11442225495_9d9cc1cbc4_zToday we continue our statistical preview of the Illinois Supreme Court’s upcoming September term with State of lllinois v. American Federation of State, County and Municipal Employees, Council 31. AFSCME poses a question with potentially significant ramifications across a range of cases: are the State’s contractual promises in contracts with its employee unions conditional on the Legislature actually appropriating the necessary funds? Our detailed summary of the underlying facts and lower court rulings in AFSCME is here.

In 2008, the State agreed to a four-year collective bargaining agreement with AFSCME. The CBA provided for small twice-yearly wage increases in 2009, 2010 and 2011. AFSCME ultimately agreed to defer the mandated wage increases in 2009 and 2010 in light of the State’s budget woes. In early 2011, then-Governor Quinn proposed a budget which included sufficient appropriations to fund the required raises for that year. But the legislature ultimately appropriated enough money to finance the raises for employees at only 14 agencies.

AFSCME sought arbitration, arguing that the State had a contractual duty to pay the mandated increases. The State responded that it had no contractual duty to perform unless and until the legislature appropriated the funds. The arbitrator ordered the increases paid.

The State filed a complaint seeking to vacate the award. In the weeks that followed, the legislature made supplemental appropriations, and several agencies experienced enough attrition to pay the required raises to the remaining employees. Employees in six agencies remained without the increases. The trial court held that the contract was invalid unless the legislature appropriated the funds, but ultimately found that the agencies had enough money to partially perform. While that decision was on appeal, the State reached an agreement with the union to withdraw the appeal, but the Attorney General refused to do so. The Appellate Court reversed, rejecting the State’s argument that its contractual promises were contingent on legislative appropriations.

Like Blumenthal, which we previewed yesterday, AFSCME comes to the Supreme Court from the Cook County Circuit Court. Cook County has consistently accounted for a portion of the Court’s civil docket roughly equal to the County’s share of the State’s population, amounting to 45.1% of the docket for the period of 2000-2014.

Since the decision in AFSCME has potential implications for the ongoing battle over employee pensions, let’s take a look at the data for both government and public pension cases. The Court has reviewed 16 conservative Appellate Court decisions – largely decisions in favor of government entities – and 23 liberal decisions. The Court’s grants among public pension cases are more evenly distributed – the Court has granted leave to appeal nine decisions in favor of employers and ten in favor of employees. The Court has reversed only 37.5% of conservative decisions involving government parties, but has reversed 56.5% of liberal decisions favoring people suing the government. The Court has reversed two thirds of public pension cases decided below in favor of employers, and exactly half of the decisions in favor of employees.

The Appellate Court decision reversing the Circuit Court comes from the Second Division of the First District. Division Two has been reversed at a rate somewhat higher than the statewide average in recent years. During most of those years, the Court’s decisions have averaged below three – and sometimes below two – votes to affirm.

We turn next to the Justices’ individual voting records for possible insight on how AFSCME might be resolved. Justice Kilbride has voted to reverse 46.2% of all conservative decisions involving the government from the Appellate Court. Justices Thomas, Karmeier, Theis, Chief Justice Garman and Justice Freeman have each voted to reverse between thirty and forty percent of conservative government law decisions (38.5%, 33.3%, 33.3%, 33.3%, 30.8%). Justice Burke has voted to reverse 28.6% of all conservative decisions.

The Justices’ voting records on liberal Appellate Court decisions (generally, plaintiffs suing the government) are in line with the overall data discussed above. Any reasonable path for the State to four votes for reversal would likely have to include the Chief Justice and Justices Thomas and Karmeier, who have voted to reverse 66.7%, 68.8% and 80% of liberal government law decisions. Justices Kilbride, Theis and Burke have frequently voted to reverse in such cases as well (70.5%, 80% and 62.5%, respectively). Only Justice Freeman has split his votes, voting to reverse 52.9% of the liberal Appellate Court decisions he has participated in involving government law issues.

Meanwhile, only Justice Burke has voted for the employers’ side in more than 50% of the Court’s public pension cases (58.3%). Chief Justice Garman and Justices Freeman, Karmeier and Thomas have all voted for the employer in slightly over forty percent of cases (47.1%, 44.4%, 42.9%, 41.2%). Justices Kilbride and Theis have voted for employers the least often – 29.4% for Justice Kilbride and 25% for Justice Theis.

Join us back here next week as we resume our study of the data from the Court’s oral arguments in civil cases since 2008 – does the Court tend to ask more questions in cases rising from districts of the Appellate Court?

Image courtesy of Flickr by LendingMemo.com (no changes).

1598056760_0c04904449_zToday, we begin a short break from our analysis of the data from the past seven years’ oral arguments to offer a statistical preview of two of the highlights of the upcoming September term. We begin with Blumenthal v. Brewer, set for the September 23 call of the docket, which poses the question of whether Illinois courts can decide property disputes between same-sex domestic partner relationships. Our detailed summary of the facts and underlying court decisions in Blumenthal is here.

Allegedly, the parties in Blumenthal become domestic partners in 1981 or 1982. One partner attained a law degree, the other a medical degree, and as the years passed, the couple allocated work and family responsibilities to care for their three children.

The parties’ relationship ended in 2008. The plaintiff (a physician) filed suit seeking to partition the home the two parties owned. The defendant, by then a judge, counterclaimed for a constructive trust over the residence and the physician party’s net earnings. The plaintiff successfully argued that Illinois law doesn’t recognize property claims between unmarried domestic partners, citing the 1979 decision Hewitt v. Hewitt. The Appellate Court reversed, holding that Illinois law no longer disfavored the parties’ relationship, and the defendant was merely seeking to assert cross claims for relief that are available to anyone else.

Blumenthal comes to the Court from the Cook County Circuit Court. Year after year, the Cook County Circuit Court has accounted for between forty and fifty percent of the Supreme Court’s civil docket. For the entire fifteen year period from 2000 to 2014, forty-five percent of the Court’s civil docket comes from Cook County.

The Court has decided multiple domestic relations cases every year since 2000. In all, the Court has decided fifty-one domestic relations cases – more than eight percent of the Court’s civil docket. Forty-two percent of the cases accepted by the Court are conservative Appellate Court decisions; fifty-seven percent are liberal decisions. The Court has reversed 73.7% of the conservative decisions it has accepted for review, while reversing only fifty percent of the liberal Appellate Court decisions.

Blumenthal comes from Division Five of the First District of the Appellate Court. For most of the period covered by our database, Division Five has had a lower-than-average reversal rate before the Court. The three-year floating reversal rate for Division Five was at or below fifty percent until 2006, and only briefly exceeded the average reversal rate in 2011-2012. By 2014, the three-year floating average reversal rate had dropped back to only 33.3%. In 2014, Division Five’s cases averaged 5.5 votes for affirmance.

The individual Justices’ voting records in domestic relations cases are consistent with the overall data. Since the Court is reviewing a liberal Appellate Court decision, we first consider the Justices’ voting records in such cases. The data suggests that Justices Burke, Karmeier, Freeman are the most likely to be receptive to reversal in Blumenthal; Justice Burke has voted to reverse 53.3% of liberal Appellate Court decisions, Justice Karmeier 47.4% and Justice Freeman 44.8%. The remaining Justices – Chief Justice Garman and Justices Kilbride, Thomas and Theis – have all voted to reverse in forty-two percent or fewer of the liberal Appellate Court decisions they’ve participated in. (Garman, Kilbride and Thomas 42.3%, Theis 40%).

Not surprisingly given the Court-wide data we reviewed above, the Justices’ have been significantly more willing to vote to reverse conservative Appellate Court decisions. Justices Theis, Burke and Kilbride have all voted to reverse more than eighty percent of the conservative Appellate Court decisions they’ve voted on (83.3%, 81.8%, 81.3%). Chief Justice Garman has voted to reverse in three-quarters of conservative Appellate Court decisions, and Justices Karmeier and Freeman only slightly less (72.7%, 70.6%). Only Justice Thomas has voted to reverse less than seventy percent of the conservative decisions he’s voted on, at 66.7%.

Given the number of years which have passed since Hewitt, the Court’s numbers in deciding domestic relations cases and the individual Justices’ voting records, the appellee would seem to have a clearer potential road to a four-vote majority than the appellants do before the Supreme Court.

Tomorrow, we’ll address another case from the Court’s September 23rd call of the docket: State of Illinois v. American Federation of State, County and Municipal Employees, Council 31.

Image courtesy of Flickr by Mikael Altemark (no changes).

15189303264_acd2284885_zLast week, we looked at how often the first question was asked by the author of the majority opinion. Today we broaden the question – how likely is it that the first questioner to both sides is writing something – the majority, a special concurrence or a dissent?

As we discussed last week, if there were no relationship between writing and questioning, we would expect to see a top line of 14.3%. Compared to that baseline, the evidence suggests that it is comparatively likely that the first question is asked by a Justice writing an opinion; with the exception of 2010, nearly one third of first questioners wind up writing.

The data for writers of majority opinions is similar. With the exception of 2008 and 2010, authors of majority opinions ask the first question of the winning side nearly one third of the time. With the exception of 2010, majority opinion authors are even more likely to ask the first question of the losing side – exactly one third of the time in 2008 and 2011, and nearly forty percent in 2012.

In contrast, there is no apparent relationship between Justices writing a special concurrence and the likelihood that the Justice asks the first question. The Court writes relatively few special concurrences each year, so year-by-year data is uninformative, but for the entire seven year period, only 6.82% of writers of special concurrences were the first questioner of one side or the other – less than a random result.

Dissents, however, do somewhat increase the likelihood that a Justice will ask the first question. Remember that we concluded a few weeks ago that the Justices are likely to average more questions to the side they disagree with. A dissenter disagrees with the winning side, and sure enough, between one quarter and one third of all dissenters between 2008 and 2014 (with the sole exception of 2009) asked the winning side the first question. On the other hand, dissenters were significantly less likely to ask the first question of the losing side (the side they agreed with). In four of the seven years during our study period, a random fraction or below of the dissenters led off questioning of the losing side.

Table 57 A

Next week, we’ll take a one-week break from our data analytic work for a statistical preview of leading cases from the Court’s upcoming September term.

Image courtesy of Flickr by Thomas Galvez (no changes).

5275403364_6396c50053_zLast week, we continued our look at the question patterns in civil cases at the Illinois Supreme Court by analyzing whether there’s a higher-than-random likelihood that the Justice asking the first question to either side is writing the majority opinion.

Today, we continue our analysis with a somewhat different issue: do the Justices tend to ask more questions in cases involving certain areas of the law? Keep in mind, we’re talking total questions here. While we determined earlier that losing parties are likely to get more questions than winners, this doesn’t address the question of which cases will get 10-15 questions per side and which ones might get 30-35 per side.

It’s easy enough to imagine a reason why the area of the law might influence the level of questions from the Court; the Justices might well tend to ask more questions, all else being equal, about areas they seldom hear and are therefore somewhat less familiar with in comparison to the mainstays of their docket, such as tort and civil procedure.

Let’s see whether the data offers any support to our conjecture.

Below are the average questions asked to each side for all civil cases argued between 2008 and 2014, classified by the principal area of law involved in each case. Notice that far from sparking a significantly reduced number of questions, tort and civil procedure cases bring something close to the overall average number of questions. In tort, appellants were asked an average of 14.82 questions, while appellees were asked 14.58 questions. Questioning in civil procedure cases was a bit lighter – 12.56 to appellants, 11.33 to appellees. Questioning in constitutional law, insurance, domestic relations and government law cases – which were all about equally common on the Court’s docket – was higher, but only slightly so.

So the evidence is at best equivocal as to whether the area of law strongly influences the level of questioning. Although contract and election law cases were comparatively rare on the Court’s docket during these years – two contract cases and seven election law – questioning was relatively heavy. Appellants were heavily questioned in environmental, employment, trusts and estates, public pensions and judicial cases, all of which are relatively uncommon on the Court’s docket. But on the other hand, administrative law, arbitration, commercial and surety law cases – also somewhat rare on the Court’s docket – drew comparatively fewer questions than other areas did.

Table 55With the exception of torts, most of these areas of law appear on the Court’s docket five times or fewer per year. For that reason, it’s not particularly informative to report year-by-year averages; when an area is only heard three or four times, a minor variation from one argument to the next will have an outsized influence on the mean.

However, tort law represents about a quarter of the Court’s docket year in and year out, so we report the year-by-year data below. Although the two sides’ level of questions were almost identical measured over the entire seven years, we see that the numbers varied fairly significantly from year to year. Average questions to appellants dropped somewhat in 2012 and 2013 – a phenomenon we’ve seen elsewhere in the data – but during the same years, questions to appellees actually increased. Tort cases were especially heavy on the docket in 2011 and 2012, but there’s no apparent effect on the average questions.

Table 56Tomorrow, we’ll return to our focus on the connection between the first question and writing by broadening last week’s question – are Justices who are writing something, whether a majority, special concurrence or dissent, statistically more likely to ask the first question?

Image courtesy of Flickr by Ryan Milani (no changes).

7095563439_8130f11eea_zYesterday, we reviewed the record-holders from the last seven years of oral arguments at the Illinois Supreme Court: most and least total questions, most and least questions for appellants and appellees, and so on.

Today, we turn to an entirely different question: what’s the likelihood that the first question came from the author of the majority opinion?

As a baseline, keep in mind that the likelihood of an entirely random Justice asking the first question is 14.29% – one in seven. Of course, that assumes that each Justice is equally likely to ask the first question. As we’ll begin discussing in a few weeks, that’s clearly not true.

Still, for many of the possible combinations over the years, the likelihood of the first question coming from the author of the majority opinion is significantly over the random selection number. For example, in 2011, more than four of every ten first questions to the losing appellee came from the author of the majority opinion. In 2012, approximately one in every three first questions to each party came from the author of the majority. In 2014, first questions for about one third of the winning parties came from the Justice who later wrote for the Court. On the other hand, only a few years earlier in 2010, only about one in every ten first questions came from the author of the majority opinion.

Table 54Next week we’ll expand our analysis to a related question – how likely is it that the first question came from a Justice who is writing something – the majority opinion, a special concurrence or a dissent?

Image courtesy of Flickr by Zeevveez (no changes).

8671901426_a940c6940f_zToday I thought we’d take a bit of a break from all the number-crunching to do something a little different: talk about the records, the “mosts” and “leasts,” from the 233 oral arguments at the Illinois Supreme Court we reviewed and charted?

The ground rules are the same as they’ve been throughout our investigation of the Court’s arguments – civil cases only, less habeas corpus cases, attorney discipline and juvenile cases. And only cases where at least one counsel appeared for both sides are eligible.

So here we go:

Most total questions, both sides:

  • 81 – In re Sophia G.L., May 22, 2008. Whether Illinois should register an Indiana child-custody determination awarding a child’s grandparents temporary custody;

Least total questions, both sides:

  • 8 – Russell v. SNFA, April 18, 2013. Whether maker of custom tail-roter bearing had minimum contacts with Illinois so as to bring the action within long-arm statute where, although defendant did not have direct U.S. customers, defendant had multiple sales of its products in Illinois and regular business dealings with an Illinois company;
  • DeHart v. DeHart, March 21, 2013. Whether Illinois should recognize claims for contract to adopt and equitable adoption;

Biggest margin between Questions to Losing Party and Questions to Winning Party:

  • 45 – In re Marriage of O’Brien – Whether a petition to substitute a judge for cause after a substantive ruling is evaluated by the statutory standard of actual prejudice or by the standard of appearance of impropriety;

Biggest margin between Questions to Winning Party and Questions to Losing Party:

  • 33 – In re Sophia G.L.;

Most Questions to Appellant:

  • 57 – In re Sophia G.L.;

Least Questions to Appellant:

  • 0 – Taylor v. Pekin Insurance Co., November 20, 2008 – Whether the requirement under the Workers’ Compensation Act that an employer pay the employee’s attorney 25% of the lien recovered by the employer applies where the employee has received benefits under the uninsured motorist provision of his employer’s automobile liability policy;

Most Questions to Appellee:

  • 43 – People ex rel. Madigan v. Illinois Commerce Commission, November 20, 2008 – Whether the Commerce Commission’s regulations require an electronic filing to be transmitted before the Commission’s office closes in order to be timely filed and vest the Appellate Court with jurisdiction to hear an appeal;

Least Questions to Appellee:

  • 0 – Ferguson v. Patton, March 21, 2013 – Whether municipal inspector general to whom city ordinance has given authority to issue subpoenas lacks authority to retain private counsel and bring an action against the City’s law department seeking to enforce the subpoena.

