Data Integrity

Frequently asked questions on how we build, validate, and make decisions about our data

Data integrity is at the core of everything we do

It's guided by our eight core data principles, backed by three pillars: our Global Reporting Principles, developed with leading global pension funds; Data Scorecards, built for institutional investors; and a Seven-Step Validation Process, refined over more than 30 years.

Below, we answer the questions as a part of the ongoing "Opening the Black Box", our ongoing series on the data treatments, assumptions, and methodologies behind our analyses.

Want new answers as we publish them?

Sign up

Opening the Black Box

“Opening the Black Box” pulls back the curtain on how CEM builds, validates, and applies the data behind everything we do – benchmarking, research and analysis. The series walks through everything from data governance and quality processes to taxonomies, methodologies, and the tools that power our analytics, giving subscribers the context to understand their results and ask sharper questions. Have a question you want answered? We’re listening, so see what we cover and how to reach us. Read the series overview blog.

Governance

Philosophically, we believe that nothing is more important than the quality of our data. Our philosophy cascades into our efforts through our eight foundational data principles. Through the application of these principles, we ensure data is complete, comparable and accurate, and that results are timely and transparent. We maintain provenance over the data to keep it confidential and secure. Read more about our data governance principles. 

Your data is governed by CEM’s eight data principles, owned by CEM’s Data Governance Committee. No decision about your data is made unilaterally. Dedicated data owners work together to keep every data decision accountable and aligned with our principles. Read more about how we govern your data. 

We introduced CEM’s Global Reporting Principles (GRP) as the global industry standard for institutional investment and administration reporting. The principles are informed by existing regional regulation and developed with an international Advisory Board. Adherence to these principles is assessed through CEM’s Data Scorecard, giving subscribers gold, silver, or bronze ratings on the completeness and timeliness of their reporting. Read more to see where your reporting stands. 

Our subscribers play three active roles in data governance: providing feedback through one-on-one conversations with your CEM Relationship Manager, participating in our Client Advisory Boards, and reacting to our product roadmap. Learn more about your role in data governance. 

Delivery

We have a proprietary and rigorous seven-step data validation process. It starts with third-party data sourcing (optional), and goes through a multi-step data cleaning process, including automated checks from our rules engine and expert analyst review. You then receive a Data Scorecard for review before we issue your draft results. Finally, your CEM Relationship Manager reviews the results with you before the final report is issued. Walk through the full process.

Most data is sourced directly from subscribers. To support our subscribers, CEM sources, where possible, data from public sources, recordkeepers, custodians and asset managers in order to pre-fill a subscriber’s CEM survey. We also use this data to enhance our insights, in, for example, our performance and risk models. Covered in more depth in our blog here.

We have multiple use cases that require peer grouping. Investment and administration benchmarking peer groups are primarily constructed based on size, because scale drives economies of scale. Talent peer groups are based on operational complexity. Communities peer groups are formed based on operating models and organization types. Custom research and analytics peer groups are built around specific subscriber needs. Where practical we bias towards keeping peer groups stable year-over-year. CEM’s Data Governance Committee holds the final approval on all peer groups to maintain benchmarking integrity. Read how each peer group is built in our full blog here.

Investments

We benchmark seven broad asset classes: public equity (stock), fixed income (bonds), multi-asset strategies (including hedge funds), real assets (including real estate and infrastructure, among others), private equity, private credit, and derivatives/overlays. Within these asset classes, we benchmark more than 70 different sub-asset classes (for example, emerging equity, long bonds, commodities) and nine different implementation styles (for example, internal management, co-investments, fund of funds). Read the detailed blog detailing the asset classes and implementation styles included in CEM’s investment benchmarking.

When providing data, subscribers are asked to report average net asset values at a total fund and asset class and/or mandate level. If a subscriber can only provide year-end holdings, we calculate an average. For new or terminated mandates, averages must be provided for the full year. For private asset limited partnerships, we also ask for the fee basis, the amount fees are based on, which is typically committed capital during the commitment phase and unreturned or invested capital thereafter. We will cover how this plays out for private assets in detail in our blog which will come out on September 30, 2026. To receive the blog directly in your inbox, sign up for the blog series here.

For explicit leverage, NAVs should be reported gross of leverage with offsetting negative cash. For implicit leverage, NAV should be reported net of leverage. CEM adjusts for explicit leverage by constructing benchmarks on levered net asset values, which allows costs to be reflected on the gross positions. Read the full logic behind this approach in our blog, which will be released on October 14, 2026. To receive the blog directly in your inbox, sign up for the blog series here.

CEM shows comparisons for holdings and costs converted to the subscriber’s local currency. On the performance side, returns can be shown either converted to local currency or unconverted. Net value added comparisons, however, reduce currency conversion noise by comparing returns in excess of benchmarks reported in the same currency. Where currency is hedged at a total fund level, returns and benchmarks are asked to be reported on a hedged basis. At an asset class level, CEM asks for returns and benchmarks to be reported unhedged. See how hedged and unhedged returns are treated in our blog releasing on October 28, 2026. To receive the blog directly in your inbox, sign up for the blog series here.

To be comparable, a subscriber’s investment costs must be complete and reported on the same basis as peers. If fees are missing or incomparable, we impute fees following a process overseen by our Data Governance Committee. We include the imputations in total fund costs, but do not include their impact on the subscriber’s cost position relative to peers; we only benchmark the costs actually provided. For transparency, all imputed values are clearly disclosed in the reporting appendices. Read the details of how imputed fees are handled in our blog releasing on November 11, 2026. To receive the blog directly in your inbox, sign up for the series here.

