A Quasi-Theoretical Case for Centralization

The general effect of the sorts of IT innovations that have been put into operation over the last couple of decades has been to increase the level of centralization in the organizations in which they’ve been installed. A particular effect is then a steadily increasing clustering of authority over resource allocation decisions in top-management functionaries. Now, if there is any reason for reservations about this trend, it would arise from the recognition that currently available technical instruments —both informational (IT) and analytical (DT)— cannot be relied on to support ‘rational’ allocationsexcept in relatively non-demanding decision situations…that is, those where predicates are dominantly “hard” or objective, where the focus is mainly on the short run and local, and where decision criteria are not ambiguous or incompatible (as, for example, with a health care system commanded to simultaneously serve both economy and quality of care). Because the allocation decisions that need to be made in the content of the typical business firm will be non-demanding, technical strictures will generally not be posed as a critical constraint on centralization. This is happy news, as there’s a good reason why we want to have commercial-industrial enterprises obtain relatively high levels of centralization. The reason derives from this proposition: Allocative rationality increases with altitude!

Again, what allocative rationality requires is the investment of some quantum of (presumably scarce) resources over an array of alternatives in such a way that aggregate impact of these investments is effectively maximized with respect to whatever outcome criterion/criteria is in force. Though there are some qualifications that eventually need mention, there's a benefit to considering more rather than fewer alternatives: Increasingly rich arrays of alternatives are increasingly likely to include the best possible uses to which any set of resources might be put. Here's where the assertion that "allocative rationality increases with altitude" comes into play. It draws its inspiration from what's come to be called the "systems approach". The signal feature of the systems approach is its emphasis on perspective. It urges expanding our view of things, considering factors and causes and effects at some remove, both temporally and spatially. The managerial variant on perspective is what can be referred to as breadth of executive purview. If then allocative rationality can be said to increase with altitude, it's because executives situated at higher levels of any managerial hierarchy will have a greater breadth of purview —a wider field of view, as it were— than those located at lower elevations. To the extent that increases in breadth of executive purview can be expected to translate into an awareness of a larger number and/or greater diversity of investment alternatives, we have the rationale for investing allocative authority in higher- vs. lower-order decision makers.

The AWACS Analogy

All other things considered equal (and this is a qualification that needs to be examined in some detail downstream), allocation decisions made by authorities operating at higher 'altitudes' should then entail a higher level of rationality than otherwise. The quasi-axiomatic character of these arguments is apparent in their conditional phrasing…all the "shoulds", etc. Nevertheless, there is some practical theoretical force behind the proposition that allocative rationality increases with altitude, as illustrated in the Figure 1 (below).

The set A consists of an array of alternative uses to which resources might be allocated. The entire set A is available for inspection by an executive sitting at the higher of the two elevations, 2. On the other hand, an allocative authority positioned at the lower altitude, 1, has only a subset of A within his purview, or{ai ak} [A]. Given the aforementioned rule, allocation exercises undertaken from elevation 2would be presumed to entail a higher potential for rationality than those executed at 1.

The empirical referent that perhaps most purely points out the pragmatic implications of this reasoning is the by now widely-familiar employment of AWACS-equipped aircraft in military engagements. An AWACS plane, popularly referred to as an "eye in the sky", is a radar platform sent aloft and positioned over an area where air-to-air combat is to take place. For any given angle of view, the higher the altitude at which the AWACS is operating, the more extensive the area it can oversee. This is analogous to the idea of increasing the breadth of executive purview, and the analogy extends also to the allocative advantages of doing so. The AWACS task is comprehensible as an exercise in dynamic or 'real time' resource allocation: To assign and vector friendly fighters to intercept/interdict an array of intruding enemy aircraft, which in the AWACS scenario play the role of alternatives. Positioning the AWACS plan at higher altitudes thus allows it to recognize more enemy aircraft and get an earlier fix on the relative magnitude of the threat they represent (in terms of the offensive weapons they carry, their speed and proximity to sensitive targets), than would otherwise be available. The effect of increases in altitude is thus more information on which to predicate allocative decisions, and hence the likelihood of making what turn out to be more favorable deployments of defensive assets by seeing that the prospectively most virulent threats are the first attended, properly matching the capabilities of friendly vs. adversary fighters, etc.

As another, more conventional case (involving static or planning-based vs. dynamic resource allocation), consider the allocative procedure that might be employed by a business firm to determine its next-period product mix: R x  x P  {R1(P1), R2(P2) … Rn(Pn)}, where:

R = the set of available resources (capital assets, factors of production, etc.)

 = a maximum-likelihood profit impact estimator

P = an array of alternative product-market possibilities

Ri(Pi)v = the resources committed for the production of the ith product at volume v.

If allocative decisions of this sort tend to be left for the highest-level executives in firms, it's because of the likelihood of their recognizing a broader range of product-market alternatives than those situated at less lofty levels. And to apply a variant on an argument introduced earlier, the greater the range of product possibilities included in P, the greater the likelihood of the firm's allocation schedule yielding the best possible return on invested assets. Take this reasoning to its next step and we get the rationale for the emergence of the supraexecutive…an example being the central allocative authority employed by shipping conferences.