KS Liberty revisionist history…

The same KansasLiberty article I addressed yesterday also included the following quote:

“The court’s only evidence was a discredited school audit that had been dismissed by the Legislature, which had commissioned it, as impractical and dangerously expensive. The audit’s figures had been compiled by asking school administrators how much money they would like to receive to achieve a range of outcomes.”
Wow… that’s one twisted interpretation. While the Augenblick report was far from stellar research by any stretch, it was, as you note the report that had been commissioned by the legislature. It was pursued on recommendation from a Governor’s Task Force (which also included legislators). Further, the charge of the report was specifically linked to the constitutional language which had also been referenced in the Task Force recommendation. And while the report had been shelved by the legislature, they never criticized it openly until after the case went against them in the lower court. It had not been discredited by legislators when the case went to trial and it was not discredited by any of the experts at trial.

Most of the evidence cited by Judge Bullock as basis for determining the constitutionality of the school finance system was unrelated to the Augenblick study. It was evidence related to disparities in funding across districts, which harmed certain populations and were associated with large persistent disparities in educational outcomes. The Augenblick study was cited primarily as an option for achieving constitutionality – a guideline – the legislature’s own guideline – one possible guideline for achieving constitutionality. It was the only one on the table. Would you have preferred the Judge to make up his own? Seems logical and fair to use the legislature’s own target.

Further, when the legislature was granted a do-over in 2006 by the Supreme Court, they commissioned two additional studies overseen by the Division of Post Audit. From these studies they got roughly the same answer they had gotten previously, and once again chose to ignore the findings.

Eventually, the court let them off the hook after a modest (and still inequitable) infusion of some additional funding, dismissing the case while leaving open the question of whether the system is actually constitutional.

For a thorough review see:

Click to access kprv27n2.pdf

Kansas Study Finds Huge Infusion of Cash Didn’t Help Schools

Today, the self-proclaimed non-partisan news outlet KansasLiberty.com reported that a scholar at the University of Kansas Business School has come to the brilliant conclusion that the massive infusion of funding that went into Kansas public schools from 1997 to 2006 has led to no substantive improvements in student outcomes. The author of this report presents a handful of graphs and a few crude regression models of the relationship between spending change and outcome change between 2004 and 2006 for test scores and 1997 to 2005 for “persistence” measures.

Click to access TR08-1205–EducationSpending_Neymotin.pdf

Setting aside academic nitpicking over the absurdly crude statistical analysis provided by the author, lets take a look at the basic question of whether there even was a massive infusion of funding to Kansas schools from 1997 to 2005.

Data on unified K-12 school districts from the Fiscal Survey of Local Governments – Public Elementary and Secondary Education Finances suggest otherwise – In fact, Kansas fell further and further behind national averages during that very period. And, any court ordered funding that went into the system, did so after the period investigated.

year Nation Kansas Gap
1990 $ 4,422 $ 4,105 $ 316
1991 $ 4,683 $ 4,252 $ 431
1992 $ 4,779 $ 4,536 $ 243
1993 $ 5,072 $ 4,944 $ 128
1994 $ 5,242 $ 5,157 $ 84
1995 $ 5,415 $ 5,261 $ 154
1996 $ 5,568 $ 5,436 $ 132
1997 $ 5,751 $ 5,550 $ 201
1998 $ 6,020 $ 5,786 $ 234
1999 $ 6,336 $ 5,946 $ 390
2000 $ 6,722 $ 6,243 $ 479
2001 $ 7,191 $ 6,547 $ 644
2002 $ 7,525 $ 7,048 $ 477
2003 $ 7,854 $ 7,291 $ 563
2004 $ 8,123 $ 7,519 $ 604
2005 $ 8,530 $ 7,706 $ 825

Okay… so let’s take a second look. Here’s Kansas versus the nation after adjusting simultaneously for regional differences in labor costs and for inflation in labor costs, using the National Center for Education Statistics, Comparable Wage Index. Now… Kansas is higher than the national average, because Kansas is a very low cost state.  But, notice that Kansas current spending has been pretty much flat over the period after controlling for changes in competitive wages.

year National Kansas
1998 $ 6,558 $ 7,251
1999 $ 6,579 $ 7,102
2000 $ 6,609 $ 7,041
2001 $ 6,801 $ 7,078
2002 $ 6,748 $ 7,238
2003 $ 6,858 $ 7,277
2004 $ 6,851 $ 7,290
2005 $ 6,987 $ 7,328

So why does this matter? Let me connect the dots here. If there was no massive infusion of cash into Kansas schools, one could not possibly measure an effect of such an infusion on educational outcomes. There must be a shock to the system to measure the effects of that shock on the system. Pretty simple. Further, even if funding had crept incrementally upward across all districts while outcomes simultaneously crept incrementally upward, one could not distill statistically a relationship between the two – Certainly not over a period from 2004 to 2006, only a few years later. Interestingly in this case what we have are outcomes drifting upward with resources staying constant (with respect to competitive wage growth) or negative with respect to education spending growth nationally.

One earlier, more rigorous, peer reviewed study in the Economics of Education Review, addressing the infusion of funding to Kansas schools during the 1992 reforms, found via fixed effects (change over time) regression:

“Using panel models that, if biased, are likely biased downward, I have a conservative estimate of the impact of a 20% increase in spending on the probability of going on to postsecondary education. The regression results show that such a spending increase raises that probability by approximately 5%.”
Deke, J. (2003). A Study of the impact of public school spending on postsecondary educational attainment using statewide school district financing in Kansas. Economics of Education Review, Volume 22, Number 3, June 2003 , pp. 275-284(10).

Other studies on Massachusetts, Kentucky and Vermont have produced similar findings following major structural changes to aid formulas. I’ll gladly share the citations for those interested.


NOTES from a School Finance Curmudgeon

On my daily commute today I had the pleasure of listening to a talk radio discussion about the option for providing tuition tax credits in New Jersey to be used for sending talented elementary and high school students from poor urban New Jersey school districts to either affluent suburban public schools or to urban catholic schools.

As is standard fare in any heated political debate, many “facts” and “figures” were thrown around by the show host in an effort to support the move toward tuition tax credits.

Here is a quick synopsis of the facts and arguments presented, along with my response:

The host noted that Abbott districts (poor urban districts that won a lawsuit against the state legislature) spend $25,000 per kid per year.

Fact Check: The average current operating expenditure per pupil in Abbott’s in 2007-08 was $15,584 with a range of about $12,000 to $22,000. Current operating expense is the most relevant figure to compare. http://www.state.nj.us/education/guide/2008/abbott.pdf

The host also suggested that Abbotts spend much more on administration than other districts.

Fact Check: Both Abbotts and non-Abbotts spend about $1,000 per pupil on administration with Abbotts slightly outpacing non-Abbotts by less than $100 per pupil. Abbotts spend a smaller share of their budgets on administration (7.5% as opposed to 8.8%), which is probably the most relevant comparison. Using NJDOE comparative spending guide (2007-08) definitions (indicator 9). http://www.state.nj.us/cgi-bin/education/csg/08/csg9.pl

The host also argued that we might save money by sending kids to Catholic schools which “cost” about “$3,500 per kid (as the host opined).” The $3,500 figure comes from nowhere (It is a relatively standard figure tossed around, but I’ve yet to really track any legitimate source for it). Further, if this was even near the tuition price that tuition price would not represent the actual cost of schooling that child, even in the private catholic school – which relies heavily on church subsidy.

Few good analyses exist on the actual total per pupil costs of private schooling. One of my doctoral students a few years back went into 4 private Catholic high schools and detailed out their cost structure finding them to be somewhat lower than area public schools in Kansas City (spending about $6,046 per pupil in 2004). These findings are reported in the school finance textbook I authored (Financing Education Systems, Merrill-Prentice-Hall). Salaries are somewhat lower and class sizes slightly larger (catholic compared to public) leading to a slight reduction in cost. The change in student body composition (relative to poor urban schools) allows them to accomplish this without substantially compromising outcomes.

But… Let’s say Catholic schools in this area actually cost about $10,000 to $12,000 per pupil to run. If we gave $6,000 vouchers to kids to go to Catholic schools we would be asking the Catholic Church to cover $4,000 to $6,000 of their cost. For 10,000 kids we’re now expecting an additional 40 million to 60 million in church sponsored (or additional) philanthropy in addition to corporate philanthropy which generated the voucher amount. As the program scales up, the additional philanthropy required to meet actual cost simply can’t be achieved. To make something like this scalable, you have to willing to pay the full cost, not tuition rate. That might be feasible for Catholic schooling, at roughly break even (a) setting aside additional transportation and program management costs and (b) assuming sufficient supply of catholic schooling.

Regarding actual private school expenditures for private schools that are required to file an IRS form 990. Here’s how the numbers shake out in New Jersey.

Public schools by their District Factor Groups

DFG Total Current Exp. PP (2006-07)
A $14,524
B $12,247
CD $10,481
DE $10,520
FG $11,068
GH $11,894
I $11,826
J $12,179

Private schools by affiliation (IRS 990 2007, for 2006)

Independent Private $21,929
Hebrew/Jewish Day Sch. $17,130
Christian (AACS, ACIS) $5,902

Based on a sample of 45 private day schools, serving over 18,600 children (2006-07). Includes “program expenditures” and “administrative expenditures” as reported on form IRS 990 (www.guidestar.org). By comparison, Charter schools served approximately 13,000 and DFG J districts approximately 50,000.

Sadly, Catholic schools which are the largest group do not report their finances. We do know from regional and national data that the conservative christian schools that are members of AACS and ACIS, at their relatively low expenditure, actually have salaries that are over $5,000 per teacher lower than catholic school teachers at same degree and experience level, and comparable assignment (in the Norheast, in 2003-04 based on the National Center for Education Statistics Schools and Staffing Survey). It stands to reason that Catholic school expenditures are somewhat higher than AACS and ACIS schools. 4 years ago, in Kansas City (a much cheaper labor market), Catholic high schools spent over $6,000 per pupil.

We also know that conservative christian day schools have the academically weakest pool of teachers (high shares from the least competitive colleges and very low shares from the most competitive colleges – typically a much weaker pool than in public schools) and lowest student outcomes (corrected for student characteristics). You get what you pay for – though most who attend these schools are buying religious separation – not academic advancement. Catholic schools have teaching staff quite similar to public schools in terms of qualifications, and adjusted test score outcomes that are about the same (acknowledging that “adjusted” is not what matters to individual children. As much as anything, it’s about getting kids into a setting with different kids).

In addition, private school headmaster salaries outpace both big city superintendents in NJ and the superintendents of the mainly affluent suburban districts where those private schools exist – this despite the fact that these private schools serve total school enrollments of 300 to 1100 students. That’s what the free market for non-church subsidized private schooling supports.

Big City Supts. $196,039
Headmaster Mean $214,309
Host Supt. Mean $184,802

Note: Private school headmaster compensation from Guidestar.org, IRS 990 for 2006. Local Superintendent compensation for district that is geographic home to private school. Supt. Comp based on 2006-07 (1 yr later than Headmaster Comp.). Headmaster and Supt. Comp. include salary and cash-basis benefits (not health-care, retirement contributions, etc.). Big City Supts. Includes Newark, Camden, Jersey City, Paterson and Trenton.

Here’s the distribution of private school enrollment slots in NJ:

Type Enrollment
Catholic Schools 100,562
Other Schools 17,453
Christian Schools 11,939
Independent Schools 9,598
Special Emphasis 5,249
Jewish/Hebrew Schools 4,996
Friends Schools 1,480

Data Source: NCES Private School Universe Survey. Note that there appears to be substantial under-reporting of private independent schools in this survey.

I could go on… But I need to actually get some other number crunching done today.


%d bloggers like this: