Oregon is changing the way it calculates how much money to give each school district to help poor kids. One aspect is clearly correct: It will use up-to-date demographic information instead of the long-outdated 2000 Census.

But early estimates based on the new metric show extreme volatility from one year to the next, raising questions about accuracy and creating planning difficulties for districts.

Bend-La Pine Schools Superintendent Ron Wilkinson told us the first estimates in December, based on 2011 data, showed his district would lose $650,000. The next round came early in the new year, based on 2012 data, and showed Bend gaining a bit back. In late January, the latest estimate showed a reduction for Bend of $279,886.

That’s a lot of variation over a period when there was little change in the district’s free-and-reduced lunch percentages — the only poverty indicator the district can measure.

Bend wasn’t alone in seeing large swings in the estimates. For the Salem-Keizer School District, the first estimate was a gain of $4.1 million, while the more recent number showed an increase of $2.3 million, according to the Statesman-Journal newspaper.

The new state allocation formula is based on the U.S. Census Bureau’s Small Area Income Poverty Estimate (SAIPE), which considers IRS and food stamp data in addition to Census information. While sending fewer dollars to Bend-La Pine, the new calculation will increase the Redmond School District’s allocation for next year by nearly $2 million and hike distributions to other Central Oregon districts as well. Across the state, 80 districts will receive additional funding while 117 get less.

Wilkinson said if the big swings continue, districts will have difficulty planning programs for poor kids, possibly prompting some to put the money in their general funds rather than focus specifically on kids in need.

Another issue, he said, is that the formula doesn’t consider the impact of Oregon’s school choice program, which allows students to move to another district. Students in poverty will be counted in their home district even if they are attending school in a different district.

In approving the new approach, the State Board of Education decided to accept the volatility, rejecting pleas for a phase-in or averaging that would avoid abrupt changes. It may need to revisit that choice if there’s evidence the volatility will interfere with needy children truly getting help.