Executive Summary
|
In July 2020, Lenox Park released our initial White Paper detailing the Investment Industry’s first standardized demographic Organization and Firm assessment methodology: The Lenox Park Diversity Impact Score (“LPI”). To develop the original LPI 1.0, we started by using advanced tools in our RoundTablesTM platform to aggregate detailed survey data around demographics. We then applied statistical analysis, domain expertise, and context from a broad network of institutional stakeholders to establish a robust scoring methodology that could also evolve as the market demand more components, different use case capabilities, and greater filtering and benchmarking. Now we are ready to announce an evolution in our Scoring Methodology, one that (i) allows for greater flexibility around components to create scores more tailored to Users’ needs and (ii) increases the pool of potential organizations that can score themselves and peers regardless of their industry or structure. LPI 2.0 is a modular scoring system that builds on the core componentry of LPI 1.0, yet allows for greater comparison across all industries by providing building blocks that can be assembled to meet any organization’s unique needs. To do this, we consulted with a Client Advisory Board to advise on evolving customer needs, hired leading data science and data visualization expertise to supercharge the data package, and worked with our network of econometricians to help bring the new framework to market. We believe the flexibility and digestibility of our data and its packaging will lead to even better outcomes in managing organizational risk around the relative diversity of teams, for any organization wishing to access these types of analyses, from Asset Management, to Supplier Diversity, to Small Businesses across the globe. |
Goals for LPI Score 2.0For the next evolution of LPI, our objectives have shifted to meet the changing needs of the market and an expanded client base: √ Establish a modular, building block component set that can be customized to suit an organization’s needs – regardless of industry or structure √ Create a universal base component – the LPI Prime score – which accounts for Leadership and Total Staff, categories common to any organization √ Increase components & categories and boost benchmarking analytics across the offering √ Provide backwards compatibility by maintaining ability to analyze and view scores in the same fashion as LPI 1.0 √ Establish a significantly enhanced data visualization package which maximizes data clarity, reporting messages, and real-time customization √ Recognize and architect analytics through the prism of Risk Management |
LPI 2.0 is built around the same philosophy as LPI 1.0, but is now organized as hierarchical roll-ups of each metric (Base Components) into carefully chosen aggregates: either template Module Scores and composite combinations of Module Scores (Composite Scores) proven to be useful to the market, or custom Composite Scores chosen by Users. The original LPI 1.0 metric methodology is still available (now called LPI Private, a Composite Score), but we now offer additional scores to provide different perspectives on each organization, with each being based on a common base Module: LPI Prime.
|
How do we gather and process LPI 2.0 Data? LPI 2.0 data is gathered with a significantly upgraded technology stack and database management through our cloud-based platform, RoundTablesTM, now entering its own 2.0 release. Firms that warehouse their DEI statistics privately in RoundTables are able transmit that information in response to surveys and diligence questionnaires in a consistent format, while tracking period-over-period changes and maintaining complete data permissioning control. Any updates to their data occur in one location, and changes to those metrics are reflected to LPI subscribers in real-time. Upgraded data analytics are powered by advanced tools like Tableau and Amazon Redfish, and our privacy and security match class leading standards, such as SOC 2 Type 2. |
A Module-based Approach
LPI 2.0 is a modular approach to DEI scoring that starts with a universal base building block Module: LPI Prime, which evaluates DEI metrics for senior Leadership and Total Staff, elements that are common to any organization, regardless of its size, corporate and / or ownership structure, or jurisdiction. Thus, LPI Prime will be a universal base Module that allows comparison of any organization versus another given the universal applicability of those categories; further, more detailed score composites can be constructed off of LPI Prime’s base. For example, LPI Prime can be combined with either LPI Ownership to form LPI Private for privately-held organizations (a score that is 100% backwards compatible with LPI 1.0 for comparison purposes and ease of transition for legacy Lenox Park clients and Users), or with LPI Governance (a governance board score) to form LPI Public for publicly-traded organizations. It is also flexible as a base for other future bolt-on modules as the market demands more attributes for assessment.
Each of the components can be pulled apart or combined in multiple ways, all while dynamically recalculating scores and benchmark comparisons on a 0–10 scale. In the future, we plan to add elements of LGBTQ+, Veteran, and Disabled categories (see Figure 11), among others, that will form new Base Components that can be combined with LPI Prime to form custom scores that similarly rescale dynamically.
