A Cou­ple of Non-​business Examples

Pros­e­cu­to­r­ial Offers of Immu­nity – con­sider crim­i­nal cases — either in real life or on tele­vi­sion — where a dis­trict attor­ney or other pros­e­cu­tor pro­vides an indicted accom­plice with immu­nity in exchange for his truth­ful tes­ti­mony against a co-​conspirator.

The ben­e­fit is that the pros­e­cu­tor increases the prob­a­bil­ity of con­vict­ing the partner-​in-​crime.

What are the costs? First, a (pre­sum­ably) guilty per­son goes free and is not penal­ized for his or her crimes. Sec­ondly, at the mar­gin, morally-​challenged indi­vid­u­als are more will­ing to join con­spir­a­cies know­ing that if they are caught their penal­ties can be reduced by coop­er­at­ing with author­i­ties. This is equiv­a­lent to shirk­ing effort within a firm. (Of course, oper­a­tors of crim­i­nal orga­ni­za­tions know this, too, and attempt to develop their own con­trol mech­a­nisms that gen­er­ally promise ret­ri­bu­tion for squeal­ing, e.g., omertá or code of silence.)

Teenage Drink­ing – To reduce the prob­a­bil­ity that their teenager is in an alcohol-​related auto­mo­bile crash, some par­ents adopt the pol­icy of no ret­ri­bu­tion for hon­est rev­e­la­tion. In other words, the par­ents instruct their teen to call home if the child or his or her dri­ver has been drink­ing or is intox­i­cated. The par­ents com­mit not to pun­ish the child despite the fact that the behav­ior is illegal.

On one hand, con­di­tional upon the fact that the teen has been drink­ing, this pol­icy may reduce the pos­si­bly severe, adverse con­se­quences of dri­ving while impaired or rid­ing with an impaired driver.

On the other hand, com­mit­ting to elim­i­nate the pun­ish­ment for such anti-​social behav­ior also sig­nals to the child the parent’s (at least) tacit approval. Like the pros­e­cu­tor who real­izes that mak­ing deals increases the like­li­hood of crime, par­ents should expect an increased like­li­hood of their child con­sum­ing alco­hol. Unin­tended con­se­quences arise when the child is too impaired to think clearly enough to call or the impaired state leads to other behav­ior and neg­a­tive con­se­quences; take your pick with the ways that drunken teenagers can get into trou­ble. ⁶

In our exam­ples, the infor­ma­tion seeker must be able to cred­i­bly com­mit to his or her pol­icy. If the dis­trict attor­ney reneges on his pledge after the fact and attempts to pros­e­cute the infor­mant, then he for­sakes the use of the tool for the dura­tion of his term. In most locales, there aren’t that many defense attor­neys, and that type of infor­ma­tion trav­els quickly among them.

If a drunken child tele­phones his par­ents and is then pun­ished, the par­ents should not expect the call the next time the teenager drinks. (Of course, the child may be less likely to drink because of the expected punishment.)

These exam­ples pro­vide evi­dence that there are no sim­ple solu­tions. Such prob­lems are rarely elim­i­nated (with­out the often pro­hib­i­tive cost of full cen­tral­iza­tion). At best, they may be mit­i­gated. So, we are left with the basic question.

Does one seek infor­ma­tion or wish to motivate?

The opti­mal choice of whether to elicit infor­ma­tion or moti­vate effort (and to what degrees) depends upon the var­i­ous costs and ben­e­fits of each alter­na­tive, includ­ing long-​term impli­ca­tions, and these costs and ben­e­fits are deter­mined by cer­tain char­ac­ter­is­tics of the organization’s envi­ron­ment and strategy.

It is worth not­ing that it cer­tain set­tings, a sin­gle piece of infor­ma­tion — a sin­gle rel­e­vant fact — may be worth mul­ti­ples of a subordinate’s annual salary, and the per­son may get a bonus for seem­ingly “doing nothing.”

So what’s the solu­tion? We rec­om­mend hir­ing a con­sul­tant and have a par­tic­u­lar one in mind: some­one who real­izes that there is no sin­gle solu­tion for all set­tings and that all decen­tral­ized solu­tions to the con­trol prob­lem will induce some eco­nomic level of agency costs with which one must learn to live. Costs that mature, thought­ful, and dis­ci­plined man­ager must per­mit for the cho­sen solu­tion to the con­trol prob­lem to remain optimal.

Epi­logue

In the spirit of char­ity and good­will towards the reader, we present char­ac­ter­i­za­tions of two types of prob­lem man­agers in the decen­tral­ized (or par­tially decen­tral­ized) struc­tures. Based upon our obser­va­tions, we will label them “The Ide­al­ist” and “The Extrem­ist.” ⁷

These two types dif­fer by inten­sity but also by tem­pera­ment; the Ide­al­ist is the kinder, gen­tler ver­sion and gen­er­ally has a sin­cere demand for the infor­ma­tion and effort. The Extrem­ist is often a mis­an­thrope but tries to dis­guise by con­sid­er­ing him­self a hard-​charging, high-​intensity man­ager. There are ener­getic, decent folks try­ing to do the right thing. We are not describ­ing those good people.

