Four of the statewide citizen initiated measures (60-63) and all three of the referenda propose to amend the state constitution. Two of the citizen initiated measures (101 and 102) seek to pass new state statutes.
Referenda Pass More Often Than Initiatives
What kind of track record does each kind of measure have with Colorado voters?
Since 1960, Colorado voters have approved 37 out of 120 initiatives (30.8%), and 65 out of 99 referenda (65.7%). The percentages are very similar for the more limited period of 1980 to 2008 (28/93 initiatives and 40/61 referenda), and for 1960-1978 (9/27 initiatives and 25/38 referenda).
Almost two-thirds of the measures that have passed were proposed by the legislature, even though about 55% of the measures considered were citizen initiatives.
Some referred measures from the legislature were statutory, rather than involving constitutional amendments. Eight of the sixty-five referred measures passed in the last forty years were not state constitutional amendments. Some were TABOR authorization to spend measures, and issues of special relevance to the public including a direction to the attorney-general to sue the federal government over immigration related expenses, creation of the lottery, powerball lottery participation, the question of whether to hold a presidential primary, the question of whether Colorado should have a death penalty, and the decision to adopt daylight savings time have been put to the voters by the legislature and passed. Ten proposed statutory referred measures (most TABOR voters) have been defeated by voters. Thus, the legislature's track record on referred statutory measures is a 44% passage rate, while the legislature has had a 70% success rate with referred constitutional amendments.
Nine of the thirty-seven citizen initiated passed by voters in the last forty years have been statutes. They have concerned a renewable energy mandate, background checks at gun shows, hog farm regulation, parental notification for abortion, campaign finance, black bear hunting, voter registration, RTD elections, and financial disclosures by public officials. Twenty-three citizen initiatives proposing statutory changes have failed in that time period, as have sixty citizen initiated constitutional amendments. Thus, for citizen initiatives, the passage rate for statutory changes is 28% and the passage rate for proposed constitutional amendments is 32%. By comparison, in California: "Voters adopted 32% of the initiatives on the ballot in the 1970s, 48% in the 1980s, and 39% in the 1990s. Between 2000 and 2006, the approval rate dipped even lower to 30%." Colorado voters are a little bit more skeptical of citizen initiated ballot measures than Californians.
Of course, that means that the Constitution of the State of Colorado has been amended 85 times in the last 40 years, 28 times by citizen initiative and 57 times in measures referred to voters by the state legislature (a 67% share). Given the voters are being asked to consider seven more state constitutional amendments this year, clearly it is not yet a flawless document notwithstanding all of the efforts that have been made to revise it.
The greater success rate for referrenda is partially a product of the fact that the legislature requires supermajority support in both the state house and state senate to put a measure on the ballot and need only simple majorities to pass ordinary legislation, which insures that only measures with bipartisan support are referred to the voters by the legislature. Many of the legislature proposed constitutional amendments are simple house keeping measures and most are quite timid.
In contrast, citizen initiated ballot issues often involve hot button political disputes, it isn't uncommon in Colorado for citizens to propose measures with little or no hope of passage, in order to arouse partisan enthusiasm and foster public debate. This strategy sometimes works. For example, a number of measures defeated by voters the first time that they were considered, were followed by similar measures that voters approved. Daylight savings time and TABOR are two examples of this happening. Similarly, while a measure to allow Sunday liquor sales failed at the ballot box, a similar measure was later passed by the legislature.
Still, it does somewhat defy conventional wisdom to learn that constitutional amendments are more likely to be adopted by the voters than statutory changes.
The Magic Eight Ball strongly suggests that initiatives will perform below their long term average of 30% (which would translate to two passages this year), and that at least one or two of the three referrenda facing voters will pass, which is about par for the course.
Initiatives Are Not Infrequently Invalidated By Courts
I can't identify any legislature proposed referendums passed by citizens that were later found to be unconstitutional. And, ordinary legislation passed by the legislature is likewise rarely found to be unconstitutional. But, citizen initiatives are not infrequently held to be unconstiutional or to require narrowing interpretations by courts to be constitutional.
At least three of the citizen initiatives that did pass were later found to be unconstitutional in whole or in part. One sought to limit home rule powers of local governments by prohibiting them from enacting measures favorable to gay rights (Amendment 2 by the U.S. Supreme Court in Romer v. Evans), and two limited campaign finance (Amendment 27 (campaign finance limitations) as a result of Citizen's United, and the Amendment 54 (pay to play) by Colorado courts).
Another (Amendment 41) was litigated in Colorado's courts and interpreted in a more narrow way than a plain reading of the measure might suggest.
As I understand the matter, the rate at which Colorado ballot measures are found unconstitutional is actually a rather low figure compared to California, for example, where almost 40% of citizen initiatives passed are held unconstitutional, at least in part.
