Retirement and 401k Plan Benchmarking
For our database of retirement and 401k plans, we conduct a rigorous ranking and benchmarking process to compare plans against both the entire universe of plans, and against potentially more relevant benchmarks. This gives an overview of the approach and the algorithms used to generate the key insights.
Ranking Algorithm
The ranking algorithm for “Plan Rank” is essentially a stacked percentile ranking of all plans across a set of dimensions, ranked from 1 to 99, with 1 being the worst and 99th being the top percentile. For the example on the right, this plan is in the bottom 6th percentile for participant contributions, but in the 78th percentile for employer contributions.
We also calculate an “overall” plan rank score by averaging across all the dimensions, and then stack ranking the result across all plans in the database. So for example, this plan, on average, has performance rankings that put it right in the middle of all other plans based on the average of all the dimensions.
Benchmarking Algorithm
For benchmarks, we attempt to create a peer set that might allow more relevant comparisons across plans. For the set of dimensions, we compare the target plan with those of the entire dataset (average), those within the same state (state), those in the same industry (industry), and those in the same employee size cluster (size.) Finally, we use an averaging across both employee size and industry group to create a “Peer” average which can be used for more direct comparisons.
