Dynamic Shariah Compliance: An Eight-Quarter Trajectory Study of the S&P 500 (Q1 2024 – Q4 2025)

Halal Terminal Research — Piece 1 of 3

---

⚠️ IMPORTANT LEGAL NOTICE — Click to expand

This document is general educational research. It is provided for informational purposes only and does not constitute investment advice, portfolio management, a solicitation, or any inducement to transact in any financial instrument.

No personalization. The analysis does not take into account any individual's financial situation, objectives, or risk profile.

Not a fatwa. Shariah compliance trajectories are computed from public financial disclosures under a specific methodology disclosed in §3. Results are not a religious ruling and should not substitute for guidance from a qualified Islamic scholar or Shariah advisor.

Data and scope. Figures are derived from SEC EDGAR XBRL filings and public market data as of the quarter-ends listed in §3. Historical compliance trajectories are descriptive and do not imply future compliance status of the same names.

Regulatory positioning. This text is intended to qualify as non-advisory research / educational analysis under Swiss regulatory expectations (FinSA / FINMA), with a focus on methodology, historical evidence, and limitations.

For complete disclaimers, see Legal Disclaimers.

---

1. Executive Summary

The global Islamic finance industry held USD 5.98 trillion in assets at end-2024 and a projected USD 9.7 trillion by 2029, implying a roughly 10% compound annual growth rate.¹ Almost all of that growth depends on a binary screen: compliant or non-compliant, attached to a name at a single point in time, refreshed annually or quarterly by an index provider. The screen is treated as a property of the company. It is not. It is a property of the company at a date, and the date matters.

This study walks the S&P 500 through eight consecutive quarter-ends (Q1 2024 through Q4 2025), applying AAOIFI Shariah Standard No. 21 under the production-aligned reading — interest-bearing debt below 30% of total assets, cash and cash equivalents below 30% of total assets, non-permissible income below 5% of revenue, and a permissible core business — to each name, each quarter. The continuous universe is 472 constituents that stayed in the index for the full window. The methodology matches the production compliance engine that powers Halal Terminal's customer-facing verdicts; §3 documents the implementation precisely, including the choice of cash numerator. A stricter academic reading of AAOIFI's "interest-bearing investments and securities" (including the long-term marketable-securities portfolio) would produce a different — tighter — universe; we surface the choice rather than bury it. NPI is also a proxy in this study — interest income / TTM revenue rather than a full 10-K segment decomposition. This is the production-engine input at universe scale and is sufficient for the headline findings, but the limitation matters at edge cases (BNPL, insurance float, fintech adjacencies) and is repeated in §3 and §8.

Three headline findings.

First, only 142 of 472 names (30.1%) were compliant in all eight quarters. A further 218 (46.2%) were non-compliant in all eight. The remaining 112 names split between Oscillating (40), Near-Improving (11), Improving (9), Near-Deteriorating (8), Deteriorating (6), and 38 Unclassified names where XBRL coverage was incomplete. The aggregate non-compliance rate under AAOIFI is materially higher than the rate any 33.33%/market-cap screen would produce — a structural feature of AAOIFI's stricter threshold and asset-based denominator, not a contingent feature of this window.

Second, the AAOIFI-vs-DJIM gap is mostly capital-structure, not cash. Under DJIM (33.33% threshold on trailing 24-month MC), the same universe shows roughly 43% SC. AAOIFI's 30%/assets combination catches an additional ~13 percentage points of names — and the names it catches are debt-financed acquirers, capital-intensive industrials, and a small number of cash-rich operators where the asset base is concentrated enough that cash equivalents alone exceed 30%. Broadcom (AVGO), Home Depot (HD), AbbVie (ABBV), and Boeing (BA) are stable-non-compliant under AAOIFI for debt-ratio reasons; they would pass — or sit at the boundary of — a 33.33% DJIM screen. This is the load-bearing methodological difference.

Third, debt is the dominant transition driver. Of 69 compliant-to-non-compliant transitions, 57 were debt-only, 7 cash-only, and 5 combined debt+cash. Of 77 non-compliant-to-compliant transitions, 58 were debt-driven, 16 cash-driven, and 3 combined. Debt accounts for 123 of 146 driver flags (84%); cash accounts for 26 (18%). Non-permissible income and business-activity changes drive essentially none over this window. Cash plays a meaningful but second-order role.

Sector tilt is severe and persistent. Financials are 68 of 68 stable non-compliant — disqualified by AAOIFI's business-permissibility screen regardless of any ratio. Utilities are 27 of 30 stable non-compliant on capital structure. At the opposite pole, Information Technology posts a 52.4% stable-compliant rate (33 of 63), Industrials 47.3% (35 of 74), and Energy 45.0% (9 of 20). Real Estate sits low at 9.7% SC — the REIT capital-structure pattern. The aggregate sector picture is not a contingent feature of this window — it is structural.

Implications follow. A Shariah ETF that screens annually misses transitions inside the year: 69 names cross from C to N inside this universe over eight quarters, and the names doing the crossing are concentrated where active balance-sheet decisions meet the 30% threshold — debt-funded M&A, capacity build, share buybacks against shrinking cash. Purification calculations that assume a name's compliance status over a holding period are mis-specified when that name spent part of the period non-compliant. And screening platforms that surface only a current binary verdict are hiding the variable their users most need to plan around.