The Ide­al­ist: Wouldn’t It Be Nice if…

We often observe supe­ri­ors who wish they pos­sessed their sub­or­di­nates’ pri­vate infor­ma­tion or could observe their sub­or­di­nates’ hid­den efforts. Often such sen­ti­ments are not expressed as these pos­i­tive desires; instead, they are com­mu­ni­cated as com­plaints like “I know he is lying to me;” “He says he doesn’t know, but that’s not pos­si­ble;” or “He is so lazy.” (Is he lazy or is he behav­ing opti­mally for him­self and the orga­ni­za­tion? What could he know that you don’t?)

Usu­ally, the ide­al­ist is attempt­ing to man­age a decen­tral­ized struc­ture in a cen­tral­ized way, and spends too lit­tle time think­ing about the design and effects of con­trols and too much time focus­ing on their sub­or­di­nates’ observed behav­iors, which any rea­son­able per­son may admit might be annoying.

In extreme cases, we have seen ide­al­ists attempt to elim­i­nate all dys­func­tional behav­ior. Such “man­age­ment” is almost always self-​defeating.⁸

Now the clever ide­al­ists may ask rhetor­i­cally, “What are you writ­ing about? There is no prob­lem. Near the begin­ning of the essay (and else­where on the web site), you, your­self, men­tion that the opti­mal level/​degree of auton­omy will change as envi­ron­men­tal fac­tors changes. When we fer­ret out the infor­ma­tion and direct work­ers regard­ing spe­cific tasks, we are respond­ing to changes in the envi­ron­men­tal and (implic­itly) chang­ing the orga­ni­za­tional structure.”

If such direct man­age­ment results from a con­scious deci­sion to cen­tral­ize, so be it. How­ever, attempt­ing to elim­i­nate all dys­func­tional behav­ior is de facto mon­i­tor­ing, and it elim­i­nates the ben­e­fits of del­e­ga­tion and auton­omy. More­over, if the struc­ture is decen­tral­ized and you — as a supe­rior — are attempt­ing to mon­i­tor rather the design effi­cient mech­a­nisms and poli­cies, then you may be the incen­tive prob­lem — not your sub­or­di­nates, e.g, that knee-​jerk reac­tion to pun­ish oppor­tunis­tic behav­ior — though sat­is­fy­ing — need not be opti­mal, and that makes the Ide­al­ist the incen­tive problem.

Ide­al­ists need to be trained about the nature of their man­age­r­ial respon­si­bil­i­ties, which extend beyond supervision.

The Extrem­ist: No pain, no evi­dence of management.

Before describ­ing the Extrem­ist we note that in dynamic envi­ron­ments, with­out incen­tive prob­lems, the goal to oper­ate at 100% capac­ity is almost never opti­mal. The basic idea is that a lit­tle bit of slack allows one to take advan­tage of prof­itable (and pos­si­bly ephemeral) oppor­tu­ni­ties as they ran­domly arise. This isn’t always true, but in the absence of addi­tional facts, it is a rea­son­able work­ing premise.

Now, Extrem­ists are often unaware of this or will­fully ignore it; sub­or­di­nates will just have to worker harder or longer or both. Hey, they should be happy that they have jobs. They don’t deserve them, after all. Extrem­ists push for max­i­mum effort, regard­less of the costs (or ben­e­fits) to the sub­or­di­nates or the organization.

Unfor­tu­nately, Extrem­ists never iden­tify them­selves as such. Worse, yet, is the fact that their supe­ri­ors may not iden­tify them either. Instead Extrem­ists tend to pool them­selves — see adverse selec­tion — with earnest folk and refer to them­selves are hard-​charging or cost-​conscious or efficiency-​minded, but their moti­va­tion is rarely pos­i­tive. They are often busy­bod­ies and prospec­tive empire builders, and for them schaden­freude is a way of life.

We could go on, but there is no need. The chance that such a per­son will man­age effec­tively in a decen­tral­ized orga­ni­za­tion is about zero. The chance of redeem­ing them through train­ing is very close to zero, as well. They are incen­tives prob­lems and will cre­ate addi­tional prob­lems for their orga­ni­za­tions (while try­ing to blame oth­ers for any fail­ures). Unfor­tu­nately, for orga­ni­za­tions lack­ing strate­gic con­sis­tency the Extrem­ists’ absence of man­age­r­ial dis­ci­pline may go unno­ticed as they may pro­duce short-​term gains at the cur­rent expense of their sub­or­di­nates and the future cost of the organization’s abil­ity to func­tion effec­tively and efficiently.

Copy­right © 2008 Spero Consulting

---

Foot­notes: