Original Research

So Just How Far Off Were My Tuition Projections?

Back in February 2011 I made a bold prediction: Full-time tuition costs at private ABA law schools would increase.

Talk about sticking to your guns and throwing conventional wisdom out the window!

But enough self-congratulations and I-told-you-sos. I offered projections for each law school, which proved so popular that a handful of Web sites even reported on them, motivating me to update them annually. Although I’m always pleased to receive positive press, I ceased making new projections when it became clear that tuition growth was going to slow down due to the applicant crash. (Also, the methodology posts were mind-numbing to write.) No tweaks to the methodology would create accurate results, so that was that. Nevertheless, time has passed; we’ve caught up to the first projections, and I’m curious how far off (or on?) they were. Maybe we’ll learn a lesson.

My original projection methodology estimated that mean-average private-law-school tuition (excluding the two private Puerto Rico schools, as always) would rise from $38,097 in 2010 to $47,598 (25 percent) by the 2015-16 academic year. Later, I believed that methodology produced results that were too inaccurate, so I revised it, giving a mean-average tuition of $46,341 (22 percent) by 2015.

These growth rates are impressive, and when I offered them I chose not to adjust them to inflation because I didn’t want to predict inflation and I was convinced that consumer-price inflation wasn’t really playing much of a role in law-school tuition anyway. In fact, the consumer price index has only risen by 8.7 percent in this five-year time period. In hindsight it may’ve been appropriate to compare tuition to the CPI’s higher-education cost index, but I think no one is worse off.

So how did I do? Thanks to the ABA’s 509 information reports, I get $44,413 mean-average tuition at the private law schools that were around in 2010—and were private law schools in 2010, for some have been socialized, e.g. Texas A&M, which used to be Texas Wesleyan. On average, tuition is 17 percent higher than 2010, so my average was high by 4 percent. Yikes, but at least the savings went to law students.

But as we learned once again recently, the mean average isn’t useful without the dispersion. Yeah, that damn average is made up of real observations that may indicate patterns of their own. So here’s the variance.

Percent Variance of Projected Private-Law-School Tuition (2015)

You can see there are a few outliers, which I’ll go into, but overall, the horizontal zero line cuts fairly closely through the body of the points. In fact, the median projection was off by less than 3 percent. Variances below the zero line, however, tend to clump together more.

So who are these outliers?

No. 1 is La Verne (82.5%), which gave up on merit scholarships a few years back in favor of flat costs for all. I figured it’d charge $48,027 in 2015, but in fact it cost $26,323. You can take this as evidence that tuition can’t go up forever.

No. 2 is the school I thought would’ve been number 1, Faulkner (45.6%), which doubled its price tag within a few years of receiving ABA accreditation. This was certain to throw off my methodology, so no one believed it would cost $51,045 today. Still, I didn’t expect it to go up by only 13 percent since 2010 ($35,050). That barely beats inflation.

No. 3 is a school I didn’t expect to see on the list, Ohio Northern (41.5%), which cut its tuition from $31,264 in 2010 to $26,030. I thought it’d charge only $36,820.

No. 4 is another unexpected tuition reducer, Roger Williams (33.3%), which costs $34,596 now rather than the $46,128 I expected.

Nos. 5 and 6, Elon (24.8%) and New Hampshire (23.2%), barely raised their tuition at all, so it’s no wonder their projections were off.

No. 7 is another tuition reducer on the list, Brooklyn (23.1%), now $46,176 when I thought it’d be $56,862. It charges only 1 percent less than in 2010, and that’s nominal dollars.

I’m not going through the rest, but the one law school I expected to see further up was New England (11.5%) because like Faulkner it raised its costs by quite a bit in the mid-2000s. I guess it just kept going. Finally, among the for-profits, Arizona Summit, Atlanta’s John Marshall, Charleston, and Charlotte all came in below their projections by at least 10 percent. Florida Coastal and Western State overshot theirs by about 5 percent.

The bottom-end variances aren’t as noteworthy, but congratulations are in order to the most expensive law school in the land, Columbia (-3.7%), for raising its costs more than I thought. A bunch of other expensive, well-regarded law schools also outdid my methodology. Good job. Not.

However, that lone dot way below the zero line at $47,980 is … WMU Cooley (-23.4%). I thought it would charge only $36,680. I’m not in the mood to research why it jacked its price tag so much, but it probably has to do with the school’s large part-time program. Maybe it’s trying to deter people from the full-time program?

Overall, I would’ve guessed the median variance would’ve been over 3 percent, but in general the projections tended to skew higher rather than find their marks. Meanwhile, I count 28 law schools (about one-fourth of them) that varied from their projection within the -2-to-2 percent band. That’s about $900 at the median law school.

