I HAVE long held a vision of what a good society would include. Its most important elements are liberty, environmental quality and communities of people who value both. But only certain institutional arrangements foster these ends. History is unlikely to show that yesterday's election moved us toward these goals.

But elections do cause thoughtful people to carefully separate their hopes from their expectations. We hope that people will weigh the public interest more heavily than their narrow self interest. However, elections usually teach us to expect concentrated interests to prevail and government to grow.

Organized interests solidify their gains by creating institutions and agencies. With these in place, even modest reforms become difficult to enact. We see this in the protracted efforts to increase the price of grazing federal lands. For grazing on our family's Montana ranch we charge more than $10 per cow and calf per month. On federal land near ours, permit holders such as Metropolitan Life Insurance pay $1.85 for their Forest Service and BLM grass.

This is a predictable consequence of political allocation. We would surely be far better off if the grazing rights recognized the value of wildlife habitat and were freely exchangeable on the open market. But today's political incentives work against such rational outcomes. This is just one example of a program that violates reasonable standards of economic efficiency, ecological sensitivity and social equity.

America's founders understood that when governments control large sectors of an economy, governance becomes an exercise in relatively peaceful, although involuntary, wealth transfers. Stealth and duplicity are great assets in this game. It often takes great skill and dedication to discover these transfers.

But the benefits of reform are widely diffused. No individual has an economic incentive to advocate reform as strongly as the beneficiaries do to preserve existing institutions. Large permit holders such as Metropolitan Life have more to gain from their subsidies than any single one of us will gain by fighting them. Today, only the congenital do-gooder has incentives to expose the waste and inequities of political mischief. I propose a modest reform: Let's create a Predatory Bureaucracy.

The logic is simple and straightforward. It is part of the convergence developing between ecological and economic theory. Essentially, the predatory bureau would receive a small percentage of any budgetary proposal offered by an agency. Here's an outline of the process I initially envisioned.

For most of us the term "bureaucracy" carries a crust of derision. Rigidity, inflexibility, insensitivity and persistence are their defining characteristics. However good our intentions in creating an agency, its budgetary imperatives ultimately lead it to stray from these goals. Over the long haul, bureaucracies are usually run for the convenience of its leaders and the clientele on whom they depend for political support. For those of us wanting reform, our design problem is to link this persistent tendency with the public interest. The predatory bureaucracy does so.

We can learn a lot from a relatively successful example. The U.S. Post Office is closely and consistently monitored and faces increasing competition from Federal Express, fax, etc. Despite our complaining, the Post Office actually does quite well because it faces the kind of feedback we expect in the market. What can we do about agencies that do not have these checks on their performance? Here is where our "predatory bureau" can help advance the public interest.

A predatory bureaucracy survives only by preying upon other bureaucracies' budgets. Like coyotes to rabbits and wolves to elk, it will eliminate many vulnerable programs.

The basic idea is simple enough to write into a state initiative. We should create an agency whose income will only come preying upon the specific program budgets of other agencies. Suppose an agency proposes a new program, or seeks to continue an obsolete one. Our predatory agency may challenge this proposal. It will use the press, interest groups and other political avenues.

Unlike most of us, this agency has a vital interest in defeating a wasteful idea because (unlike most agencies) it only gets funding from its successes. If the other agency's proposal loses, it receives 1 percent of the defeated agency's budget proposal. And the defeated agency loses that 1 percent and gets a proportional amount taken from its operating budget.

After an initial two-years' funding, this agency will cost taxpayers nothing. Its income comes only from other agencies. Because all agencies will begin to act more wisely, on pain of being attacked by our predator, enormous amounts of tax money will be saved due to weak programs not being proposed.

Incorporating a predatory bureaucracy into government may do more than elections to control government waste. It's a way to harmonize the public interest with the self-interest of experts in the political process. Let's consider this idea while waiting for next November.