Tomorrow we’ll begin looking at the likelihood that the Court’s first question comes from the Justice writing the majority opinion.

Image courtesy of Flickr by Martin Pettitt (no changes).

9647972522_eb1f0c3ca7_zAs we’ve been discussing for the past few weeks, the academic literature has consistently concluded that the losing party, on average, gets more questions in oral argument than the winning party. Yesterday, we addressed whether the result – affirmance or reversal – or the vote margin had the greatest impact on the difference in questions to one side and the other.

We wound up with a surprising result: losing parties only consistently average more questions when the Court ultimately affirms – not in the case of reversals. We’ll be looking more deeply at possible explanations for that result in the next several weeks, but one possible theory is that affirmances are, in a sense, a reversal of fortune at the Supreme Court. The appellant “wins” by persuading the Court to grant the PLA, but then loses on the merits when the Court affirms. Perhaps the Court tends to question appellants in an affirmance more heavily in order to make sure that there is in fact no basis for reversal in an opinion which the Court agreed to review.

Today, we consider a somewhat different question: just how long are the odds against winning for a party that gets more questions?

The table below plots the percentage of unanimous and non-unanimous civil decisions in which the winning party received more questions from the Court, year-by-year from 2008 through 2014. There is no consistent difference between unanimous and non-unanimous decisions. Parties getting the most questions were more likely to win non-unanimous decisions in 2008, 2010 and 2014. The more heavily questioned side won only 16% of all unanimous decisions in 2008 and 2010, 18.75% of non-unanimous decisions in 2012, one third of non-unanimous decisions in 2008 and 2014 and one-third of unanimous decisions in 2011 and 2014. In 2009, the more heavily questioned party won nearly forty percent of all decisions. For the entire seven year period, the more heavily questioned party won 27.27% of all non-unanimous decisions, and only slightly more – 31.88% – of all unanimous decisions.

Table 53

Next week, we consider the extremes of the Court’s oral arguments – the most questions in a single argument, the most questions asked to a single party, and so on. Then, we turn to the question of whether the first question is statistically likely to be from the author of the majority opinion.

Image courtesy of Flickr by Raymond Bryson (no changes).

202872717_a8a4799419_zAs we discussed a few weeks ago, the academic literature on question patterns in appellate oral argument has consistently shown that the losing party tends to get more questions. We’ve shown that that result also holds for the Illinois Supreme Court’s civil caseload, considered as a whole, between 2008 and 2014.

Today, we continue our investigation by looking at the data in a slightly different way: does the margin between questions to one side and the other depend more closely on the result (whether the case is affirmed or reversed) or the voting margin?

The expected relationship, with losing parties receiving more questions on average, often does not hold in cases where the Court reverses. Winners received more questions in unanimous or one-dissenter reversals in 2009, 2010, 2011 and 2013. Winners received more questions in 2 to 3-dissenter reversals in each year since 2008. Nor is there any consistent relationship between the margin on unanimous and one-dissenter reversals, as opposed to two- and three-dissenter reversals. Interestingly, there is some indication that total questions are declining at least slightly in reversals over the entire seven-year period:

Table 51

On the other hand, losing parties do tend to get the most questions when the Court affirms. Losing parties averaged more questions in unanimous and one-dissenter affirmances in every year between 2008 and 2014. Similarly, losing parties averaged more questions in two-and three-dissenter affirmances in each year between 2009 and 2013. The difference in questions to losing and winning parties have been relatively constant each year during the period:

Table 52

Tomorrow, we continue our investigation by looking at the cases in which the winning party received the most questions from the Court.

Image courtesy of Flickr by Duncan Hull (no changes).

