Site icon IMK

Covered California’s Voter Registration Numbers Don’t Add Up

Kevin Knauss
Vote, Registration, California

Covered California reports show thousands of people are requesting to register to vote. But the Secretary of State doesn’t report the same numbers.

Covered California must offer all individuals applying for health insurance the opportunity to register to vote online, have a voter registration card sent to them, or decline the invitation. As a Certified Insurance Agent for Covered California I must offer my assistance to any individual I am helping enroll in health insurance with the voter registration opportunity. When the voter registration program was rolled out several years ago I was skeptical. But from the Covered California records, they have helped over 122,000 people either register to vote or change their voter registration. While the Covered California numbers look great, the Secretary of State’s office shows far fewer voter registration inquiries coming from Covered California.

Covered California Voter Registration Agency

Covered California is designated as a National Voter Registration Agency (NVRA) like the California Department of Motor Vehicles. They must offer each person who comes into contact with the agency the opportunity to register to vote. If a person declines to register, the agency must ask the person to fill out a declination form, which is kept on file at the agency office. As a Certified Insurance Agent, I must take a yearly training course on how to offer people the ability to either register to vote online, through the Secretary of State’s website, or request a voter registration card be sent to them.

I was curious about how many people enrolling in health insurance through Covered California actually register to vote. In August of 2017 I submitted a Public Records Act request to Covered California for the statistics of their voter registration requests. I was supplied with monthly Excel spreadsheets. I decided to review the July 2016 through June 2017 reports to look at a full one year cycle. Each 30 day monthly report actually spans two months. They report the collected data from the 16th of the first month through the 15th of the next month (June 16th – July 15th).

Just like any well designed website, Covered California is able to track a variety of metrics associated with each application for health insurance and how the consumer interacts with the voter registration page. The voter registration interaction is first broken down by user. For example, if a Certified Insurance Agent is performing the enrollment through his or her Covered California account, those voter registration interactions are put under the user Agent. Captured users include –

The CalHEERS (California Health Eligibility Enrollment and Retention System) program (aka the Covered California online application) logs the responses the consumer provides on the voter registration opportunity page.

The Covered California monthly voter registration reports also breaks down the applicants by their eligibility status.

The reports also indicate if the consumer was a new applicant, reporting an address change or renewing their coverage. The renewal only occurs usually during the Open Enrollment Period. Each report has two identical worksheets; one for applications in English and another with the same categories for the Spanish language enrollments.

I created my own spreadsheet comparing reporting periods July – August 2016 through June – July 2017. On average, only about 10% of the people applying for health insurance through Covered California indicate they either want to register to vote online or would like to have a voter registration card sent to them.

Covered California Reported Total Monthly Applications and Voter Registration Requests

Perhaps most surprising to me was which category of user had the most voter registration requests. Even though Certified Insurance Agents account for approximately 44% of all enrollments, they only assist with between 12% and 22% of the total voter registration requests in any given month. Individuals who start and complete their own application with no assistance, (who make up about 40% of all enrollments) accounted for between 71% to 79% of the voter registration requests over the 12 month period.

Percent Covered California Applications Requesting Voter Registration and User

While the percentages for voter registration requests and the percentages based on the types of user were fairly consistent throughout the year, the 2016 October – November report might have erroneous numbers. During this period overall requests for voter registration dipped to 6% of the total applicant, only 35% of Individuals requested registration and 55% of the requests were assisted by agents. The data was either entered incorrectly or the Covered California automatic renewal system that starts during Open Enrollment may have skewed the numbers some how.

In the months of October through December 2016, both Covered California and consumers were renewing their health insurance enrollments for 2017. The applications tallied during these months for renewal ranged between 35% and 42% of the total for voter registration. It’s possible that during the renewal process not all of the data was captured because not all renewals may have had to pass through the voter registration page. The converse is also possible in that the CalHEERS program may have assigned a positive value to the request for voter registration for a consumer even though no such request was made.