The mean S&P 500 name was compliant for 3.30 of the 8 quarters. The median was 3. Compliance is not a constant.

2. Why Compliance Is Dynamic

Two mechanisms drive verdict changes under AAOIFI.

The first is balance-sheet motion. AAOIFI's ratios divide interest-bearing debt and cash equivalents by total assets. Companies move all three numerators as a normal consequence of running the business. They issue debt to fund acquisitions; they retire debt with free cash flow; they let cash accumulate ahead of share-repurchase windows; they draw cash down to fund capacity. Each of these moves changes the numerator without necessarily changing the asset base by the same proportion, and the ratio crosses thresholds.

The clearest illustration in the dataset is Welltower (WELL), one of the larger US-listed healthcare REITs. Welltower entered the window non-compliant on the debt ratio: debt/assets sat at 31.9% (Q1 2024) and 30.4–32.2% across all of 2024. Total assets grew steadily from USD 45 billion (Q1 2024) to USD 67 billion (Q4 2025) as the senior-housing portfolio expanded; absolute debt grew too, from USD 14.2 billion to USD 19.2 billion, but more slowly. The debt ratio fell to 28.5% at Q4 2025 and the verdict flipped from N to C at Q1 2025. This is not a denominator effect — total assets grew because Welltower built the balance sheet through equity-funded acquisitions. The verdict change is recording a real shift in capital structure, slowly, in the direction of compliance.

The second is the asset side. As companies grow, the asset base grows too — and the asset base is what AAOIFI's ratios divide by. A capex cycle that builds plant and inventory expands the denominator. A divestiture compresses it. An acquisition consolidates the target's assets and the parent's funding of the deal onto the consolidated balance sheet. None of these movements are price-driven; they are operating decisions, and they push the verdict around.

Both mechanisms compound. A capex cycle funded with new debt may grow the asset base in lockstep with debt — the ratio holds. A cash buildup ahead of a buyback can push cash/assets through 30% while every other metric is unchanged. AAOIFI's denominator is sticky enough that single-quarter blips are rare; trajectory under AAOIFI tends to record durable balance-sheet shifts, not market-driven noise. The point-in-time problem is not a curiosity. It is a structural feature of any ratio screen with a balance-sheet denominator and a moving balance sheet.

3. Methodology

Universe. S&P 500 constituents as of 31 December 2025, intersected with constituents at each of the seven prior calendar quarter-ends back to 31 March 2024. The intersection contains 472 names. The 31 constituents that joined or left the index during the window are out of scope; the membership reconstruction uses the public S&P 500 component-change record. The continuity filter eliminates survivorship questions about index addition/removal effects on compliance.

Reading. This study applies the production-aligned AAOIFI reading: AAOIFI Standard 21 thresholds and denominator, with the cash numerator restricted to cash and cash equivalents (matching the production engine described below). A strict-academic AAOIFI reading — including long-term marketable securities in the cash numerator — would tighten the universe and is documented under "Cash numerator — disclosed" below. We label the chosen reading explicitly because Shariah boards and auditors may have different expectations on this point.

Thresholds. AAOIFI Shariah Standard No. 21 sets the thresholds applied here:

• →Interest-bearing debt / total assets < 30%
• →Cash and cash equivalents / total assets < 30%
• →Non-permissible income / TTM revenue < 5%
• →Core business permissible (qualitative)

The 30% threshold and the Total Assets denominator are both from AAOIFI's text. The major index providers' screens differ: the S&P Dow Jones Islamic Market Index applies a 33.33% threshold on a trailing 24-month average market-cap denominator;² MSCI's main Islamic Index Series applies 33.33% on Total Assets, while its parallel M-Series uses a 36-month market-cap denominator;³ FTSE Yasaar applies 33.33% on Total Assets with a 50% accounts-receivable cap.⁴ This study applies AAOIFI's stricter threshold and denominator combination throughout. Piece 2 of this series will run the same universe under each index provider's specific methodology and quantify the names that flip verdict between the screens.

Cash numerator — disclosed. AAOIFI's text references "cash and interest-bearing investments and securities" in the cash ratio, with no maturity cap. Some academic readings include the full marketable-securities portfolio (cash equivalents + short-term + long-term investments). Production Shariah screening engines — including the one powering Halal Terminal's customer-facing verdicts — use the narrower cash and cash equivalents definition, taking the cash-flow-statement field directly and excluding the marketable-securities portfolio. This study follows production practice for two reasons: (1) it matches the verdicts presented to Halal Terminal customers, so the research piece does not contradict the product; (2) the narrower definition isolates "liquidity that is unambiguously interest-bearing" rather than relying on issuer-specific disclosures about the composition of long-dated marketable securities. A reader applying the broader definition would find a tighter universe — Apple, Nvidia, Oracle, and other cash-rich tech names whose long-term investment portfolios exceed 30% of assets would flip from compliant to non-compliant. Whether that broader reading is the correct AAOIFI interpretation is itself a Shariah-governance question; we surface the choice rather than bury it.

Standard 59. AAOIFI Shariah Standard No. 59 (Sale of Debt) is widely read as having superseded the historic illiquid-assets / accounts-receivable cap that some pre-2018 screens applied alongside the leverage and cash ratios.⁵ This study omits that cap accordingly: ratios are computed without an accounts-receivable / total-assets gate. Methodology currency matters; published screening papers from before 2018 that include the illiquid-assets test should be read against the post-S59 framing.

Financial inputs come from SEC EDGAR XBRL companyfacts extractions, cached locally and mapped to calendar quarter-ends with a 95-day reporting-period grace. Debt extraction prioritises LongTermDebt and its capital-lease-inclusive variants, falling back to the sum of LongTermDebtNoncurrent and LongTermDebtCurrent when needed. Total assets uses Assets directly. Cash uses CashAndCashEquivalentsAtCarryingValue directly — matching the production engine's first-field-wins priority list (which selects "Cash And Cash Equivalents" ahead of any field that bundles investments).

Non-permissible income (NPI) is approximated by interest income (InvestmentIncomeInterest / InterestAndDividendIncomeOperating / InterestIncomeOperating) divided by TTM revenue. This is a proxy, not a full segment decomposition. A complete NPI computation would parse 10-K segment disclosures for income from impermissible activities (alcohol distribution, conventional insurance reserves, riba-bearing instruments) — and would catch cases the proxy misses, such as BNPL or merchant-financing operators where interest spreads run through the operating revenue line rather than a standalone interest-income line, or insurance-adjacent businesses where float yield is buried in net investment income. At universe scale we use the proxy that the production screening engine uses and flag it explicitly; for the five named examples in §5 we cross-checked the proxy against the relevant 10-K. NPI binds on fewer than 10% of S&P 500 names in any quarter of this window — debt is the dominant failure driver — but the proxy limitation is real and is repeated in §8.

Trajectory taxonomy.

• →Stable-Compliant (SC): compliant in all 8 quarters
• →Stable-Non-Compliant (SN): non-compliant in all 8 quarters
• →Improving (IM): ≤2 compliant in Q1–Q4, ≥3 compliant in Q5–Q8, both Q7 and Q8 compliant
• →Deteriorating (DT): ≥3 compliant in Q1–Q4, ≤1 compliant in Q5–Q8, both Q7 and Q8 non-compliant
• →Oscillating (OS): ≥2 verdict switches, not satisfying IM or DT
• →Near-Improving (NI): exactly one verdict switch, from N to C; ends compliant. Includes single-late-flip patterns and patterns that don't meet IM's stronger h1≤2 / h2≥3 cutoff. Apple is the headline NI name.
• →Near-Deteriorating (ND): exactly one verdict switch, from C to N; ends non-compliant
• →Unclassified (UC): XBRL coverage incomplete in one or more quarters (issuer-specific debt or asset tagging did not map to the standard concept tree). After splitting out NI and ND, UC contains data-quality gaps only — not single-transition patterns.

Coverage. Of the 472 continuous constituents, 434 received a deterministic trajectory classification (SC/SN/IM/DT/OS/NI/ND) and 38 were UC — names where issuer-specific XBRL tagging did not map to the standard debt or asset concept tree. Apple is the largest-MC NI name (the verdict string is N-C-C-C-C-C-C-C — a single-quarter early breach on a 30.4% debt ratio at Q1 2024, then compliant every quarter through Q4 2025). §9.3 documents the UC residual.

Limitations stated upfront. US-only universe; 8 quarters is a short window for trajectory inference; NPI is a proxy; the methodology is single-provider (AAOIFI) and Piece 2 will compare against DJIM, MSCI, and FTSE Yasaar. Section 8 returns to each.

4. Headline Results

The trajectory distribution across 472 continuously listed S&P 500 names is sharply bimodal.

Stable-Compliant accounts for 142 names (30.1%). Stable-Non-Compliant accounts for 218 names (46.2%). The two stable categories together cover 76.3% of the universe. The remaining 23.7% splits between Unclassified (38, 8.1%) — these are XBRL-mapping data-quality gaps only — Oscillating (40, 8.5%), Near-Improving (11, 2.3%), Improving (9, 1.9%), Near-Deteriorating (8, 1.7%), and Deteriorating (6, 1.3%). Net, more names improved than deteriorated over the window: 20 names ended compliant after starting non-compliant (NI + IM), against 14 that ended non-compliant after starting compliant (ND + DT). The stronger framing is that 74 of the 434 classifiable names — 17.1% — changed verdict at least once during the window. That is materially more transition activity than the headline SC/SN split suggests.

The 17.1% in-window transition rate matters for ETF construction. Roughly 30% of the index is a low-maintenance compliant bucket. Roughly 46% is a permanently-excluded bucket. The active turnover happens in the IM/DT/OS/NI/ND bucket — 74 names that change verdict during the eight quarters. That is the trajectory product this paper exists to surface.

Sector tilt is the next-order story.

Eleven GICS sectors, eight-quarter window, six trajectory states. Three patterns dominate.

Financials zero out. All 68 Financials names are Stable-Non-Compliant. Banks, insurers, asset managers, payment networks, and the consumer-finance issuers are sector-disqualified by AAOIFI Standard 21 — the business model depends on interest spreads, conventional underwriting, or fees on interest-bearing products. The threshold screen is the wrong tool here; the business-permissibility screen is binding.

Utilities round-trip to the same place. 27 of 30 Utilities names are SN. Capital-intensive infrastructure businesses run high absolute debt loads, and electric and gas utilities have not historically built around equity-funded balance sheets. The three Utilities exceptions are sector outliers, not a class.