4860959981_209d1142f8_zOver the past few weeks, we’ve been looking at the data on the patterns in the Illinois Supreme Court’s questions at oral argument to see what we could conclude about the likely future progress of a case. Yesterday, we looked at whether particularly active questioning from the Court suggested that the case might be under submission from oral argument to opinion for a longer time.

Today, we address a similar question: does active questioning suggest that the Court’s majority opinion will be longer?

We can imagine scenarios at oral argument in which the answer might be yes. For example, a particularly difficult case, or one with a large number of issues, might lead both to heavy questioning from the Court and to a longer-than-average majority opinion.

As we noted yesterday, total questions from the Court is a widely dispersed variable, with a mean value of 31.155 questions, and a spread ranging from a low of 1 to a high of 81 questions. The standard deviation of the total questions variable is 15.136.

Because the Illinois Supreme Court tends to write concise opinions, the data on majority opinions shows considerably less spread. The mean value for pages in the majority opinion is 13.422. Opinions have ranged from one-page per curiams to a high (in civil cases) of only 41 pages. The standard deviation for pages in the majority opinion is 6.259 – less than half the spread in total questions.

The scatter plot seems to show a slight positive relationship between total questions and opinion length:

Table 50The correlation between total questions and length of the majority opinion confirms the plot. The correlation is 0.08997. So unlike our findings with time under submission yesterday, the sign of the relationship is positive. Although the two variables are more strongly correlated that either of our previous posts (questions margin and losing margin in votes; and total questions and time under submission), total questions asked at oral argument is at best only a slight indicator that the majority opinion might be longer.

Image courtesy of Flickr by Michael Graf (no changes).

8880371801_ae981b8b19_zLast week, we investigated several questions about the data on oral argument question patterns at the Illinois Supreme Court: (1) is the party who gets more questions statistically more likely to lose; (2) do the odds against winning get longer as the difference in the number of questions asked one side and the other increases; and (3) is lopsided questioning a reliable predictor that the ultimate decision is likely to be unanimous?

This week, we continue our analysis of what the Court’s question patterns have indicated over the years about the ultimate decision. Today – do more questions suggest that the case will be under submission between oral argument and opinion for longer?

One can easily imagine scenarios which suggest that the answer might be yes. For example, it wouldn’t be surprising to see a case presenting an unusually large number of issues, or particularly difficult issues, both leading to a particularly “hot” bench, and a somewhat longer lag time until publication of the opinion. Appellate judges and justices often say that oral argument does, from time to time, cause them to view a case in a different light. To the extent that the several members of the Court get responses from counsel they weren’t expecting, one can imagine the Court might revisit issues, slowing down the ultimate opinion.

We begin with the basic descriptive statistics. Both total questions per oral argument and total days under submission are widely dispersed variables. The mean number of questions asked per argument from 2008-2014 is 31.155. Observations range from a low of 1 to a high of 81. The standard deviation of questions per argument is 15.136.[1]

The average days under submission for civil cases between 2008 and 2014 (both unanimous and non-unanimous decisions) is 134.595. Observations range from a low of 15 days to a high of 715 days. Consistent with that, the standard deviation is 95.145.

Based on these results, we would expect the scatter plot to be widely spread out, and sure enough, it is. In the table, we plot total days under submission against total questions asked by the panel. If the two variables were strongly related, with more questions tending to be in cases which took longer to write opinions, we would expect the scatter to follow a fan anchored at the lower left hand point, with more total questions meaning more days. But in fact, the vast majority of the Court’s cases are handed down between 100 and 200 days after oral argument, with no clear relationship between total questions asked and time under submission:

Table 49The correlation data confirms the implications of this plot. The correlation between days under submission and total questions is -0.0197. What this means is that particularly active questioning from the Court tells us little about how long a case is likely to be under submission, and to the extent there is any relationship at all between the variables, it is the reverse of what we were expecting – heavy questioning suggests a quicker decision.

Tomorrow, we’ll address another question: does heavy questioning from the Court suggest a longer majority opinion?

Image courtesy of Flickr by LinkHumans (no changes).

 

——————————————-

[1]               Standard deviation is the square root of a data set’s variance. Variance measures the degree to which the variables are dispersed from the mean. Thus, a variance of zero indicates that all observations are equal to the mean, while a large variance suggests that observations vary widely.