But when you look at the months outside of Open Enrollment, or during the Special Enrollment Period, the percentage of applicants requesting voter registration and assistance from agents is fairly uniform. One trend is indisputable and that is people prefer to have the voter registration cards sent them versus registering online. The request for the voter registration information by mail was anywhere from two to eight times larger than the online requests.

There were also relatively few Spanish applications requesting voter registration material. The 2017 January – February report had the most Spanish applications requesting voter registration at 70. That’s pretty small number when there was a total of 19,663 requests for voter registration material out of 196,172 applications in that time period.

Secretary of State Data Comparison

The big bureaucracy that is supposed to hold all the data for the voter registration requests is the Secretary of State. I figured that I could compare Covered California’s numbers to that of the Secretary of State’s data to verify the voter registration requests. Unfortunately, the numbers don’t match. The numbers reported by the Secretary of State are only a fraction of what Covered California is reporting. Who do we believe?

As with Covered California, I submitted a Public Records Act request for voter registration data that was initiated through the Covered California CalHEERS program and hopefully captured by the Secretary of State’s office. I was sent the following reply.

SOS response to PRA request

Thank you for your email. As a part of the National Voter Registration Act (NVRA) and Senate Bill 35 (2011-2012), the Secretary of State’s office works with designated NVRA agencies and county elections officials to track and report the number of voter registrations that are captured on a monthly basis. As you mentioned in your email, your clients are able to link from the Covered California application to the Secretary of State’s online voter registration application or receive a hard copy application in the mail. Each month, these numbers are compiled and posted in several reports that are available on the Secretary of State’s website at http://www.sos.ca.gov/elections/voter-registration/nvra/reports/ .

To view statistics capturing the number of clients that chose to register to vote using the link to the Secretary of State’s online application through the Covered California website, visit http://www.sos.ca.gov/elections/voter-registration/nvra/reports/ and view the report titled: “California Online Voter Registration (COVR) by all NVRA Agencies, April 2014 to Present”.  You will find Covered California statistics under the “NVRA” category.  Please note, that we are currently revising this report and will post the most current data soon.

To view statistics capturing the number of clients that chose to register using a hard copy voter registration application that was mailed to them, please visit http://www.sos.ca.gov/elections/voter-registration/nvra/reports/

and view the report titled: “SB35/NVRA Monthly Reports by County and Agency Office/Site”. This report includes a spreadsheet with a worksheet for each of the 58 counties and lists an entry for “California Health Benefit Exchange” (or similar title) providing the number of hard copy voter registration cards received by that county in the time period identified.

Voter Registration Card Requests

There are numerous reports on the Secretary of State website. The report that has the data for requests for voter registration cards (VRCs) to be mailed to the resident is SB 35/NVRA Monthly Reports by County and Agency Office/Site. It is the only report that specifically lists Covered California under the Public Assistance category as the agency that generated the request. As best I can tell, each county fills out their own spreadsheet, sends it to the Secretary of State, and someone in that office just pastes it into the monthly workbook. Some counties list a hundred different departments that initiate a request for a voter registration card. Some counties only list 10. In other words, the level of detail, and possibly the veracity of the data, varies between California’s 58 counties.

Over the course of the 12 months, there were monthly reports were some counties reported no requests for VRCs from Covered California. But there were some counties that always seemed to report some activity such as Alameda, Orange, San Bernardino, and San Diego. What was odd was that the largest county, Los Angeles, had months where they reported no requests from Covered California. On the other end of the spectrum, San Bernardino almost always reported the highest number of VRCs from Covered California in almost every monthly report.

There were some months when over half of the requests for VRCs from Covered California was attributable to the San Bernardino report.

The highest request for Los Angeles was 135 in August 2016. Most counties reported an average of 0 to 20 requests for VRCs from Covered California per month. Counties with larger relative populations, such as Orange and San Diego, would report between 50 and 100 requests. That is why the numbers from San Bernardino were so off-the-chart. Yolo County even reported receiving over 1,000 requests in October, but in other months they made not have had any requests for the VRC from Covered California.