Additionally, LPI Prime, as a standalone score, opens the market for comparison among a broad range of organizations beyond Asset Management: publicly traded or private, manufacturing or industrial, non-profit or for-profit. It is a universal comparison methodology that is compelling from both a building block perspective and as a universal assessment metric.
A Flexible Hierarchy of Scores
The flexibility of LPI 2.0 comes from the ability to combine a set of Base Components into different levels of modular scores. These levels form a hierarchy, shown in Figure 1.
LPI Private is equivalent to LPI 1.0 total score *
Base Components: The Building Blocks
At the bottom of the hierarchy are the Base Components used to construct all other scores. Every Base Component is measured as a proportion from 0–1. For example, the Race Staff Base Component measures the share of an organization’s workforce who are PoC. Base Components are the combination of an attribute (either Race or Gender) and a Module metric. Metrics are split into three groups: Prime, Ownership, and Governance.
There are currently 14 Base Components across the three Module template category groups.
| Module | Base Component | Base Component Definition |
| Prime | Race Leadership | The share of an organization’s leadership who are PoC |
| Race Staff | The share of an organization’s total workforce who are PoC | |
| Gender Leadership | The share of an organization’s leadership who are non-male | |
| Gender Staff | The share of an organization’s total workforce who are non-male | |
| Ownership | Race Owner Count | The share of an organization’s current owners who are PoC |
| Race Ownership Duration | Of the total years of tenure for an organization’s current owners, the share of those years served by PoC owners | |
| Race Ownership Equity % | The share of a total equity currently held by an organization’s PoC owners | |
| Gender Owner Count | The share of an organization’s current owners who are non-male | |
| Gender Ownership Duration | Of the total years of tenure for an organization’s current owners, the share of those years served by non-male owners | |
| Gender Ownership Equity % | The share of a total equity currently held by an organization’s non-male owners | |
| Governance | Race Share of Board | The share of an organization’s current board who are PoC |
| Race Board Duration | Of the total years of tenure for an organization’s current board, the share of those years served by PoC owners | |
| Gender Share of Board | The share of an organization’s current board who are non-male | |
| Gender Board Duration | Of the total years of tenure for an organization’s current board, the share of those years served by non-male owners |
| Future Module(s) | Base Component | Base Component Definition |
| Plus | Veteran Status | The share of an organization’s total workforce who served in the active military, naval, or air service |
| Disability Status | The share of an organization’s total workforce who have a physical or mental impairment that substantially limits one or more life activity | |
| LQBTQ+ Identification | The share of an organization’s total workforce who identify as LGBTQ+ |
Every score in the hierarchy is computed by scaling and summing various Base Components. Scaling is done so that every score is on a 0–10 scale, with the included Base Components equally weighted. This construction of scores is described in more detail below.
Sub-module Scores: Scoring by Attribute and Metric
The scores above Base Components in the hierarchy are the Sub-module scores. A Sub-module score is a combination of either both the Race and Gender Base Components for a given metric, or all the Base Components in a group with the same Race or Gender attribute.
For example, the Leadership Sub-module score is constructed by combining the Gender Leadership and Race Leadership Base Components. While the Ownership Gender Sub-module score combines the Gender Ownership Count, Gender Ownership Duration, and Gender Ownership Equity Share Base Components.
The Base Components that comprise a given Sub-module Score are scaled so that when they’re summed, the resulting Sub-module Score is on a 0–10 scale. For example, the Gender Leadership and Race Leadership Base Components are each scaled by 5 before summing to compute the Leadership sub-module score. Similarly, the three Base Components in the Ownership Gender sub-module score are each scaled by 3.33 before being summed.
Module Scores: Templates Separating Universal and Structure-specific Components
In the level above the Sub-module Scores are the Module Scores templates: LPI Prime, LPI Ownership, and LPI Governance. These scores combine all the applicable Race and Gender Base Components beneath them. As discussed above, these Module Scores are a key feature of LPI 2.0.
