Wednesday, October 31, 2007

Measuring Failure

Instinctively -- when you encounter one -- you know that you are looking at a bloated bureaucracy; the real question, is how to prove it?

We obviously need some simple form of cost-based accounting, that gives us a decent metric on which to compare various organizations and their effectiveness. The key here is the the measure be simple and that it cannot be subverted by those trying to apply it. Clearly, it is in a bureaucracies best interest to hide its own inner ineffectiveness; that is the very essence of why the organization is defective in the first place.

To keep it simple we only need to deal with two basic observations. The first is that at the end of the day all organizations produce either products or services, nothing else. Regulatory monitoring for instance, is just a service for the public to insure that the rules are followed. Building roads is a product, while maintaining them is a service. Everything falls into one or the other category.

The second observation is that all organizations have a cash intake. Internally they can play around with hiding pools of money from previous periods to 'alter' the appearance of getting more or less money, but in an absolute way these are just various ways of applying a debt financing service to the structure. Each period they collect a fixed amount of money, which at the end of the period is entirely known.

So applying our understanding, we can easily create a product/service accounting model that fits over each and every organization. We do this by setting forth a limited number of categories for the product/services. The less, the better. The creation of new categories needs to be tightly controlled, as it is the place most likely to be subverted by those who are reporting. So there is a very limited number of these 'official' categories.

Each category was a name -- such as "road building" -- and a metric -- such as kilometers. For each category, we find out exactly how many kilometers were built for that organization, for that period.

Say for example, a provincial government created 10,000 kms of new roads over the course of a year. There is nothing difficult in gathering that number, if there were, then it would show a serious problem with the organization's internal structure that needs to be immediately fixed; if your paying people to build roads, but you don't know how many they have built, your in serious trouble.

To each category a portion of the incoming cash flow is assigned.

That may be a little more difficult, but generally you can figure it out via capital expenditures and various budgets. Again, the powers that be will try to subvert this, so the rules about subdividing up the cash into the various product/service 'buckets' needs to be formal and tightly controlled. You can't for instance shift some of the cash onto other services, if the underlying budgets don't confirm the percentages.

There are of course two main difficulties with these ideas.

The first is administrative overhead, while the second is variable pools of money. In the first problem, most organizations will attempt to shift the various portions of their overhead around to make them look more or less effective. To counter thing, the allocation of monies to various categories should not vary significantly from one period to the next. Where expenditures are absolutely provable, they cannot be shifted to other categories. Both of these rules will have to be able to track the effectiveness of changes over an extended periods.

Organization budgeting is often extremely complex. The idea here is not to get 'tricked' but that complexity. If an organization sets aside money for later, that in itself is a service. If an organization borrows money to fund a project, that too is part of a service. Then for example is we get $100, and borrow $20 the incoming cash is $120, which is spend on whatever original service we were performing and also a debt service that is costing us a rate such as 10% per year, put whatever costs were incurred to arrange the debt.

The key idea in this accounting approach is that all of the money is allocated to a limited fix set of products or services. There should be nothing left over that isn't allocated. It all has to go somewhere. The mechanisms for allocation should be simple, and they should consistent. If we get a result that one government is spending $200 per km to pave roads and another is spending $400 dollars, the inference that the second organization is twice as inefficient should be correct; there might be differences in 'conditions' or the 'environment', but the underlying relationship is truthful.

Once we have these metrics in place, finding and fixing the problems in our organizations will be easier. Bureaucracies will no longer have the option of putting 'spin' on their results. They will no longer be able to bury their mistakes, errors or corruption. They will be open and visible. Instead of watching as these organizations steadily decline, we'll be able to assess their health and whether or not specific changes are improving or making it worse. We'll be able to compare two organizations with similar services and see which one is operating more effectively.

We will finally be able to fix the unfixable.

No comments:

Post a Comment