In all, there was a nugget of accuracy to the initial projections, but I don’t take credit for predicting that; rather, I was right that the projections would overshoot reality. Private law schools slowed their tuition increases over the last five years. That’s small potatoes for the students though.

Below the fold, here’s a list of private law schools by cumulative cost increase between 2010 and 2015, along with information on their projected 2015 tuition.

(more…)

Office of Management and Budget: +$1.1 Trillion in Direct Loans by 2026

…Which is down from $1.4 trillion by 2025 as predicted last year.

Every year in July the Office of Management and Budget (OMB) publishes its Mid-Session Review of the budget, which includes the Federal Direct Loan Program, and projects its future. The federal government’s direct loans consist primarily of student loans, but there are a few other programs in there. However, federal direct loans do not include private student loans, but these are a small percentage of all student loans. Thus, the OMB’s measure is both over- and under-inclusive of all student debt, but it covers most of it.

The OMB classifies direct loan accounts as financial assets totaling $1.144 trillion in 2015. According to the office’s projections, by 2026 this figure will grow to $2.213 trillion—93 percent.

Projected Direct Loan Balances (OMB, Billions Current $)

(Source: OMB FY2017 Mid-Session Review (pdf))

As with previous years, the current direct loan balance is below the OMB’s past projections. For FY2012, it predicted the balance would be $1.363 trillion by 2015, $219 billion (19 percent) higher than what actually occurred. Even last year, the OMB’s estimate for 2015 was still high by 4 percent. Here are the OMB’s direct loan projections going back to FY2010.

Direct Loan Balance Projections (OMB Billions Current $)

Because the OMB expects GDP to grow as well over this time period (we’d have bigger problems than student loans if it didn’t), the ratio of direct loans to GDP will level off below 8 percent over the next decade.

The OMB’s measure of direct loans is the net amount owed to the government, and the annual changes to that amount are not the same as the amount lent out each year to students. The Department of Education tracks its lending, and I last discussed it here. As of 2015, fewer students were borrowing from the federal government, so lending appears to be declining. The newly implemented gainful employment rule might further reduce student lending as well. These factors may explain why the OMB’s projections keep falling short. Consequently, I don’t believe student debt will exceed $2 trillion.

Let’s Shift St. Paul, Minnesota’s, Property Taxes Onto Land!

At last, I’ve found time to write! This will be the first in a while for a while. A few months back I threatened readers with posts on real-estate mapping and assessment, and I make good on my threats.

Background: When I moved to the Twin Cities, I joined Minnesota’s Common Ground chapter, and this year its intern introduced some of us to the black arts of geographic information system/science (GIS) software. GIS is a discipline full of jargon, like “vectors” and “table joins,” that I find silly, but they’re certainly more sensible than “JD Advantage.”

The purpose of our inculcation is to help Common Ground advocate for land-value tax districts in Minnesota cities as a pilot program toward enabling municipalities to adopt split-rate taxation like Pennsylvania’s. The goal is to observe the effects of removing taxes on improvements and replacing them on locations to encourage development. Connecticut enabled a pilot program like this in 2014 (and it’s being extended), so LVT of a kind is in the policy air. So far, Common Ground’s efforts have successfully resulted in the introduction of a bill in the Minnesota House of Representatives.

Oh, and I claim zero credit for any of this.

But I do have newfound technical skills to unleash on readers (actually, a lot of the work is done in MS Excel, which is old hat around here), so with the open source QGIS and the state’s wonderful MetroGIS’s database in hand, here’s how a property tax shift would affect St. Paul, Minn.—because I was born there, and it’s the city we started our training with.

Boom:

St Paul Shift Small

Click to enlarge.

What you’re seeing is the distribution of the percent changes in property taxes for each parcel, divided into negative changes (blue, decreased charges) and positive changes (red, increased charges), and excluding tax-exempt properties. Because more parcels’ property taxes would be cut by the shift (yay!) than raised, the number of blue/red parcels isn’t equal.

Here’s a histogram of the percent changes of property taxes by percentile for single-unit residential lots claimed as homesteads, which dominate the city’s land use by nearly two-thirds of all properties and are probably the most salient politically.

Distribution of Property Tax ChangesThe results for these properties is inauspicious. A bare majority, 52 percent, would get breaks, and the remaining homeowners would pay more by comparison.

I’ve included a macro table for reference below, but generally, the tax shift would move the property tax to residential parcels and off commercial lots. Vacant lots would pay more—as they should—but there aren’t many of them in St. Paul.

One of the biggest conceptual problems with estimating the effects of a land-tax shift is that the current property-tax system discriminates among property classes. Residential lots on average pay less than commercial lots—by design. Single-unit homesteaders pay on average 1.52 percent of their assessed values in property taxes, and commercial owners pay 4.24 percent. Meanwhile, vacant residential lots pay 8.08 percent, illustrating existing progressivity in the property-tax system.

Consequently, much of the effect of a property-tax shift is really just equalizing the tax rate on all parcels, eliminating discrimination among property classes. In fact, a land-value tax shift starting from a hypothetical flat property-tax rate that includes buildings is better for single-unit residential landowners. Unsurprisingly, Common Ground Minnesota explores what happens when residential parcels are treated differently than other types in its advocacy.

Let’s return to the map above. Comparing the top 10 percent of property-tax reductions to the top 10 percent of increases, the property-tax burden is lifted most from the northeast part of St. Paul and downtown, and moved to the southwest part of the city, which is mostly residential.

These results don’t fill me with warm fuzzies, but the underlying issue is not who would pay but who isn’t currently paying, i.e. it’s the land-value assessments. Minnesota requires properties to be assessed at their fair-market values. These estimates are fed into multiple formulae to arrive at properties’ tax capacities, and then tax authorities apportion levies against all properties based on these tax capacities. When properties are under-assessed, they receive a hidden tax break; when they’re over-assessed, they opposite is true. Between commercial landowners and homeowners, guess who gets the hidden breaks?

To illustrate why I think St. Paul’s real estate isn’t properly assessed, here’s a map of St. Paul’s land values per square foot, including only the bottom 10 percent of parcels and top five percent.

St Paul LVPSF

Click to enlarge.

I’ve zoomed the map to the southwest part of St. Paul, and while you can see that the downtown cluster includes much real estate in the top 1 percent by value, some of it streaks west along Grand Ave. There are a few other peculiar concentrations of 1 percent real estate: a few blocks south of St. Catherine University (St. Kate’s to the locals), and a few blocks south of the University of St. Thomas’ divinity school (the law school’s campus occupies a surprisingly large chunk of land in nearby downtown Minneapolis). Slightly less valuable real estate lies along University Ave. (where the light rail connects the Twin Cities) and the L-shape along Cretin Ave. and Ford Parkway.

Most of the top .01 percent is downtown (~$56 per square foot), but some of it is still in these areas. Naturally, you may be wondering why property owners with the most valuable land aren’t demanding their properties be rezoned so they can build office towers in western St. Paul. The answer is that much commercial real estate is under-assessed, and many parcels’ values are malapportioned between buildings and land. (I have an acquaintance who recently bought a decades-old house on a plot of land valued at a mere $5,000.)

Although Minnesota’s tax authorities go to great lengths to ensure assessments are fair, notably sales-ratio equalization estimates of numerous parcels, these methods only include properties that were sold in arms-length transactions. One ongoing problem I’ve identified is that commercial real estate transactions differ substantially from those of owner-occupiers. The wealthy simply buy land differently from the rest of the populace.

For example, on November 10, 2015, the First National Bank building sold for $37.25 million but its assessed value as of January was only $25.5 million. Why? Because according to the Ramsey County Assessor’s Office, it was a “Not Typical Market” transaction, so it was disqualified from the sales-ratio equalization analysis.

Here’s another fun one: A Walgreens on Larpenteur Ave. (which might actually be in Roseville) sold for $11.2 million in April 2013, but was excluded from the sales-ratio equalization study because of “unusual financing.” The same goes for the Walgreens on Ford Parkway, which sold for $13.9 million but was assessed at $3.2 million.

Other times sales that qualify for the sales-ratio equalization analysis still result in assessments that are below their sales prices, e.g. the lot on 240 4th St. East, which sold for $800,000 in March 2015 but was assessed at only $286,300 in 2016. The biggest offender I’ve found in my casual search is ten vacant lots along Dunlap St. that sold for $7.5 million and have been assessed at about $3,200 per parcel. You’d think qualified sales of vacant real estate would be assessed at something close to their sale prices, but they’re simply not. This results in large property-tax breaks for wealthy landowners and an increased burden on everyone else.

If I were a St. Paul homeowner (and I know a bunch), I would grab my pitchfork and march on the Ramsey County Assessor’s Office and demand the city’s land values, especially its commercial land, be properly assessed according to law. I’m confident that would shift some of the property-tax burden away from homeowners and onto downtown landowners without affecting their property values. Municipalities should also rely more on mass building-residual assessments to arrive at more accurate land (and building) values, echoing the negative corporate land values in 2009 that I wrote about a few months back. I believe better assessments would make land-value-only property taxes more attractive to single-unit homesteaders than the current system illustrates.

