Bureaucracies
Jan. 4th, 2006 08:51 amThere's a certain inevitable progression in bureaucracies. An evolution, to use the dreaded "E" word, in how they are created, age and die. The pace may vary, but the sequence is pretty much set in stone.
In logic terms, the organization of a bureaucracy forms a virtual neural net. At first, it is simple and direct, but if it survives the organization of a bureaucracy slowly grows more complicated. In the early stages this is a good thing; the complexity forms multiple redundant pathways, providing reserve capacity in case of trouble. If Joe is out sick, Carol can handle the same work. If there's a flood of extra work, Carol and Jill can help take up the slack.
However, as neural nets mature they become ossified, with preferred pathways forming. Any slight advantage of one particular pathway will result in it being used more and more often, while other pathways between the same origin and destination atrophy or are diverted to other tasks. Now, if Joe is out sick or there's extra work, Carol doesn't even know the proper procedures or any of the people working under Joe, and Jill retired three years ago and wasn't replaced. And when Joe retires - no matter how much warning there is of the event and no matter now much preparation has been made to replace him - the result is disruption of operation. (Figurehead managers being an exception, here. As long as there are assistants who actually do the work for titular heads such as political appointees, the bureaucracy will continue with business as usual. No matter how many promises of reform were made by the new administration.)
Worse, since there is now inherently less flexibility, any change in conditions will cause similar problems, because there are no resources in the neural net for handling anything different from what it has been doing for so long. Because the evolutionary process has eliminated the redundant and less-efficient (with "efficient" being defined by how well something matches the majority of the neural net) pathways and unused capacity. Unexpected changes are especially bad, but even expected changes can leave the bureaucracy floundering. Particularly hidebound bureaucracies may not even realize that the new situation requires a change in procedures, and claim everything is fine (which it is, in the operation of the bureaucracy, that being how it now measures success) when the result of the organization's work is actually poor, zero or even negative. (Think about FEMA's reaction to Katrina. The head of the organization was reporting that everything was under control, with no problems, when the agency was actually hindering rescue and relief efforts. And the director most likely didn't know this until he saw the problems on TV news, because he believed the bureaucracy's reports that it was operating normally.)
Adding to the problem, more and more of the resources of the organization become dedicated to supporting its operation, instead of helping actually do the work it was created for.
Reforms rarely help an old bureaucracy to change; due to the nature of mature neural nets, bureaucracies react to change - any change, even obvious improvements - as damage and attempt to repair it. Replacing a bad manager or reorganizing a section may result in temporary improvement, but eventually the rest of the organization will force the modifications to change back into something similar to what it is used to. What is needed is an organization-wide housecleaning.
In logic terms, the organization of a bureaucracy forms a virtual neural net. At first, it is simple and direct, but if it survives the organization of a bureaucracy slowly grows more complicated. In the early stages this is a good thing; the complexity forms multiple redundant pathways, providing reserve capacity in case of trouble. If Joe is out sick, Carol can handle the same work. If there's a flood of extra work, Carol and Jill can help take up the slack.
However, as neural nets mature they become ossified, with preferred pathways forming. Any slight advantage of one particular pathway will result in it being used more and more often, while other pathways between the same origin and destination atrophy or are diverted to other tasks. Now, if Joe is out sick or there's extra work, Carol doesn't even know the proper procedures or any of the people working under Joe, and Jill retired three years ago and wasn't replaced. And when Joe retires - no matter how much warning there is of the event and no matter now much preparation has been made to replace him - the result is disruption of operation. (Figurehead managers being an exception, here. As long as there are assistants who actually do the work for titular heads such as political appointees, the bureaucracy will continue with business as usual. No matter how many promises of reform were made by the new administration.)
Worse, since there is now inherently less flexibility, any change in conditions will cause similar problems, because there are no resources in the neural net for handling anything different from what it has been doing for so long. Because the evolutionary process has eliminated the redundant and less-efficient (with "efficient" being defined by how well something matches the majority of the neural net) pathways and unused capacity. Unexpected changes are especially bad, but even expected changes can leave the bureaucracy floundering. Particularly hidebound bureaucracies may not even realize that the new situation requires a change in procedures, and claim everything is fine (which it is, in the operation of the bureaucracy, that being how it now measures success) when the result of the organization's work is actually poor, zero or even negative. (Think about FEMA's reaction to Katrina. The head of the organization was reporting that everything was under control, with no problems, when the agency was actually hindering rescue and relief efforts. And the director most likely didn't know this until he saw the problems on TV news, because he believed the bureaucracy's reports that it was operating normally.)
Adding to the problem, more and more of the resources of the organization become dedicated to supporting its operation, instead of helping actually do the work it was created for.
Reforms rarely help an old bureaucracy to change; due to the nature of mature neural nets, bureaucracies react to change - any change, even obvious improvements - as damage and attempt to repair it. Replacing a bad manager or reorganizing a section may result in temporary improvement, but eventually the rest of the organization will force the modifications to change back into something similar to what it is used to. What is needed is an organization-wide housecleaning.