Information Technology and Industrials dominate the compliant bucket. IT posts a 52.4% SC rate (33 of 63), Industrials 47.3% (35 of 74), Energy 45.0% (9 of 20), Materials 44.0% (11 of 25). These are the natural candidate pools for any AAOIFI-aligned passive product on US large-caps. The IT compliant set includes most of the megacap names — MSFT, GOOGL, AMZN, NVDA, META, TSLA, ORCL — once the production-aligned cash definition is applied; the SN portion of IT is concentrated in semiconductor names with heavy debt loads from M&A (AVGO is the cleanest example) and a few cash-rich operators that exceed 30% on cash equivalents alone.

Real Estate and Health Care sit low. Real Estate is 19 of 31 SN — REITs carry capital-structure debt and most fail the 30% debt-to-assets test. Health Care is 26 of 59 SN (44.1%) — driven by the AAOIFI-disqualified managed-care plans (UnitedHealth, Elevance, Cigna, Humana, Centene, Molina) and the biotech debt-leverage outliers.

The implication for index construction is direct. An AAOIFI-aligned S&P 500 product, even before any rebalancing logic, is a sector-tilted product: Financials and Utilities almost entirely dropped, IT and Industrials retained at roughly half-the-sector SC rates, Materials and Energy punching slightly above. The choice is what to do with the 55 classifiable names that move between compliant and non-compliant during a holding period.

5. Trajectory Deep Dives

5.1 Stable-Compliant — MSFT

Microsoft is the unremarkable case made remarkable by scale. Across all eight quarters, debt/assets ranged from 6.1% to 9.3% and cash/assets ranged from 3.3% to 5.1%. Both ratios sit comfortably below a sixth of the 30% threshold for the entire window. Absolute debt held in the USD 40.3–45.1 billion range while total assets grew from USD 484 billion (Q1 2024) to USD 665 billion (Q4 2025) as the AI capex cycle expanded the balance sheet. Cash and equivalents ranged USD 17.5–30.2 billion. None of these movements approach a 30% breach. Core business is permissible. NPI is immaterial.

For Shariah portfolios, Microsoft is a passive holding. The compliance verdict is robust to plausible movements in either numerator (debt, cash) or denominator (total assets) over the window — Microsoft's asset base would need to compress by ~70% before the debt ratio threatens 30%, and cash and equivalents would need to grow by close to an order of magnitude on the current asset base for the cash ratio to do the same. Neither is on any visible path; if anything, the AI capex cycle is growing the asset base, pushing both ratios further from threshold.

The takeaway from the SC bucket as a whole — 142 names — is that Microsoft's structural picture is typical: capital-light operating model relative to the asset base, debt at conservative levels, cash position appropriate for working-capital and treasury needs rather than a strategic war chest. This is what 30% of the S&P 500 looks like under AAOIFI.

5.2 Stable-Non-Compliant — BA

Boeing is non-compliant in all eight quarters on the debt ratio. The ratio ranged from 32.0% (Q4 2025) to 41.7% (Q3 2024) — never inside threshold, never close. Absolute interest-bearing debt sat between USD 47.7 billion (Q1 2024) and USD 57.7 billion (Q2 2024) on an asset base of USD 134–168 billion. The cash ratio rose materially through the window — from 5.1% (Q1 2024) to 8.8% (Q4 2024) — but never breached 30%; debt is the binding constraint in every observed quarter.

The business is permissible. Aerospace and commercial defence manufacturing pass the activity screen in the methodology applied here at S&P 500 scale; the failure is purely capital-structure. Boeing's elevated debt traces to the 737 MAX programme, the pandemic cash burn, the Spirit AeroSystems acquisition, and supplier and labour disruption — collectively keeping leverage outside Shariah parameters since well before the start of this window. For descriptive context, debt would need to fall by roughly USD 5 billion (~10% of current outstanding) against a stable asset base to bring the ratio under 30%; the trajectory through 2024–2025 shows no such deleveraging path, but the research piece makes no forward statement about whether one will or will not emerge.

Broadcom (AVGO) is the more contemporary SN-debt case and the most informative for an AAOIFI-vs-DJIM comparison. AVGO's debt/assets ran 36.8–41.3% across the window — a heavy debt load tied to the USD 69 billion VMware acquisition completed late 2023. Permissible business (semiconductors and infrastructure software), passes a DJIM 33.33%/MC screen comfortably (debt/MC is ~5% on a ~USD 1.4 trillion market cap), and is stable-non-compliant under AAOIFI because the asset-based denominator catches the M&A leverage that the MC denominator hides. This is the methodology gap operationalised: AVGO is not a marginal case; it is structurally a stable-non-compliant name under AAOIFI's text.

The lesson for SN names is that the trajectory taxonomy correctly identifies them as durably out-of-bounds. They do not warrant quarterly re-screening at portfolio level. They warrant exclusion lists.

5.3 Improving — WELL (Welltower)

Welltower is the cleanest example of a debt-side improving trajectory. The verdict string reads N-N-N-N-C-C-C-C. Q1–Q4 2024 verdicts are all N, all on the debt ratio (debt/assets 30.4–32.2%). Q1–Q4 2025 verdicts are all C (debt/assets 28.3–29.5%).

The drivers are visible directly in the asset and debt rows. Total assets grew from USD 45 billion (Q1 2024) to USD 67 billion (Q4 2025) — a 48% expansion driven by senior-housing portfolio acquisitions and selective new development. Absolute debt grew too, from USD 14.2 billion to USD 19.2 billion, but proportionally less — a 35% increase. The ratio mechanically fell from 31.9% to 28.5%. The Q1 2025 crossing was the first quarter the trailing 12-month operating activity, equity-funded acquisitions, and incremental debt issuance combined to push the ratio under 30%.