The Number of Ballot Measures Is Typical This Year
This year's voters are considering half as many statewide ballot measures as they did in 2008, when there were 14 initiative and 4 referrenda on the ballot (2 initiative and 2 referrenda passed, and key provisions of one of the initiatives were later held to be unconstitutional). The only year since Colorado first began considering ballot measures, in 1908, with more measures on the ballot than in 2008, was in 1912, which appears to be the first year that the citizen imitative process was available, so there was apparently at lot of pent up demand.
The number of state ballot issues before voters in this two year election cycle, nine, is actually a little below the forty year average of 11 per two year election cycle. But, the number of statewide ballot measures voters in Colorado will consider in the 2008 and 2010 election cycles combined (i.e. 27) is just a little bit more than the average over the last 40 years for a four year election cycle (i.e. 22).
Never Used Powers
Many of the state's direct democracy provisions appear to have never been used.
Referals of Legislature Passed Legislation
State laws passed by the legislature which do not contain a "safety clause" reciting that they are "necessary for the immediate preservation of the public peace, health or safety, and appropriations" do not take effect for ninety days, during which citizens have the right to insist that the legislation be put to a statewide vote. But, to my knowledge, this provision has never been utilized.
State voters have the power to recall all (or at least almost all) state and local elected officials. Colorado's laws on the subject are typical of states that allow recalls. Recalls of local elected officials aren't uncommon in Colorado. The last successful recall election in Colorado removed four of six targeted city council members in the City of Nunn, Colorado in 2009. In December 2005, District Attorney Colleen Truden, whose multi-county 9th judicial district office that includes Aspen made her technically a state elected official rather than a local elected official was recalled. Arapahoe County Clerk and Recorder Tracy Baker was successfully recalled in February of 2004. But, to my knowledge after an admittedly incomplete search, there has never been a statewide recall election in Colorado, or a successful recall of a state legislator in the state.
Successful statewide recalls of Governors took place in North Dakota in 1921 (the attorney general and the commissioner of agriculture were also removed from office in the same election) and in California in 2003. A 1998 attempt to recall the Governor of Arizona was mooted when the Governor was impeached.
Courts have held that federal elected officials are not subject to recall, notwithstanding state law allowing such recalls, despite a couple of serious attempts to do so. Mayors and other local elected officials are not infrequently the subject of successful recall efforts in states other than Colorado as well.
State legislators have been removed from office by recall in other states in a few instances, but success in these efforts is quite rare:
Prior to Californias' 2003 recall election, the only sucessfull recall In California, where 107 recall efforts were initiated from 1911 to 1994, only four qualified for the ballot. A state senator was recalled in 1913. In 1914, one senator was recalled and another survived a recall attempt. Not until 1994 was another state recall election held, and the senator involved in that attempt - David Roberti - won 59 percent of the vote. In 1995, two Assembly members were recalled. Recall efforts against two Michigan state senators in 1983 were successful - for the first time in that state's history. An Oregon state legislator was recalled in 1988.
Votes Not To Retain Appellate Judges
Also, citizens have the power to vote not to retain judges. Voter have chosen not to retain roughly one percent of trial judges facing retention elections, but have never failed to retain a Colorado Court Of Appeals judge or Colorado Supreme Court justice, the only positions that face voters statewide.
In part, this is because voters have no reason to remove them from office:
Since judicial retention evaluations began in 1988, there have been 953 judges on the ballot. Fifteen got recommendations against retention and 10 got “no opinion.” Another seven were voted out. In the last retention election in 2008, 102 out of 103 judges were retained by voter approval.
Colorado's experience since it adopted the system in 1966 (with the first judicial retention election actually held in 1970) is typical of the ten states that have had judicial retention elections (an earlier iteration of the report is found here). In the period from 1964 to 2006 (a complete data set up to that point since the system was first established in Missouri in 1964), judges have not been retained in only 56 of 6,308 judicial retention elections. Of those, 86.7% of the elections were for "major trial court" seats and 91.1% of the defeated judges were trial court judges. (A study on our practice of making recommendations on judges since 1990 in Colorado can be found here).
Thus, there have only been five appellate court judge non-retentions in the history of the system anywhere in 839 retention elections for appellate judges. At least one of those was in 1984 in Illinois, where Court of Appeals Judge Gordon Maag, who was running for partisan office on the state supreme court was not retained. Another case, in 1996, involved Nebraska Supreme Court Justice David Lanphier whose death penalty and term limits decisions attracted public interest. Most of the non-retentions (51.8%) have taken place in Illinois (which had 31.3% of the election) because it requires a 60% vote to retain a judge, rather than the 50% threshold set in other states. Only one of the twenty-nine judges defeated in Illinois had less than a 50% "do retain" vote. There have been twenty-seven judges have been defeated in retention elections outside Illinois since the first retention election was held in 1964, out of 4,334 retention elections in those states.