Friday, March 23, 2007

Stanford report #4 - Misunderstanding Incentives; Ignoring Reasons for the Evisceration of School Funding

Stanford Report #4
Understanding the Incentives in California’s Education Finance System, Duncombe, William and Yinger, John (2007), Syracuse University.

In two sections -- [my comments in blocks and italics; boldface is my highlighting their text]

I. UNDERSTANDING INCENTIVES [summary of state financing follows the incentive section, which is the reverse of the actual report #4]

[From Appendix D]
A ten percent increase in the API requires a 7.1 percent increase in spending, all else equal. We combine this result with various assumptions about school district efficiency to obtain estimates of the cost of reaching the 800 API target. [and this is useful info how?]

[HOW DO THEY ARRIVE AT THESE KINDS OF CONCLUSION?! Look at a paragraph from another part of the paper below to see how they deal with assumptions upon which they base these math formulas -- notice “but this adds amazing complexity” which they don’t do because they need to keep the math formula “manageable” – so human actions, motives, behavior are simplified so they can fit it into a math formula that predicts human behavior, ie what kind of tax would the voters support????]

. . . Finally, this issue is complicated because it introduces another margin—whether or not to increase property taxes. All existing models have a single margin and that margin determines the tax price. We don’t think it makes sense to have two tax prices. It might it make sense for a voter to pick the smallest tax price (that is, to increase the revenue source with the smallest impact on her). But this adds amazing complexity to the model and to the empirical work. We would have to calculate two tax prices and use the smallest—both in determining monitoring and in the demand model. To keep the analysis manageable, we assume that voters do not perceive borrowing for operating expenses to be the relevant margin when they make their monitoring decisions. Instead, they see money raised through this route as a contribution to augmented income. This assumption is somewhat awkward because the alternative margin, the parcel tax, requires a 2/3 vote and is rarely used. But the restrictions on borrowing for operating expenses are severe and the parcel tax logic is straightforward, so we keep the parcel tax at the core of the model. This leaves open the possibility that borrowing for operating expenses is a “margin” in the demand equation. In other words, the possibility of borrowing for operating expenses might influence the demand for school quality either through a price incentive or though augmented income, depending on voter perceptions. We can estimate it both ways.

An education cost equation indicates the amount of money a district must spend per pupil to obtain a given level of student performance. This type of cost equation is analogous to a cost equation in private production, which is a central tool in economics. [Output = API score, Input = teachers’ salaries. ] . . . A cost equation indicates how much a school district would have to spend to achieve a given performance level if it used the best available technology, that is, the best available teaching methods and management policies. We cannot observe costs in this sense, however, but instead observe actual spending, which may not reflect the best available technology. To put it another way, the dependent variable in our analysis, spending per pupil, is equivalent to educational costs divided by an index of school district efficiency. As a result, we need to control for efficiency to preserve the cost interpretation of student characteristics and other cost variables. . . . . spending on art is likely to a source of inefficiency in the production of mathematics performance. Our main analysis measures school district performance using API.

[I don’t even know where to start with taking apart all the horrific assumptions that underly the statements in the above paragraph]

[do they really believe that this kind of talk is helpful??] School officials in high revenue/high performance districts may work especially hard to maintain the performance level in their district despite the decline in their revenue. Similarly, school officials in low revenue/low performance districts may be able to use the relatively large increases in their revenue limits to increase their relative performance even without being relatively efficient.


[the report summarizes prop 13 and Serrano, but doesn’t explain how the commercial real estate loopholes have contributed to an evisceration of school funding – without explaining that, they don’t reveal an important way to increase funding – close the commercial loopholes in prop 13. PICO attempted a campaign to do so several years ago, and CTA began a campaign last year. Both were told by the CBR not to do it and were stopped by the power of CA CEOs]

Proposition 13 sets the parameters of the expected local contribution . . . . In a fundamental sense, these restrictions on local supplementation transform “local” revenue into “state” revenue, that is, they take decisions about the property tax and other local revenue sources out of local control.. . . the Serrano decisions reduced revenue-limit differences across districts . . . . the amount of revenue is much higher in New York, but that, in some cases, the variation in revenue is higher in California. . . . [the variation] shows up in state categorical aid and in federal revenue.

… heavy reliance on categorical aid, as in California, gives state lawmakers more control over the allocation of school district budgets, but it also limits local flexibility and innovation and raises the share of resources that are devoted to bookkeeping instead of education. This trade-off is recognized in California. The six AB 825 block grants, which were passed in 2004, each combine several categorical aid programs and give school districts more flexibility in deciding how to spend the aid funds (Goldfinger, 2006). We believe these block grants were a step in the right direction. . . . . Categorical aid programs may therefore add to district inefficiency not only by adding constraints, but also by being unpredictable.

this “poverty gap” [Ed trust’s study between 25 districts with most and 25 with least poverty] is considerably lower in California than in the average state or the average big state. Nevertheless, this gap is still quite large, $534 per pupil, and it is larger than the gap in other West Coast states, $309 per pupil.8 To some degree, the relatively small absolute poverty gaps in California reflect the state’s relatively low spending per pupil. [and why are there high poverty and low poverty districts to begin with ...? oh, that's not part of the discussion is it?]

Overall, there is still plenty of room for debate about the features of each component in the California state revenue system, but the mix of revenue sources does not appear to be in need of a major overhaul. . . . Table 7 also indicates that taxes other than the property tax make up a very small share of local school revenue in California, as they do in other states. . . . the main source of this difference appears to be that in California, unlike other states, private educational foundations make large contributions to public schools in several school districts. . . . the parcel tax provided 8.0 percent of local non-property-tax revenue in 2002-03, compared with 5.6 percent in 1995-96. This growth in the reliance on this tax is likely to continue, because the parcel tax is one of the few revenue sources with clear expansion possibilities, but the 2/3 voting requirement obviously has minimized reliance on this tax up to now. . . . According to the criteria developed by public finance economists, a parcel tax is a poor substitute for a property tax. First, a parcel tax does not meet basic standards of fairness . . . it is very regressive.

Revenue from private foundations is distributed in a relatively inequitable manner, because districts with richer residents can attract more contributions. Unlike state aid given to wealthy districts, however, these contributions do not come out of the same budget that must provide funds to the neediest districts. Moreover, private contributions may be difficult to regulate; even if contributions directly to schools were prohibited, parents could make equivalent contributions through tutoring, art, music, sports, or other programs run outside the school system.

The foundation amount in California is not adjusted in any way for the higher cost of education in some districts. This situation is consistent with the Serrano guidelines for a minimal revenue-limit deviation across districts, which is expressed in terms of actual spending per pupil, not in terms of real spending per pupil (that is, spending per pupil adjusted for educational costs). The California approach is not consistent, however, with the central normative argument for a foundation formula, which is an education finance system should lead to an adequate student performance in every district. This view requires not only a decision about the resources needed to reach this performance level in a typical district, that is, a decision about the basic foundation amount, but also a decision about the best way to account for variation across districts in the cost of education.

[notice in this paragraph the paradigm IN which they are operating—the criterion for creating an incentive for financing should be “performance level” – in other words, standardized test scores. ]

[photos from Rethinking Schools]

No comments: