Thursday, January 18, 2007

Inflating the Minnesota budget 

Gary has posted something on the new budget proposal, featuring tax relief, that Michael reported earlier, and on which I wrote the other day. Commenter "Deminn" -- a regular on the Times chat board and a rather iconoclastic fellow, not one to take dogmatic positions -- started his note thus:
There is a �surplus� because of Republican magic which at one time attempted to deny the laws of economics and decree that inflation is something that only happens to Minnesota Citizens,...
You can look at this two different ways. One, if you believe that any level of government activity budgeted once is budgeted forever, then not budgeting government spending with inflation adjustment would be seen as a cut, and you'd want to portray expenditures with the inflation adjustment. But why would you choose this as the basis? If private companies always budgeted costs to rise with inflation without any increase in activity -- and ergo, revenue -- these firms would soon dry up of money. I'd as soon see governments engage in zero-based budgeting than use budget forecasts either with or without inflation adjustment, a zero basis meaning no need for a forecast, but at least zero-inflation budgeting puts some pressure on government servants to be more efficient.

Second, once one decides to budget for inflation, the question is which inflation rate to use. This isn't a simple choice between CPI and a few alternatives; it is a question of how one calculates the price of a government service. CPI itself includes governmental fees and some taxes, so there's a feedback loop between government size and CPI. And since the largest part of government spending at the state and local level is personnel, and since health care costs rise faster than the CPI generally, there would be a push for a higher inflation adjustment ... without any assurance that the money actually goes there.

Even if you could agree to a particular index, the question then is how to forecast it. Every forecaster deals with a loss function -- what is the cost of an error? And in particular with government budgeting, the cost to the political system of an error that requires a tax increase or spending cut in the second year of the biennium to balance the budget is much greater than cost of overbudgeting in year one and holding a surplus in year two (all the better for giving away in an election year run-up. Nice deal there -- they take credit for borrowing money from you and paying you back with zero interest!) The budgeting process would then have an inherent bias towards conflict in politics and likely to overspending and overtaxing.

The bill
should be rejected -- the current system actually works quite well, protecting almost all of government spending while allowing gradual adjustment of the budget to shifting priorities.