LPI Prime: Leadership and Total Staff Contributions
The LPI Prime score places the spotlight on senior leadership and all staff within an organization. Focusing on identities within these groups yields deeper insights, as research has concluded that the culture, performance, and wellbeing of an organization is dictated by the actions and sentiments of these two groups. Moreover, these categories are universally found in any organization. From those in Asset Management to industries beyond, from private to publicly traded firms, every organization is comprised of elements of both Senior Leadership and a Total Staff (or workforce).
|
Composite Scores: Scoring the Whole Organization At the top of the hierarchy are the Composite Scores. These are holistic scores that reflect the entire organization including certain elements that may not be universally common, but are common to a peer set. For privately-owned organizations, we combine the Base Components of LPI Prime with those of LPI Ownership to compute an LPI Private score. For publicly-owned organizations, we combine the Base Components of LPI Prime with those of LPI Governance to compute an LPI Public score. |
Continued Focus on Ownership and Seamless Transition from 1.0 to 2.0Our decision to highlight ownership characteristics as their own LPI Ownership Module, separate from LPI Prime, was discussed with various industry leaders. The breakout of ownership from the Leadership and Staff components leads to increased benchmarking accuracy within portfolios where both publicly traded and privately held organizations exist.Despite this change, LPI 2.0 scores remain backwards compatible with LPI 1.0 scores on an individual component level, and at a rolled-up level through LPI Private |
Scores as Scaled Base Components
Every Module Score at any level of the hierarchy is some combination of Base Components. Furthermore, every score is on a 0–10 scale and is an equally-weighted sum of its Base Components. Therefore, to compute any Module Score we scale its Base Components so that they will sum to a total between 0 and 10.
For example, LPI Prime, shown in Figure 2, comprises four Base Components: Race Leadership, Race Staff, Gender Leadership, and Gender Staff. Since these Base Components are all proportions from 0–1, we scale them all by 2.5 so that they add up to a 0–10 LPI Prime score.
Figure 2: Base Components of LPI Prime
LPI Ownership, on the other hand, has six Base Components. These six 0–1 components are scaled by 1.66 (or 10 ÷ 6) so that they add up to a 0–10 LPI Ownership score. See Figure 3. It is important to note that these weightings are unique to each rescaled Module or Composite Score and do not carry over when Base Components comprise other score combinations (see An Important Note on LPI Scaled Values).
Figure 3: Base Components of LPI Ownership
LPI Governance, shown in Figure 4, also has four Base Components so, likewise, these are scaled by 2.5 to add up to a 0–10 LPI Governance score.
Figure 4: Base Components of LPI Governance
Figure 5: Base Components of LPI Private
Figure 6: Base Components of LPI Public
If we go up a level to the Composite Scores, we see LPI Private is comprised of 10 Base Components: four Prime components and six Ownership components. Therefore, they can simply be added together without re-scaling to compute a 0–10 score. See Figure 5. LPI Public, on the other hand, has eight Base Components: four Prime components and four Governance components. In this case, the components are scaled by 1.25 to add up to a 0–10 score. See Figure 6.
|
An Important Note on LPI Score Scaled Values The combination of Base Components and their subsequent scaling to 10, with each equally weighted, in a Module or Composite Score, does not result in that Base Component or Module taking on that scaled value as its standalone calculation methodology. For example, common Module Scores may take on different weightings within various Composite Scores. LPI Prime, for example, as a Module within LPI Private (Figure 5) has four Base Components out of a total of 10 when the Ownership Module is added. The Ownership Module comprises six Base Components. As a result, LPI Prime in that case has a scaled value of 40% of the total LPI Private Composite Score. This contrasts with the LPI Public Composite Score (Figure 6) in which LPI Prime comprises 50%, because there are an equal number of Base Components in the LPI Governance Module (four). This does not mean that LPI Prime has a different inherent value itself – in order to make comparisons across LPI Prime scores, they must be regarded as standalone scores out of 10 themselves, with each of Prime’s Base Components being weighted as 25% of the total. |
The same scaling principles apply to the Sub-module Scores lower down the hierarchy. For example, the Staff Sub-module Score (Figure 8) has two components: Race Staff and Gender Staff. In this case, each of those components is scaled by 5 so that they sum to a 0–10 score. The Gender Ownership Sub-module Score (Figure 9), on the other hand, has three Base Components so these are scaled by 3.33 to sum to a 0–10 score.
