Other Budget Rules

Hammer's Budget Process

Hopefully each group comes up with a plan reflecting their values and their meetings are quick. Each plan becomes one of the candidates for all voters to rank in a Condorcet Pairwise tournament tally. And the winner is... a reasonable budget with majority support.

One advantage of the Hammer Budget Process is its affect on proposals. As always, some people exaggerate their own needs hoping to influence others' assessment of those needs. But people who want their plan to win, strive to understand what other voters want and to moderate their own demands.

Professor Hammer made no claim that the winner would be optimal in any sense (Pareto or Nash), only that it would be a coherent plan with majority support. While this rule does not try for consensus, it does minimize number crunching and maximize human judgment and creativity for finding win-win solutions.

(Hammer was a leading member of Twin Oaks Community in its early years. He left to teach architecture and computer science at Cal Poli. His skills blended practical group process and number crunching very effectively for community ideals.)

There are counter arguments to the Hammer Budget Process:
1) It lets a ruling majority ignore the needs of a minority.
2) It does not require each rep to rank her priorities publicly.
3) It depends on members learning a lot about others' priorities; this can come from discussions, papers or ballots for other voting rules.

Other rules, such as Fair Share Voting or those below, may contribute another economic plan to the final Condorcet ballot.

Median Voter Process

Of course, a disciplined majority faction can set all medians exactly where they want the budgets. So this is not a fair-share rule.

A rep might be allowed to give two votes to the department she cares for most. Her extra weight on that decision must be balanced by losing weight on setting a couple of other budgets.

It might not lead to a balanced budget overall, even if each voter's ballot is required to balance. (The detail page on Median Voter Process gives the simple formula that adjusts all budgets to limit their total.)

The rep whose vote is the median for a department can act alone to change its median up or down to the closest votes by other reps, just above or below hers. You could say she is the "most valuable player" in that department.

But it is relatively easy for another rep to take her power. MVP encourages vote trading by pairs of voters on opposite sides of two budgets. Alex might say to Bobby, "I'll raise my previous low vote for your committee if you'll raise yours for my committee." This is time consuming and the results are not very rational because many changes to the budgets depend on who is quickest to find a deal, not on any deep or persuasive thinking about the effects.

The median vote rule finds a rough balance of political forces in numbers where rules such as Budget Refill Voting and Voter Influence Points measure the strength of each rep's desires.  MVP can do that too.

Where MVP is used to set the initial budgets before a fair-share rule, it might tempt an off-center majority to vote extreme initial budgets -- so it is best used with direct democracy, direct representation or ensemble councils.  It might set the initial budgets only for new agencies while old agencies start from their current budgets.

VIP, Political Cards

A voter can show how strongly she supports each budget by allotting her "budget cards" to a few favorite agencies. But we must deter exaggerated votes by limiting big votes with soft constraints or hard limits.

The tabletop tally had hard constraints, limiting voters to 2 or 3 sizes of cards for self-expression. The same rule with soft constraints lets voters build large cards from small ones but discourages exaggerated votes by the "square root rule":
To raise a budget by 1 card, you just give it 1 card. But to raise a budget by 2 cards, your pile must be as deep as it is long: 2 cards long by 2 cards deep. That makes 4 cards total. To raise it by 3 cards, each card must be 3 deep, 9 total. The depth discourages dumping all cards on 1 favorite budget. So pushing for a big change uses up more of your influence than pushing for a small change.

The Hylland-Zeckhauser rule for influence points uses the square root rule. It may require voters to balance cards for some budgets with cards against other budgets. That's easy with computer ballots shown in this chapter. Otherwise it is hard for a voter to estimate the current budgets. And it is easy to see and resent voters who have hurt his favorite budgets.

The Budget-Setting Points page shows a similar soft-limit rule.

    Economic Value vs. Majority Rule

When VIP voting stops, we may see that its budgets for some agencies are not majority winners: One department may have a few large cards pushing its budget up, partially balanced by many small cards pushing down. A budget which set that agency lower could win majority support.

But an overall budget which "corrects" all departments with similar patterns would likely lose by a majority -- because the voters pushing, gently, with single cards, care more about the budgets they pushed up or down with large blocks of cards.

Standard-Score Voting

A voter casts points for or against as many line items as she likes or dislikes. The points mean the percentage change she wants in the budget. A large budget must be broken into several line items, so large and small budgets are equally hard to raise or cut.

A voter's average gets adjusted to zero. And her points are “normalized” into a rough “bell curve”. This processing puts soft limits on gaming the system with extreme votes. A voter can push very few items high or low. The normal bell curve ranges from -3 to +3. But those numbers may be multiplied by [5].

For each item, we average the points from all ballots. A zero vote counts in the average.

Benefits: Vote trading within a coalition can game the system only a little. In this way it is better than the Clarke Tax or the square-root rule by Hylland and Zeckhauser.

A voter needs an interactive ballot to show her standardized scores and make this decision rule easier to understand. PoliticalSim^™ uses this rule, with such a ballot, to let players select the election rule for a game.

Fair Shares by Turns

Unfortunately, many political situations, including these, reward good strategy more than good policy. It is as though the budgets depend on each rep's skill at poker: her speed at scanning the table, her awareness of others' goals, and her bluffing.

The tedium of voting one card a minute might be solved by automating the process with software. Each party or rep could use their own program to place their cards. But that just makes this a very fast poker game.

    Comparing MMV with BRV or VIP

A survey committee (or the voter) sets a utility curve within limits, often zero to two.   The voter may rank items to set the order in which they fill the pre-defined curve.   Or he may score items to create a unique curve.

Project-funding ballots only need to limit large votes, to prevent bullet voting by dumping all money on one or two favorites.   Agency-funding ballots need limits at the bottom also, to prevent free riding by contributing zero to an essential shared good.

An MMV tally for projects transfers money automatically from items that fail to reach a quota of support (and any excess from those that do reach quota).   A BRV or VIP tally for agencies makes voters respond to each others' ballots, transferring money personally as they create and counter budget changes.

Rules for setting policies can be used for adjusting budgets one at a time, with the previously mentioned problems inherent in that process.