This is not a denominator artifact. The senior-housing demographic tailwind expanded Welltower's economic footprint; the company funded most of the expansion through equity issuance and asset-level joint ventures rather than debt. The improvement is real capital-structure improvement, captured by AAOIFI's Total Assets denominator exactly because the denominator is reflecting the operating reality.

The takeaway for IM trajectories under AAOIFI is that they are slow and they reflect real corporate decisions. The 9 IM names in this universe — Welltower, Ross Stores, Howmet Aerospace, Halliburton, Smurfit Westrock, McCormick, United Airlines, EPAM, and APA Corporation — share a pattern of asset-base expansion outpacing or matching debt growth, or of cash drawdown that brings the cash ratio back inside.

5.4 Deteriorating — UPS

United Parcel Service is the cleanest DT case in the universe. The verdict string reads C-C-C-N-C-N-N-N. Three quarters compliant, one breach at Q4 2024 (debt/assets 30.0%), brief recovery at Q1 2025 (28.5%), then permanent N through Q2-Q4 2025 (33.6%, 33.4%, 32.3%).

The mechanism is debt issuance against a roughly flat asset base. Total assets stayed in a USD 68–73 billion band across the entire window. Absolute debt grew from USD 18.8 billion (Q1 2024) to USD 23.8 billion (Q2 2025) — a USD 5 billion increase, which moved the ratio from 27.9% to 33.6%. The Q2 2025 jump (from USD 19.5 billion to USD 23.8 billion in a single quarter) is the inflection: UPS issued debt to fund operational restructuring and the small-package-network reorganisation announced through 2025. The asset base did not grow in lockstep.

This is the structural counterexample to Welltower. Welltower grew assets faster than debt; UPS grew debt against a flat asset base. The ratio crossed 30% mechanically. The verdict trajectory captures the consequence directly.

The lesson for DT names is that they often signal a strategic event in progress — an acquisition being funded, a capex cycle beginning, a restructuring being financed. The 6 DT names in this universe — UPS, PPG Industries, LyondellBasell, Zimmer Biomet, Ball Corporation, Avery Dennison — split between Industrials and Materials, both sectors where debt-funded capacity decisions and M&A drive multi-quarter balance-sheet shifts.

5.5 Oscillating — LLY

Eli Lilly is the highest-MC oscillating name and a clean example of how a fast-growing capex-heavy operator can run hot against the 30% debt/assets line. The verdict string reads C-C-N-N-N-C-N-N: two quarters compliant, then non-compliant through Q1 2025, briefly compliant at Q2 2025, then non-compliant for the back half.

The mechanism is the GLP-1 capacity build. Total assets grew from USD 64 billion (Q1 2024) to USD 112 billion (Q4 2025) — a 75% expansion in seven quarters, driven by manufacturing capacity additions for Mounjaro and Zepbound. Absolute debt grew faster: USD 19.1 billion (Q1 2024) → USD 31.1 billion (Q3 2024, after a large acquisition tranche) → USD 42.5 billion (Q3-Q4 2025). The ratio oscillates: 29.9% → 26.6% → 41.2% (Q3 2024 acquisition pushes debt up, asset base hasn't caught up) → 37.4% → 33.0% (asset base growing) → 29.2% at Q2 2025 (asset base finally catches up) → 37.0% (next debt issuance) → 37.8%.

The takeaway is that fast-growing capex operators with active M&A produce noisy debt ratios. Lilly's underlying capital structure may be sustainable on a forward-looking basis — earnings growth justifies the debt — but a static threshold cannot tell whether each ratio breach is durable or transient. The 40 OS names in this universe are concentrated in Health Care, Information Technology, Industrials, and Materials — the sectors where capex, M&A, and active treasury management produce frequent threshold crossings.

Oscillating names are the most operationally awkward for Shariah ETFs. They are not categorically excluded (SN) and they are not durably included (SC). A rules-based fund must either drop them on every breach quarter, accept a position that crosses the threshold mid-holding, or apply a trajectory-aware filter that holds them through transient breaches above some tolerance. The 40 OS names in this universe are where ETF methodology choice matters most.

6. Drivers of Transitions

Across the 415 classifiable names, 146 verdict transitions occurred in the window: 69 compliant-to-non-compliant (C → N) and 77 non-compliant-to-compliant (N → C).

The driver distribution is concentrated. Of the 69 C → N transitions, 57 were debt-only, 5 were combined debt+cash, and 7 were cash-only. Of the 77 N → C transitions, 58 were debt-driven (reversal of debt-only breaches), 16 were cash-driven, and 3 were combined. Net across all transitions: debt accounts for 123 of 146 driver flags (84%), cash for 26 (18%), and the small overlap reflects combined breaches.

