Sum-of-the-Parts Valuation for Conglomerates

When General Electric completed its three-way split in April 2024, the combined market capitalization of GE Aerospace, GE Vernova, and GE HealthCare exceeded $215 billion--more than four times what the conglomerate had been worth before it started separating. The value was always there. The market simply refused to pay for it while the pieces were bundled together. That gap between what the parts are worth individually and what a conglomerate trades for is the conglomerate discount, and it averages 13-15% in developed markets (reaching 30% or more in extreme cases). Sum-of-the-parts valuation is the tool that measures that gap--and the real play against treating a software division, a healthcare unit, and an industrial segment as though they all deserve the same multiple.

Why a Single Multiple Fails for Conglomerates (The Core Problem)

You would never value a hospital the same way you value a cloud software company. Yet that is exactly what happens when you slap a single P/E or EV/EBITDA multiple on a conglomerate's blended earnings. A semiconductor division trading at 27x EV/EBITDA and an oil refining segment worth 5.5x cannot be averaged to 15x without destroying information about where value actually resides.

The point is: SOTP valuation disaggregates the conglomerate into its constituent business units, values each one using the methodology its pure-play peers deserve, and then reassembles them into a total enterprise value. It is the only approach that respects the fundamental reality that different businesses have different risk profiles, growth rates, and capital intensity.

This matters whenever you encounter:

• Conglomerates with distinct operating segments (Berkshire Hathaway, Honeywell, historically GE)
• Holding companies with minority stakes in other public businesses
• Companies where segment disclosure reveals fundamentally different business models buried inside one stock ticker
• Activist situations where someone is arguing that breaking up the company unlocks value (and they usually have the math to prove it)

The SOTP Formula (How the Math Works)

Total Enterprise Value = Sum of all Segment Values - Unallocated Corporate Costs - Net Debt

Each segment gets valued using whichever methodology its industry peers use:

The takeaway: you must match the valuation method to the segment, not force every segment through the same framework. This is more work than applying a blended multiple. That extra work is precisely where the edge lives.

Building an SOTP Model Step by Step (The Practitioner's Workflow)

Step 1: Map the Segments from Filings

Start with segment reporting from the 10-K. Companies are required to disclose revenue, operating income, and assets by reportable segment under ASC 280. Pull those numbers directly.

Why this matters: limited segment-level data is the single biggest constraint on SOTP analysis. If the company aggregates a meaningful business into "Other" (and many do), you have a data problem before you even start. You cannot value what you cannot measure.

Step 2: Build Separate Comp Sets for Each Segment

For each segment, identify 3-5 pure-play comparable companies and calculate median valuation multiples. You are essentially building multiple standalone comparable company analyses inside one model.

Consider a hypothetical conglomerate with three divisions:

• Technology segment: Pure-play SaaS peers trading at 28-35x EV/EBITDA
• Healthcare segment: Medical device comparables at 18-22x EV/EBITDA
• Consumer goods segment: FMCG peers at 12-15x EV/EBITDA

The point is: the comp selection for each segment matters enormously. Choose peers that match on growth rate, margin profile, and end-market exposure (not just broad industry classification). A high-growth healthcare AI company is not a valid comp for a mature medical device business, even though both sit in "healthcare."

Step 3: Calculate Segment Enterprise Values

For each segment, multiply the relevant financial metric by the appropriate multiple:

Step 4: Handle Corporate Overhead and Net Debt

Unallocated corporate costs--headquarters, shared services, executive compensation not assigned to segments--must be deducted separately. The standard approach is to capitalize them at a blended multiple (typically 8-10x, reflecting lower-quality, non-revenue-generating cost):

• Corporate overhead: $400M annually x 8x = ($3.2B)
• Net debt (total debt minus cash): ($8.0B)

SOTP Equity Value = $26.2B - $3.2B - $8.0B = $15.0B

If the company's market cap is $11.5B, the implied conglomerate discount is 23%. That gap is either your opportunity or a warning (depending on why it exists).

Step 5: Stress-Test the Discount

What the data confirms: a discount is not automatically an error in your model. It may reflect the market's legitimate skepticism about management's capital allocation, segment integration complexity, or the probability that a breakup actually happens.

Ask yourself three questions:

1. Is management subsidizing underperformers? If cash flow from the technology division is funding a struggling consumer goods turnaround, investors are right to discount.
2. Is there an activist catalyst? Elliott Management's $5 billion stake in Honeywell in November 2024 (their largest single-stock position ever) is the kind of catalyst that can close the gap.
3. Would separation destroy synergies? Shared R&D, cross-selling, and centralized procurement create real value that pure SOTP ignores.

The Conglomerate Discount (Why Markets Punish Diversification)

Empirical research consistently finds that diversified companies trade at 13-15% below their SOTP value in the U.S. and Western Europe. A 2024 German study analyzing roughly 6,000 firm-years found discounts of 7.9-11.5% even after controlling for selection bias and measurement effects. In emerging markets (particularly India), discounts can reach 30-50%.