Appendix:

Here’s the macro table of what the LVT change would do to the majority of the city’s parcels.

Macro Tax Shift Table

Click to enlarge, if you dare.

********************

CLASS OF 2015 EMPLOYMENT REPORT

[UPDATE: As with last year, it appears the ABA took down the employment spreadsheet by late Friday afternoon, making this post … an exclusive. There may be substantial revisions to come.]

At last, something to write about! (And time to do it too!)

On Friday, the ABA updated its Employment Summary Report Web site, which provides employment data for each law school class going back to 2010. Many if not all law schools have uploaded their individual reports, and some intrepid researchers have already dug into them, but I prefer to wait until the easy-to-use spreadsheet comes out. Note: There may be revisions to these data, but this first, preliminary cut gives a good sense of the class of 2015’s employment outcomes. Also, I diligently account for all accredited law schools, so researchers should recognize that Concordia Law School must be inserted manually. Indiana Tech has no data.

39,423 people graduated from 204 ABA-accredited law schools outside of Puerto Rico roughly between September 1, 2014, and August 31, 2015. The employment information is good as of about March 15, 2016.

Here’s the employment status distribution.

Class of 2015 Employment Status Distribution

Surprisingly, many of the employment status categories’ percentages are identical to last year, even though the absolute numbers have fallen. I almost thought I was looking at the 2014 data by mistake. Notably, the employment status tables added a section for “Employed – Law School/University Funded” jobs. It’s probable that a good chunk of these jobs were classified as “JD Advantage” until now, further clouding the validity of that category.

The display tables appear below the fold to conserve blog space.

**********

(more…)

2016-04-09 Site Update

I’ve updated the site’s most popular attraction, the Lawyers Per Capita by State page, thanks to updated population and lawyer counts. Kudos to the state bars for all reporting attorney numbers on time. Back in 2014, the Census Bureau’s press department honored the page by using its contents in a “Profile America” feature on the ABA’s foundation.

And for fun, here’s a time series chart depicting the saturation of lawyers by Bureau of Economic Analysis region.

No. Attorneys Per 10,000 Residents by BEA Region

That’s all.

Wage-and-Salary Lawyer Employment Slows in 2015, Incomes Flat

The Bureau of Labor Statistics (BLS) usually completes its updates of its many measures of occupational employment for the previous year by April. Data for 2015 are now available, allowing a comprehensive summary of lawyer employment for the year. For detailed discussion of what the BLS datasets are and how they address lawyer employment, I recommend the lawyer overproduction page [updated!].

For context, according to the Current Population Survey (CPS), the number of people who reported working as lawyers in 2015 grew 2.5 percent to 1,160,000. The employment projections program (EP program) placed the number of lawyer positions at 778,700 in 2014. The discrepancy between these two measures has existed for a long time and has yet to be explained. Although the CPS is considered more reliable, the EP program estimate is appropriate for discussing future lawyer employment. The CPS measures the number of people in an occupation, but the EP program estimates the number of positions in that occupation, including people holding multiple jobs. Both measures include part-time lawyers and self-employed lawyers in all industries.

The CPS also estimated 803,000 people working as lawyers on a wage or salary basis, an implausible 9.0 percent growth from the previous year (+66,000 lawyers). By contrast, the more accurate Occupational Employment Statistics (OES) program found that the number of wage-and-salary lawyers grew by 1.1 percent last year to 609,930. The number of employee lawyers in the legal sector grew only a negligible 0.4 percent to 380,180.

Lawyer Employment by BLS Measure

Employee lawyers’ incomes were flat in 2015. The OES estimated a scant 0.6 percent median hourly wage growth, although the CPS registered a 4.2 percent median weekly wage increase. Going by the OES, the last peak for lawyers’ earnings was 2009; incomes are about $10,000 lower in real dollars since then. Here is an annualized dispersion.

 

10th to 90th Percentile Dispersion of Annualized OES Lawyer IncomesThese lawyer employment measures are not strong bellwethers for the value of legal education because they include many established lawyers and don’t measure recent graduate outcomes particularly well, especially those of graduates who do not promptly start careers in law. Readers are instead advised to look at my criteria for predicting improvements in law graduate outcomes for insight.