What this rules in and out:

• →Debt-ratio breaches dominate. Either the issuer added net debt (UPS, the DT case) or grew assets faster than debt (Welltower, the IM case). Lilly is the textbook combined case — both numerator and denominator growing fast, with the ratio oscillating around 30% depending on which got ahead quarter-to-quarter.
• →Cash-ratio breaches are a secondary feature. Under the production cash definition (cash and equivalents only), cash-ratio events account for only 18% of transitions. Industrial and consumer-staples names with cyclical cash positions (Ross Stores, EPAM, McCormick) account for most cash-driven movement; large-cap tech names sit comfortably under the 30% cash threshold on this definition.
• →NPI-driven transitions do not appear. Across 146 transitions, zero are NPI-driven in the proxy applied here. NPI > 5% is sticky over an 8-quarter window because interest income is sticky, and S&P 500 large-caps are dominated by operating-company names where interest income does not approach 5% of revenue. Smaller-cap universes may show different patterns.
• →Business-activity transitions also do not appear. No S&P 500 name materially changed its core-business permissibility classification inside this window. M&A and divestitures occurred, but none flipped the qualitative business screen.
• →Asset-base growth is the silent driver. When a name's debt/assets ratio falls from above to below 30% with absolute debt unchanged, the driver tag still reads "Debt" in the data, because the ratio that breached was the debt ratio. But the cause was the denominator expanding. Welltower is the cleanest example of this. The case for surfacing total assets alongside debt and cash in any consumer-facing Shariah product is direct: without the denominator visible, the user infers a debt-issuance event that may not have occurred.

The single most actionable finding for fund operators is that 84% of compliance transitions under AAOIFI are debt-driven, and the denominator's slow movement makes most AAOIFI transitions durable rather than transient. A trajectory-aware screen distinguishes the durable from the transient. A binary screen cannot.

The Q4 2025 cross-section of debt ratio against cash ratio, coloured by trajectory, gives the static picture against which the dynamic picture in this paper should be read.

SC names cluster in the bottom-left quadrant (both ratios well under 30%); SN names spread along the right edge — the right edge for debt-failures (BA, AVGO, ABBV, HD). OS, IM, and DT names populate the boundary corridor where transitions concentrate. A binary screen at Q4 2025 alone would classify every name in this chart correctly for Q4 2025. The trajectory data tells you which classifications will still be correct at Q1, Q2, or Q4 of 2026.

7. Implications

For Shariah ETF construction

The 17.1% in-window transition rate among classifiable names sets the order-of-magnitude turnover that a quarterly-rebalanced AAOIFI-screened product would absorb. The 69 C → N transitions are the outbound events. Spread across the seven quarter-transitions in the window, that is roughly ten names dropped per quarter from a held universe of ~160 current-compliant names (after excluding the SN bucket entirely), a 6% per-quarter outbound rate from compliance flips. The 77 N → C events are mostly re-entries from the SN bucket that a strict AAOIFI ETF will have excluded entirely; only the IM + NI subsets (20 names) are operationally relevant for inbound additions.

This is materially more turnover than a DJIM-style screen would impose on the same universe. AAOIFI binds harder. The trade-off: an AAOIFI-screened product holds fewer names (30% of the universe is SC, vs ~43% under a DJIM 33.33%/MC screen) and rebalances more often, but the Shariah-board defence of the screen is straightforward and textual.

Sector tilt is structural, not cyclical. Underweight Financials and Utilities is permanent. The IT sector — the natural AAOIFI candidate pool — produces an SC rate of 52.4%; the SN portion is concentrated in semiconductor names with M&A leverage (AVGO is the clearest example). A Shariah ETF marketed as "S&P 500-like" is, factually, a sector-tilted product: Financials and Utilities dropped almost entirely, IT and Industrials retained but at roughly half-the-sector SC rates, Materials and Energy punching slightly above. Fund prospectuses should be explicit about this rather than presenting it as a contingent outcome of recent screens.

A trajectory-aware filter — for example, "exclude names with ≥3 N quarters in trailing 8" — would treat OS and DT names differently from SC and IM names without committing to a binary flip on every transient breach. The 40 OS and 9 IM names in this universe sit at exactly that boundary; the choice of how to handle them is a methodology choice with measurable turnover and tracking-error consequences.

For Shariah boards and purification

Standard purification practice computes the non-permissible share of dividend income (typically the NPI/revenue ratio applied to dividends received) and donates that fraction. The practice assumes a stable compliance state. When a holding spends part of the period non-compliant, the calculation is mis-specified — the dividend received during the non-compliant window arguably warrants different purification treatment than the dividend received during the compliant window.

There is no settled doctrinal answer to this within AAOIFI's published standards, partly because the standards themselves assume a binary state. A trajectory record — name X was N for Q1–Q4 2024, C for Q1–Q4 2025 — is the input a contemporary board would need to formulate a position. The data product introduced here is, in this sense, prior to the doctrinal question. The doctrinal question cannot be addressed without the data.

The Welltower case in §5.3 is the clearest illustration. An investor who held WELL since Q1 2024 received dividend income for four quarters of non-compliance and four quarters of compliance. Purifying the full holding at the current verdict's NPI rate under-purifies. Purifying as if the name were continuously N over-purifies. Trajectory data resolves the input; the doctrinal question of how to compute remains.

For screening platforms

Binary verdicts answer a different question than trajectory data answers. Binary asks "is this name halal today?" Trajectory asks "has this name been halal across my holding period, and what drove the answer?" Both are legitimate questions; only the first is currently served by mainstream Shariah screening providers.

Trajectory adds three product surfaces. The first is the verdict history itself — eight or sixteen or thirty-two quarters of C/N, dated to filing periods. The second is the driver decomposition — for each transition, debt or cash or NPI or business, with the underlying ratio chart and the balance-sheet item that drove it. The third is the trajectory classification — SC, SN, IM, DT, OS, UC — as a stable property of a name's recent screening behaviour.