The discount reflects several forces working simultaneously:

• Complexity tax: Analysts and investors struggle to model diverse businesses, so they apply a lower multiple to the whole (rather than doing the SOTP work themselves)
• Capital misallocation: Profitable segments subsidize underperformers through internal capital markets, destroying value that external markets would reallocate
• Management bandwidth: Running unrelated businesses dilutes strategic focus--the CEO who oversees software, healthcare, and consumer goods is expert in none of them
• Investor self-selection: Investors who want pure software exposure do not want it bundled with industrial equipment (this forces selling pressure from specialists who cannot own the conglomerate)

Why this matters: the discount is not irrational. It is the market pricing in real agency costs and information asymmetry. The opportunity arises when the discount exceeds what those costs justify--or when a catalyst (activist, management change, strategic review) makes separation probable.

The Breakup Wave of 2023-2025 (Real-World Proof)

The recent wave of conglomerate breakups provides the most compelling evidence that SOTP analysis works:

General Electric: Separated into GE Aerospace, GE Vernova, and GE HealthCare between 2023-2024. Combined market cap post-separation exceeded $215 billion--more than 4x the pre-split value. The aerospace business alone was worth more than the entire conglomerate had been.

Honeywell: Under pressure from Elliott Management (which built a $5 billion position in late 2024), Honeywell announced a three-way split in February 2025 into Honeywell Aerospace, Honeywell Automation, and an Advanced Materials company. Wolfe Research's SOTP analysis projected ~$293 per share of value in the separated entities, a substantial premium to the pre-announcement trading price.

Johnson & Johnson: Split off its consumer health division as Kenvue in 2023, allowing the higher-growth pharmaceutical and medical device businesses to trade on their own merits.

3M: Spun off its healthcare business as Solventum in 2024, separating a high-margin healthcare operation from the legacy industrial and consumer conglomerate.

The point is: these are not theoretical exercises. Every major conglomerate breakup in this cycle has validated the SOTP framework's core insight--the pieces are worth more than the whole when the market can value them independently.

Common SOTP Errors (And How to Avoid Them)

Using the same multiple for every segment. This is the most frequent mistake and it defeats the entire purpose of the analysis. If you are applying a blended multiple, you are just doing a standard comp analysis with extra steps.

Double-counting corporate costs. If corporate overhead is already allocated to segments in the filings (check the footnotes), deducting it again will understate your value. If it is unallocated, handle it once and document your assumption.

Ignoring inter-segment transactions. If the technology segment sells software to the consumer goods segment, that revenue gets eliminated in consolidation. Valuing segment revenue without adjusting for these transfers overstates the sum.

Overvaluing synergies that would be lost. SOTP tells you what the pieces are worth separately. It does not tell you what they are worth together. Some conglomerates genuinely create value through shared capabilities (Berkshire Hathaway's insurance float funding other acquisitions is the classic example). The test: would each segment's margins deteriorate meaningfully as a standalone company?

Anchoring on stale comparables. Peer multiples shift with market conditions. A tech segment valued at 35x EBITDA in late 2021 might warrant 22x in a rising-rate environment. Update your comp sets quarterly, not annually.

When SOTP Signals Opportunity (Three Patterns to Watch)

Pattern 1: Market cap far below SOTP value with an activist catalyst. If your SOTP shows $100B of value and the stock trades at $70B equity value, and Elliott or Starboard has just filed a 13D, the probability of value realization rises dramatically. In 2024 alone, strategy-focused activist campaigns more than tripled their share of total activism (from 8% in 2020 to 25% in 2024).

Pattern 2: One segment dominates value. If 80%+ of SOTP value comes from a single segment, the company is really a bet on that business. The other segments are noise--or worse, distractions consuming management attention and capital that would be better deployed in the core. This was exactly the GE Aerospace story.

Pattern 3: Hidden value in minority stakes. Conglomerates sometimes hold minority positions in other public companies that are carried at book value on the balance sheet but should be marked to market in your SOTP analysis. The gap between book and market value can be substantial (and often overlooked).

SOTP Valuation Checklist (Tiered by Impact)

Essential (high ROI)

These four steps prevent 80% of SOTP errors:

• Pull segment data directly from 10-K filings (not third-party summaries)
• Build separate comp sets with 3-5 pure-play peers per segment
• Verify whether corporate overhead is allocated or unallocated in filings
• Calculate the implied conglomerate discount and compare to the 13-15% historical average

High-impact (workflow refinement)

For investors who want rigorous SOTP models:

• Cross-check multiples against DCF for the largest segments (at least two methods per segment)
• Adjust for inter-segment revenues before applying multiples
• Research activist filings (13D/13F) for external SOTP analyses you can benchmark against
• Sensitivity-test each segment's multiple +/- 2-3x to see which drives the most value variance

Optional (for catalyst-driven investors)

If you are specifically looking for breakup opportunities:

• Screen for conglomerates trading at >20% discount to estimated SOTP
• Monitor activist 13D filings in conglomerate-heavy industries
• Track management commentary on "strategic reviews" or "portfolio simplification" (these are the precursors to actual separations)

Next Step (Put This Into Practice)

Pick one conglomerate you follow--Honeywell, Berkshire Hathaway, or any diversified company in your portfolio--and pull its segment disclosure from the most recent 10-K. For each material segment (anything above 10% of revenue), find three pure-play comparable companies and calculate median EV/EBITDA. Multiply each segment's EBITDA by its peer multiple, subtract corporate overhead and net debt, and compare your SOTP equity value to the current market cap. If the discount exceeds 20%, research whether any activists have taken positions. Their public presentations often include SOTP analyses you can calibrate against--and their involvement is frequently the catalyst that closes the gap.