Diversity Score
Definition
Diversity Score measures portfolio concentration by converting actual obligor exposures into an equivalent number of perfectly uncorrelated, equally-weighted exposures. Moody's calculation groups obligors by industry, applies correlation assumptions (50-100% within industries, 0% across industries), and computes effective portfolio size. A CLO with 250 obligors averaging 0.4% each, concentrated in cyclical industries, might score 180 diversity—meaning it behaves like 180 uncorrelated exposures rather than 250. Minimum diversity scores are rating constraints: AAA typically requires 170+, BBB requires 120+. Diversity score directly impacts loss distribution tail—higher scores reduce probability of catastrophic losses through correlation effects.
Why it matters
Diversity score binds more tightly than simple obligor limits in concentrated portfolios. A manager cannot simply comply with '2% max single obligor' rules—if those obligors cluster in auto, retail, and energy (correlated sectors during recession), the diversity score plummets below rating requirements despite satisfying obligor limits. This forces genuine diversification across industries and business models. During 2020, many CLOs with compliant obligor limits violated diversity scores as defaults concentrated in retail/energy/travel, triggering rating watches. Managers had to execute forced trades—selling performing loans in concentrated sectors, buying loans in uncorrelated sectors—to cure diversity violations, crystallizing losses during dislocated markets.
Common misconceptions
- •Diversity score isn't simply a count of obligors—it's correlation-adjusted. 200 retail obligors score lower than 150 obligors across diverse sectors.
- •Higher obligor count doesn't guarantee higher diversity. A portfolio with 300 obligors in 3 industries might score 150, while 200 obligors across 15 industries scores 175.
- •Diversity requirements don't prevent concentration risk entirely—they limit correlation risk but can't eliminate idiosyncratic industry shocks that affect broad sectors.
Technical details
Diversity score calculation methodology
Industry grouping: Moody's assigns each obligor to one of 33 Moody's industry classifications. Obligors in same industry receive correlation factor (typically 50-100%). Cross-industry correlation assumed 0% (conservative simplification).
Exposure weighting: Each obligor's par amount converted to percentage of total portfolio. Larger exposures contribute more to concentration. 2% obligor position has 5x impact of 0.4% position on diversity calculation.
Aggregation formula: Within each industry, compute sum of squared exposures. Aggregate across industries. Diversity Score = 1 / (sum of squared industry exposures). Formula penalizes concentration nonlinearly—doubling exposure to single obligor reduces diversity by more than 2x.
Example calculation: Industry A (3 obligors: 2%, 1.5%, 1%) = 4.5% industry exposure, squared = 0.2025%. Industry B (2 obligors: 2%, 2%) = 4% exposure, squared = 0.16%. Continue for all industries. If sum of squared = 0.55%, Diversity = 1/0.55% = 182.
Minimum diversity requirements by rating
AAA tranches: Minimum diversity 170-180 depending on deal vintage and WARF. Reflects need for maximum tail risk protection. Cannot achieve AAA rating below this threshold regardless of subordination.
AA tranches: Minimum diversity 140-150. Allows moderately concentrated portfolios but still requires meaningful diversification across industries.
A tranches: Minimum diversity 120-130. Permits more concentration as tranche has thicker subordination cushion absorbing correlation risk.
BBB tranches: Minimum diversity 100-110. Lowest requirement as BBB positioned to absorb correlated losses from sector concentrations. Still binds for highly concentrated strategies.
Dynamic requirements: During stress, rating agencies may increase required diversity for rating maintenance. If macroeconomic correlation increases (recession scenario), agencies demand higher diversity to maintain ratings—creating forced trading pressure.
Manager constraints and portfolio construction
Industry exposure limits: Diversity score creates implicit industry limits beyond explicit concentration covenants. Even without formal caps, managers cannot exceed ~15-18% in any single Moody's industry classification while maintaining required diversity scores.
Obligor size optimization: Managers prefer uniform 0.3-0.5% obligor sizes to maximize diversity. Large positions (1-2%) disproportionately reduce diversity. This creates tension with fundamental credit selection—best ideas can't be sized aggressively without diversity cost.
Correlation arbitrage: Within Moody's classifications, some sub-sectors have lower true correlation. 'Technology' category includes semiconductors, software, and hardware—genuinely different businesses. Managers exploit classification boundaries to maximize diversity while maintaining sector views.
New issue vs secondary trade-offs: Building portfolio in new issue market (limited deal flow) often results in concentrated positions temporarily violating diversity. Managers must actively trade secondary to cure diversity, accepting transaction costs and bid-ask spread losses.
Diversity score gaming and limitations
Gaming through subsidiaries: Single corporate family with multiple borrowing subsidiaries (HoldCo, OpCo, PropCo) technically count as separate obligors but economically correlated. Moody's methodology assigns same industry classification but treats as multiple entities—artificially inflating diversity.
Private equity sponsor concentration: 20 portfolio companies backed by same PE sponsor have correlation through sponsor distress, refinancing timing, and dividend policy. Diversity score doesn't capture sponsor-level correlation—treats as independent entities if in different industries.
Covenant-lite concentration: Modern portfolios have 80%+ covenant-lite exposure. During stress, lack of early warning systems creates correlated deterioration across all holdings—effective correlation higher than historical assumptions used in diversity calculations.
Limitations during crisis: 2008-2009 showed actual correlations reached 80-90% as all cyclical sectors suffered simultaneously. Diversity scores of 180 behaved like 60-80 during systemic stress. Moody's recalibrated assumptions but methodology still underestimates tail correlation.