Secretary of State vs Covered California Online & VRC Requests

The Secretary of State reports are showing 15,804 VRCs requests for the twelve month period while Covered California is showing 106,720. The Secretary of State reports are for the full calendar month. The Covered California reports overlap months, but are still for a full 30 days. Between the San Bernardino excessively high VRC requests and Covered California’s consistently higher monthly totals, I’m not sure either data set can be viewed with a high degree of confidence that the numbers are accurate.

Online Voter Registration

The other type of voter registration request comes from an online source when Covered California consumer selects to register online via the Secretary of State’s website. The Secretary of State link to the online request data is California Online Voter Registration (COVR) by all NVRA Agencies, April 2014 to Present. There are two problems with these reports. First, all online activity for certain agencies like Covered California is grouped under the category NVRA. There is no way to separate out the online requests from Covered California and potentially other agency requesters. Second, the reports only had data through February 2017, when I downloaded the files in late August, so I could not do a full year comparison.

Unlike the VRC numbers, the online figures seem fairly consistent between the Secretary of State and Covered California. Even though we only have eight months of data to compare, and given the difference in reporting time frames (full month vs. overlapping months) the overall totals are within 741 requests.

I am not sure how much information Covered California captures when a consumer or agent opens the Secretary of States website for voter registration. It most likely captures the outbound link click, but may not record if the consumer actually registered to vote or terminated the action before completion. It’s possible that the Covered California numbers are recording instances of the Secretary of State link being opened, but the Secretary of State is actually reporting the number of people who actually registered to vote or completed the action on their website.

It’s not uncommon for consumers to open a shopping cart on an online store but never complete the transaction where they pay for the goods. This is where advance analytical software that tracks the consumer behavior on the website comes into play. Online stores want to know where the consumers drop-off and never complete the transaction. During what step in the process is the consumer most likely to terminate the transaction? Unfortunately, none of the reports from either Covered California or the Secretary of State, and the subsequent comparisons, indicate exactly what is happening in the online voter registration process.

Conclusions

After sifting through all the reports and data, I’m not sure if either agency is accurately reporting voter registration inquiries through Covered California. While the online registration numbers look comparable between the Secretary of State and Covered California, it is unclear if the data is a measurement of a click to open an online registration web page or an actual complete voter registration action. The voter registration card requests from Covered California seem wildly inflated compared to what the counties are reporting to the Secretary of State. However the Secretary of State’s figures may also be corrupt by potentially erroneously high requests posted by San Bernardino County and the lack of requests from huge counties such as Los Angeles.

This much I do know, when it comes to fulfilling Public Records Act requests, Covered California is the best government agency I have dealt with. There PRA department always gets back to me with questions (because my requests tend to be vague) and when humanly possible they get the reports or records to me in a timely basis. I have PRA requests with other agencies that just stall getting me the information or ignore my requests altogether.

Covered California has been a very professional organization when I have dealt with them seeking information. They try to be as transparent as possible, and when they can’t supply me with the information, they cite the code that prohibits the release of the information. If the Covered California VRC request numbers are inflated, it is not by purposeful design. Most likely the system is capturing events or data that really were not triggered by a human, but may be a by-product of some script running in the program.

The data from the Covered California voter registration requests is coming from one system and one agency. The Secretary of State’s data is coming from 58 different counties, which from the looks of the variations in the spreadsheet formats and fonts, may be using their own rules for counting the requests and reporting them. In other words, the data from Covered California has less exposure to errors that would either inflate or deflate the numbers.

If the tabulation of voter registration requests from designated agencies is important, then Covered California and the Secretary of State need to confer and find out why their numbers don’t match. Otherwise, these reports are just another source of wasted tax payer money for bureaucratic work that serves no purpose. Below are the data set spreadsheets I used to analyze and compare the voter registration requests. I made every effort to make sure my numbers were correct, but its possible I made have made some calculation mistakes. Other people more familiar with the reports might be able to point information I missed or was not aware of. I welcome any and all critique of this analysis.

  • Voter Registration Request Data

Voter Registration Request Data

 

 

Exit mobile version