Anthony Carnevale Has Two Years to Reemploy 15.8 Million College Grads

Two years ago I made fun of President Obama’s ludicrous claim that “more than 60 percent of jobs in the next decade will require more than a high school diploma.” It appeared Obama appropriated the statistic from Anthony P. Carnevale’s paper for the Georgetown Center for Education and the Workforce (GCEW), entitled, “Help Wanted: Projections of Jobs and Education Requirements Through 2018.” It shrieks on page 22 that 63 percent of jobs created by 2018 would require a college education: 33 percent bachelor’s degrees, and 30 percent associate’s degrees or just some college.

As I wrote in January 2014, Carnevale and his colleagues reasoned that the BLS was holding occupational credential requirements constant when they should drift with times. As non-college jobs go increasingly to college-educated workers, we should consider those jobs as requiring college education.

If you’re scratching your head wondering if Carnevale is rationalizing credential inflation, then you have no hope of employment in a D.C. think tank. (Maybe you didn’t go to college?) In the fourth appendix, the authors merely counter-argue, “BLS’ educational and training requirement data undercount postsecondary degrees by 22 million in 2008. This implies that 22 million workers are overeducated. The overwhelming consensus in the literature contradicts this.”

Thanks to the most recent publication of the BLS’s employment projections (tables 1.7 and 1.11), I get 15.8 million people with a bachelor’s degree or higher in jobs requiring a high-school education or less. On the bright side, that’s down 100,000 jobs from two years ago. That backlog won’t clear until the mid-22nd century.

It’s true that occupations can change and benefit from productivity advances, and many occupations do not require a single credential to enter them. However, the question GCEW should be asking is what jobs overqualified workers are taking. The answer isn’t too compelling.

Percent BA's in HS & Less Jobs

These twenty occupations account for half of the 12.9 million bachelor’s-degree holders working in high school or less jobs. These occupations dominate among master’s-degree and doctorate holders as well. Maybe some of these folks over 25 are in these jobs temporarily (they’d have to be for many), but at that age it’s pretty implausible that they’re on track for college-premium-magic careers.

Overall, 19.3 million college-and-higher people are qualified or underqualified for their work, and 27.4 million workers are at least somewhat overqualified, which includes PhDs working in bachelor’s jobs.

In contrast to the GCEW’s forecast, the BLS essentially says that 27.7 percent of the jobs to be created by growth and replacement over the next decade will require an associate’s degree or higher. (BA’s are at 20.5 percent.) High-school and less will account for 64.2 percent. Of the 46.5 million jobs that will be created, here’s a table of the top twenty, accounting for 16 million jobs.

OCCUPATION EDUCATION REQUIRED NO. EMPLOYED (2014) (1,000s) NO. EMPLOYED (2024) (1,000s) NEW JOBS (GROWTH + REPLACEMENT) (1,000s)
TOTAL 150,539.9 160,328.8 46,506.9
Retail salespersons No formal educational credential 4,624.9 4,939.1 1,917.2
Cashiers No formal educational credential 3,424.2 3,491.1 1,523.8
Combined food preparation and serving workers, including fast food No formal educational credential 3,159.7 3,503.2 1,364.6
Waiters and waitresses No formal educational credential 2,465.1 2,534.0 1,255.0
Registered nurses Bachelor’s degree 2,751.0 3,190.3 1,088.4
Customer service representatives High school diploma or equivalent 2,581.8 2,834.8 888.7
Laborers and freight, stock, and material movers, hand No formal educational credential 2,441.3 2,566.4 851.7
Office clerks, general High school diploma or equivalent 3,062.5 3,158.2 756.2
Stock clerks and order fillers No formal educational credential 1,878.1 1,971.1 689.0
General and operations managers Bachelor’s degree 2,124.1 2,275.2 688.8
Janitors and cleaners, except maids and housekeeping cleaners No formal educational credential 2,360.6 2,496.9 605.2
Personal care aides No formal educational credential 1,768.4 2,226.5 601.1
Nursing assistants Postsecondary nondegree award 1,492.1 1,754.1 599.0
Home health aides No formal educational credential 913.5 1,261.9 554.8
Accountants and auditors Bachelor’s degree 1,332.7 1,475.1 498.0
Maids and housekeeping cleaners No formal educational credential 1,457.7 1,569.4 459.4
Cooks, restaurant No formal educational credential 1,109.7 1,268.7 452.5
Maintenance and repair workers, general High school diploma or equivalent 1,374.7 1,458.1 443.7
Childcare workers High school diploma or equivalent 1,260.6 1,329.9 441.3
First-line supervisors of retail sales workers High school diploma or equivalent 1,537.8 1,605.4 411.3

Most of these jobs don’t look like they benefit from more education, but hey, maybe Carnevale will reemploy all 15.8 million college grads into jobs that fully utilize their credentials. He only has two years to make it happen.