Figure 7: Base components of the Leadership Sub-module Score
Figure 8: Base components of the Staff Sub-module Score
Figure 9: Base components of the Ownership Gender Sub-module Score
Understanding how components are dynamically scaled so that any score under consideration is on a 0–10 scale is important not only for understanding how scores are computed, but also for understanding the relationship between different Module scores. For example. LPI Private is comprised of the four Base Components of LPI Prime and the six Base Components of LPI Ownership. All these scores are on a 0–10 scale. Therefore, LPI Private = 0.4 x LPI Prime + 0.6 x LPI Ownership.
Understanding scaling is also import for understanding our upgraded visualizations, which will not only provide views of any given level of score, but also the components of that score. Those components will be shown on different scales depending on the score being visualized. For example, when visualizing the LPI Prime score, we can decompose it into its Race and Gender Sub-module Scores. Because each of those scores are composed of an equal number of Base Components, they’ll be shown on a 0–5 scale so that they sum to a 0–10 LPI Prime score.
See Figure 10 and the section on our enhanced visualizations below.
Figure 10: Decomposing LPI Prime into scaled Race and Gender Sub-module Scores
Future Module Score Templates
LPI 2.0 is designed to accommodate future Base Component and metrics additions. Such an example can be seen in Figure 11 where data aggregated from our client base around Veteran, Disability, and LGBTQ+ identity are represented. Similar to other Base Component calculations, each of these categories would be calculated on a standalone basis as a proportion of 0-1, and then be combined to form a new Module Score template we call LPI Plus. LPI Plus is a future release that depends on a critical mass of integral data gathered over time; at present, there is not a statistically defensible volume of data, yet. LPI Plus could be viewed on a standalone basis, as seen in Figure 11, or as a Composite Score when combined with LPI Prime. We look forward to the LPI Plus score’s release in the future, and are tracking these data closely as we continue to conduct new survey efforts.
Figure 11: Base Components of LPI Plus Module Score
Enhanced User Experience & Data Visualization
Upgraded visuals; easier comprehension and ability to analyze data through graphs and charts architected to maximize accessibility and clarity of data and key takeaways
LPI Scores are best understood not as standalone numbers, but within meaningful contexts. One of the most useful contexts to evaluate an organization’s score is in relation to others for a given comparison set (for example, by industry, AUM, staff size, or some combination of all three).
The LPI Score distribution visualization shows how scores are distributed within a benchmark group and indicates how an organization’s LPI Score ranks in that distribution.
An example of this visualization is seen in Figure 12. Each bar in the visualization represents another organization in the comparison set of organizations. The organizations are ordered by LPI Prime score, from lowest to highest. The comparison organizations are chosen so that they are evenly spaced across the set. The horizontal axis is labeled with the percentile of LPI Prime score, so that our sample organization (at the 66th percentile) has a higher score than 66% of all other organizations in the same comparison group.
For more detail, these distribution charts can be broken out into the Gender and Race Sub-module Scores, as seen in Figure 13, or into the metric Sub-module Scores, as seen in Figure 14. As explained above, these charts always scale the overall score to be on a 0–10 scale, and scale sub-components accordingly. For example, the Gender and Race panels in Figure 13 are each on a 0–5 scale.
Figure 12: LPI Prime Distribution and Rank Visualization
Figure 13: Breakdown of LPI Prime Race and Gender distributions
The benchmark group can also be adjusted to provide more relevant comparisons. These visualizations help provide an informative context to understand how an organization's LPI Scores compare to those of their peers.
Figure 14: Breakdown LPI Prime metrics distributions
LPI Evenness: Measuring of Anti-Concentration
Lenox Park uses a novel statistical methodology to measure something that is increasingly important to our client base: measuring not only representation of diverse professionals, but also assessing how the dispersion of races and genders is represented at any given firm; effectively a measure of anti-concentration. We measure this evenness in diversity using the Normalized Herfindahl-Hirschman Index (NHHI) as a supplement to the standard methodology.
Conclusion
LPI 2.0 sets out to achieve universal coverage of organizations across industries and structures, while also providing template scores that are typically used by our client base as well as the flexibility to create custom score combinations that rescale and automatically provide dynamic peer group recalculation according to the components selected. This allows for any organization, regardless of its structure, to compare itself to nearly any other organization, as in the case of LPI Prime, to more narrow selections, such as when comparing LPI Private Scores within common industries, or Asset Classes, and with similar headcounts and other common metrics. This evolution in scoring methodology, including the technology to deliver such results with integral data, will drive better and more informed decision-making for all organizations, worldwide.
Comments
0 comments
Article is closed for comments.