CEM constructs a bespoke, peer-based benchmark cost custom to each subscriber. The benchmark cost adjusts for economies of scale and asset mix differences to assess cost-effectiveness on an apples-to-apples basis. The cost comparisons include all investment costs, internal and external, including costs related to front office teams and governance, operations and support functions. We break down exactly how this works in an upcoming blog available exclusively to CEM subscribers. If you want to discuss more, please contact a CEM team member by emailing info@explorecem.com.

Cost disclosures in annual reports vary by region. Costs reported to CEM often exceed those reported in annual reports, because CEM asks for reporting on a standardized basis across all geographies. CEM does provide a reconciliation service and produces cost disclosure statements for annual reports as an enhancement to our standard report. We’ll walk through how this works in a detailed blog releasing on December 09, 2026. Sign up for the blog series to receive it in your inbox when the blog is released.

Tracking costs and offsets that are netted from returns, such as performance fees and carried interest, helps provide meaningful context to a portfolio’s or manager’s performance. From CEM’s perspective, performance fees and carried interest should be included for a meaningful total cost comparison. That said, CEM does provide the ability to view cost comparisons both including and excluding private asset performance fees. We go deeper on this in a blog releasing on December 23, 2026. Sign up for the blog series to receive it directly in your inbox.

All performance comparisons are done net of investment fees, including external investment management fees and internal costs. If returns are provided gross of fees, we calculate the net returns. Net value added is calculated against reported benchmarks for each asset class and the total fund. For private equity, CEM standardizes the net value added comparisons for peers. The impact of the standardized benchmark on private equity and total fund net value added is fully disclosed in the reporting so subscribers can understand the significance in their results. Read the detailed blog on January 06, 2027. To receive it directly in your inbox, sign up for the blog series here.

CEM has a proprietary adjustment to private equity benchmarks that is based on a mix of lagged, listed small-cap indices. It is used when normalizing PE performance across funds. The standardized private equity benchmark is needed as our research shows that, on average, there is low correlation between private equity returns and self-reported benchmarks, due to reporting lag and the often uninvestable nature of the benchmarks. We will explain the full rationale in our blog, releasing on January 20, 2027, available exclusively to CEM subscribers. If you want to discuss more, please contact a CEM team member by emailing info@explorecem.com.

CEM requires time-weighted returns rather than internal rates of return (IRRs), because IRRs are highly sensitive to the timing and size of cash flows and can be inflated depending on when capital is called or returned. Time-weighted rates of return (TWRRs), by contrast, aren’t impacted by cash flows, making them the more independent measure of manager performance for peer-to-peer comparisons. We will dig into why this matters on February 03, 2027. To receive the blog directly in your inbox, sign up for the blog series here.

Pension Administration

We benchmark costs for 24 different administration activities across the front office and governance and support functions. Service is benchmarked across four member journeys, as applicable: active, inactive, retiring, and retired members. We will break down each journey in detail on February 17, 2027. Sign up for the blog series to get the blog when it’s live.

CEM creates a bespoke, peer-based service score custom to every subscriber. The total fund admin service score is constructed from scores across each of the member journeys the plan has: active members, inactive members, retiring members, and retired members/annuitants. Each member journey score is built on concrete, measurable metrics, like call wait time, degree of personalization, and self-service tools available to members. Over 150 service metrics are assessed across more than 20 activities. We will unpack each driver on March 3, 2027 in a blog available exclusively to CEM subscribers. If you want to discuss more, please contact a CEM team member by emailing info@explorecem.com.

CEM creates a bespoke, peer-based service score custom to every subscriber. The total fund admin service score is constructed from scores across each of the member journeys the plan has: active members, inactive members, retiring members, and retired members/annuitants. We will walk through how it’s built in a blog releasing on March 17, available exclusively to CEM subscribers. If you’d like to discuss more, please contact a CEM team member by emailing info@explorecem.com.

Research & Analytics

CEM’s research and analytics work goes through the same rigorous data quality reviews as our investment and administration benchmarking products. Expert research associates manually review all data, in addition to automated cleaning checks, and new calculation models are approved by CEM’s Data Governance Committee in line with our Data Principles. Internal peer reviews, along with ad-hoc external peer reviews, are carried out before research is published. We will take you inside this process on March 31, 2027. Sign up for the blog series to read it as soon as it’s out.

Tools & Systems

Automated data cleaning is carried out through our rules engine before an expert analyst reviews the data. The rules engine is our proprietary set of data validations, owned by our Data Governance function, and the rules are evaluated annually for improvements and redundancies. As an example, one of our rules checks for private equity limited partnership (LP) fees that fall more than 15% below the database median to confirm if the fees are reported gross of GP offsets. The 15% threshold is based on our historical view of LP deal terms in our database. We will share more about our rules engine on April 14, 2027. Sign up for the blog series to get notified when the blog is live. 

CEM adheres to its AI policy, which is a part of its Data Governance Policy, owned by the Data Governance Committee, which dictates that no client data is ever fed to any public LLMs for research or learning. CEM’s AI-based platform, Query, is hosted in a private environment and is available to subscribers for querying their reports; AI is otherwise only used for internal process efficiencies. We will cover this in full detail on April 28, 2027. Sign up for the blog series to read the blog as soon as it’s out.