Case study: Large client data audit

  • Print
  • Connect
  • Email
  • Facebook
  • Twitter
  • LinkedIn
  • Google+
By Elizabeth Dey, Tyler Thornton | 17 June 2019

The challenge

Last year Milliman was chosen to provide recordkeeping services for the defined contribution plans of a multi-discipline healthcare system. The client had been with their previous provider for over 70 years and had recently undergone a merger of their plans; four affiliated 401(k) plans and a money purchase plan were merged into a single 401(a) plan and three 403(b) plans merged into a single 403(b) plan.

During implementation meetings, the client expressed frustration that they were still experiencing historic data issues three years post-merger. Complex payroll situations with reporting groups, frequencies, and participant transfers only added to the data validation issues. Additionally, there were disconnects between human resources and the prior recordkeeper. This led to inaccurate deferral election changes, which were often discovered by the participants rather than being caught ahead of time.

The implementation team faced several key issues that needed to be resolved during conversion of the client’s plans to Milliman. First, the team needed to audit the existing participant data so the plan could start on the Milliman platform with a clean slate. Second, the team needed to address the quality of the ongoing data feeds because incorrect post-live data loads would easily negate the pre-conversion data audits. Finally, the team recognized their limited time window prior to the asset conversion; they would need to perform additional ad hoc data audits after go-live to ensure continued data quality.

The solution: Conversion

To ensure the most accurate data possible was loaded to the Milliman system, the implementation team gathered data files from the prior recordkeeper, a census from the client, and historic payroll files for each employee group. The implementation team used these files to compile a full list of participants and then rebuilt each participant’s service history.

Next, the team ranked their confidence in the accuracy of each key data point, awarding high scores to data that matched across all files, and lower scores to data that varied or was difficult to reconcile. After the data had been organized and ranked, the implementation team provided summary files to the client for confirmation and approval.

Although the team had built additional time into the conversion schedule to accommodate this extensive data audit and approval process, the client was not fully able to validate all data prior to the conversion of assets. With the approval of the client, the implementation team loaded the most accurate data available at the time, with the understanding that they would perform further data audits after the conversion.

The solution: Ongoing

During the conversion process, the Milliman team discovered that incorrect participant statuses often occurred because the client terminated participants when the participant simply moved from one participant group to another. These incorrect termination statuses allowed participants to withdraw funds when not eligible for the distributions, which is an operational error that can disqualify a plan.

To mitigate this problem, the Milliman team requested a single, monthly termination date file to load accurate term dates. In addition, the conversion team required approval signatures from the client on withdrawal requests prior to processing, which prevented the processing of additional incorrect withdrawals.

The other major ongoing data issue was incorrect deferral rates. Previously, participants had multiple avenues to change their deferral election, but in practice it often resulted in a disconnect between the rate on file with the recordkeeper and the rate on file with human resources.

The solution was to only allow deferral rate changes via the Milliman system and mirror these values in the client’s system. Milliman strongly encourages all clients to use the Milliman system to track and report all deferral election changes because this allows for consistent setup and reporting of deferral elections and helps mitigate any liability for missed deferral corrections.

The outcome

The client came to Milliman with data and service issues looking for answers. At the time, they were not aware of the extent of the problems or the solutions to them. During the conversion process, the Milliman team consulted with the client to determine known pain points and corrected as many of the data issues as could be identified. Then, Milliman developed procedures to help catch any additional issues that were not uncovered during the conversion.

In the year following the conversion, the Milliman team worked with the client to tackle the immense project of fully auditing all participant data. This included verifying basic information like employment status and service history as well as more complex issues like vesting and deferral elections. After the audits were completed, the client was comfortable with moving to tracking deferral rate changes solely through Milliman and agreed to remove the temporary signature requirements for withdrawals.

The goal of every conversion is to wrap up the project confident that all data is correct and the plan will run as designed. In this case, the Implementation team knew there was the potential that the plan was not going to run properly due to the known data issues; the best available solution was to implement procedures to catch errors before they processed. In this case, anticipating data issues and taking steps to prevent additional errors was the best practice until the root cause was found and resolved.