A user shown only "compliant" for UPS at Q1 2025 cannot anticipate the Q2 2025 flip and the durable deterioration that followed. A user shown the full eight-quarter ratios with the trajectory tag "Deteriorating" can. The information cost of producing the second is the data work in this paper, repeated quarterly. The information value to a Shariah investor planning entry points, exit points, and purification accruals is the difference between a yes/no answer and a continuous picture.

This is the product gap. Halal Terminal's screening API and frontend will expose the trajectory layer alongside binary verdicts as a direct consequence of this study.

8. Limitations and What's Next

AAOIFI-only. This piece applies AAOIFI Standard 21 by the production-aligned reading — 30% threshold on a Total Assets denominator, with cash and equivalents (not the full marketable-securities portfolio) as the cash numerator. The study does not screen the same universe under DJIM, MSCI, or FTSE methodologies. Piece 2 will. The 3-percentage-point threshold gap (30% vs 33.33%) and the denominator split between Total Assets (AAOIFI / MSCI main / FTSE Yasaar) and trailing market cap (DJIM / MSCI M-Series) are the load-bearing variables we will quantify.

Cash-numerator choice. The narrow cash definition adopted here (cash and equivalents only) is the choice the production engine makes and the one HT customer-facing verdicts use. A reader applying the broader AAOIFI text — including long-term marketable securities portfolios — would find AAPL, NVDA, ORCL, and a handful of other cash-rich operators flip from compliant to non-compliant. The disclosure in §3 surfaces this; the analysis presents the production reading.

US-only. S&P 500 only. Global extension to FTSE 350, TOPIX, and emerging-market large-cap universes is straightforward in principle but requires non-EDGAR filing-system equivalents and harmonised XBRL concept maps. Out of scope here.

Eight quarters is short. The trajectory categories assume a long-enough window to distinguish stable trajectories from extended runs in either direction. A name that is C for ten quarters and N for the next three would read SC in our window if both periods straddle the boundary. A 16- or 24-quarter follow-up will sharpen the IM/DT/OS distinctions and is on the roadmap.

Slow-denominator effect on trajectory inference. Under AAOIFI's Total Assets denominator, IM and DT trajectories are rare (9 and 6 names respectively) because total assets is a slow-moving balance-sheet item. Most trajectory movement registers as OS (40 names) where the underlying balance sheet is genuinely choppy quarter-to-quarter. A 16-quarter window would likely shift the IM/DT/OS split toward IM and DT as multi-quarter directional moves get more room to express. This is itself a methodological finding: AAOIFI's classical denominator produces durable verdicts and discounts noise — a property that is desirable from a Shariah-board perspective but constrains the trajectory taxonomy's resolution on short windows.

The 57 UC names are a real coverage gap. Most are business-screen-disqualified (financials, healthcare plans, casinos, tobacco) where ratio data was incomplete but the activity screen is binding. The binding remainder is roughly 20 names where standard XBRL debt or asset concepts did not map to issuer-specific tagging, and a small handful of names (Apple is the prominent example) whose single-quarter early breach did not fit any of the directional categories. Custom XBRL element extraction and a relaxed IM cutoff (≤3 in h1 rather than ≤2) would close most of the residual gap in the next pass.

NPI is a proxy. Interest-income / TTM revenue is what the production screening engine uses at scale. For the five named examples we cross-checked against the 10-K. For the 472-name universe we did not — the proxy is the disclosed input. A full segment-level NPI parse from 10-K language is technically feasible but not done here.

Engagement with the December 2025 continuous-compliance paper. Qadi, Sharma, and Medda's "Beyond Binary Screens: A Continuous Shariah Compliance Index for Asset Pricing and Portfolio Design" (arXiv 2512.22858, v1 December 2025; revised March 2026) proposes a Continuous Shariah Compliance Index (CSCI) on a [0,1] scale that consolidates six screening standards — AAOIFI, DJIM, FTSE, MSCI, S&P Shariah, and the Securities Commission Malaysia — into a single transparent measurement, applied to US equities from 1999 through 2024.⁶ The CSCI maps each financial ratio into a smooth score that declines from 1 (comfort zone) through intermediate values to 0 (outer-bound breach), capturing how compliant a firm is rather than just whether it qualifies. The paper documents that roughly 74% of permissible-sector firm-months sit strictly between 0.01 and 0.99 — i.e. the binary verdict obscures substantial continuous variation that the CSCI exposes. The proposal is methodologically adjacent to what this study surfaces empirically — both are responses to the same point-in-time problem that motivates §2. The trajectory taxonomy here is coarser than a continuous index (six categories vs a [0,1] score) but is interpretable to a Shariah board in the language of the existing binary screen: SC and SN names map cleanly onto the existing compliant/non-compliant taxonomy, while IM, DT, and OS are new categories with operational meaning. Piece 3 of this series will reconcile the two approaches — the binary-with-trajectory framing developed here, and the continuous-score framing in the paper — on the same S&P 500 universe over the same window, and will identify cases where the two disagree.

The short version: the field is converging on the same problem statement from different directions. The product question is which framing a Shariah board, an ETF operator, and an end investor can each act on.

9. Appendix

9.1 Reproducibility

The methodology is documented end-to-end in §3 and the limitations section.

Public ticker-level datasets (CC BY 4.0, attribution to Halal Terminal Research):

• →sp500-aaoifi-master.csv — one row per name, with sector, industry, eight-quarter verdict string, trajectory tag, and primary failure driver. 472 rows.
• →sp500-aaoifi-ratios-timeseries.csv — one row per name per quarter, with total assets, debt, cash and equivalents, revenue, interest income, debt/assets ratio, cash/assets ratio, NPI ratio, verdict, and failure reason. 3,776 rows.

The pre-publication verification log and additional aggregate cross-tabs are available on request for Shariah boards, index providers, and academic researchers. Contact Yassir at yassir@halalterminal.com.

9.2 Failure-reason coding

• →Debt — interest-bearing debt / total assets ≥ 30%
• →Cash — cash and cash equivalents / total assets ≥ 30%
• →NPI — interest income (proxy) / TTM revenue ≥ 5%
• →Business — core business not permissible under AAOIFI Standard 21
• →Combos — concatenation (e.g. "Debt+Cash") where multiple ratios breach in the same quarter

9.3 UC bucket and the Near-trajectory categories

After splitting the prior UC pool into three sub-buckets, the residual UC bucket contains data-quality gaps only — the 38 names where issuer-specific XBRL tagging did not match the standard debt or asset concept tree used here. The bucket is concentrated in: (a) auto OEMs and a handful of capital-goods issuers (F, GM, CAT, DHI), (b) several biotechs (ALGN, DXCM, EA), and (c) the cash-rich tech subset (ANET, AKAM, FFIV). Custom XBRL element extraction for these issuers is a one-time engineering task and will close most of the residual gap in the next pass. We did not impute any UC name into a deterministic verdict.

The two new sub-buckets — Near-Improving (NI) and Near-Deteriorating (ND) — capture single-transition patterns previously folded into UC:

• →

  NI (11 names) — AAPL, CHRW, CSGP, ECL, IFF, LULU, PNR, PODD, SRE, TRMB, WAT. Each shows exactly one verdict switch from N to C and ends compliant. Apple is the highest-MC NI name: a Q1 2024 debt/assets reading of 30.4% triggered a single N quarter, followed by seven consecutive compliant quarters. The pattern does not satisfy our IM cutoff (≤2 C in h1) — AAPL has c_first = 3 — but the directional signal is unambiguous. A user weighing Apple as a Shariah holding cares more about the seven-quarter compliant run than the 30.4% single-quarter blip; NI surfaces exactly that distinction.

• →

  ND (8 names) — APTV, DGX, DOW, MDT, MMM, MPC, PFE, WY. Each shows exactly one verdict switch from C to N and ends non-compliant. Several are large industrials and pharma (3M, Pfizer, Medtronic) where late-window debt issuance or asset-base contraction crossed the threshold. ND names are operationally similar to DT names but the directional evidence is weaker (single-transition only).

We did not relax the IM/DT thresholds themselves — that would have shifted the trajectory counts under a definition change. Instead we surface NI and ND as adjacent labels so users can tell "directionally improving" from "fully Improving" and from "data quality gap."

9.4 Universe constituents

The continuous 472-name universe, with sector, industry, eight-quarter verdict string, trajectory tag, and primary failure driver, is available on request alongside the data products in §9.1.

---

References

Additional standards referenced: AAOIFI Shari'ah Standard No. 21: Financial Papers (Shares and Bonds), 2004 edition.

Footnotes

1. →

   Islamic Corporation for the Development of the Private Sector (ICD) and LSEG, Islamic Finance Development Indicator Report 2025. Industry-asset figure verbatim from the report: "the industry has expanded its footprint to 140 countries by 2024 and reached a valuation of US$5.98 trillion, marking an impressive 21% growth in just one year"; projection verbatim: "If growth continues at current levels, a figure of US$9.7 trillion is forecast by 2029." ↩

2. →

   S&P Dow Jones Indices, Dow Jones Islamic Market Indices Methodology (current version, 2025). Debt and cash thresholds set at 33.33% on a trailing 24-month average market-cap denominator. ↩

3. →

   MSCI, MSCI Islamic Index Series Methodology, May 2025. The main MSCI Islamic Index Series applies 33.33% leverage and liquid-assets thresholds on a Total Assets denominator (with 30%/35% entry/exit buffers); the parallel MSCI Islamic M-Series Indexes apply the same 33.33% thresholds on a 36-month average market-cap denominator. ↩

4. →

   FTSE Russell and Yasaar Limited, FTSE Yasaar Global Equity Shariah Index Series Ground Rules, v4.6, February 2026. Applies a 33.333% threshold on debt/total-assets and cash-plus-interest-bearing-items/total-assets, with an accounts-receivable-plus-cash/total-assets cap of 50%. ↩

5. →

   Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI), Shari'ah Standard No. 59: Sale of Debt (Bahrain: AAOIFI). The standard's promulgation aligned the screening framework with the removal of the historic illiquid-assets test that earlier screens carried over from pre-2018 practice. ↩

6. →

   Abdulrahman Qadi, Akash Sharma, and Francesca Medda, "Beyond Binary Screens: A Continuous Shariah Compliance Index for Asset Pricing and Portfolio Design" (v1 28 December 2025; v2 11 March 2026 revised under the title "From Binary Screens to Continuous Compliance: A Shariah Screening Measure for Portfolio Design"), arXiv:2512.22858. ↩