When divisions are worth more apart than together. Conglomerate discount, SOTP mechanics, and a worked example using IndustrialCo — a four-division conglomerate with $800M in EBITDA and four distinct businesses that the market is undervaluing as a single entity.
Sum-of-the-parts (SOTP) is the methodology you use when a single EV/EBITDA multiple would be intellectually dishonest. Applying an 11x average multiple to a conglomerate with a 14x-quality aerospace division, a 10x-quality chemicals business, a financial services unit that should be valued on P/E, and a real estate portfolio that should be valued on asset value doesn't produce a meaningful number — it produces noise. SOTP values each segment with the methodology appropriate to that segment, then aggregates. The result is almost always higher than what the stock trades at, which is the conglomerate discount in action.
SOTP is appropriate in specific situations where a blended multiple would distort value. The clearest signal is a company with divisions operating in industries where market multiples differ meaningfully — a conglomerate with both a software division (high multiple) and a commodity chemicals division (low multiple) cannot be fairly valued by applying one number. Other triggers:
The key constraint that forces SOTP is the absence of a credible single-multiple peer set. If all divisions competed in the same industry at the same margin profile, a standard EV/EBITDA comps analysis would suffice. SOTP is the solution to heterogeneity.
IndustrialCo is a publicly traded conglomerate with four divisions: Aerospace Components, Industrial Chemicals, Financial Services, and Real Estate Holdings. It currently trades at a 20% discount to what the divisions would be worth individually. Here is the segment overview:
| Division | EBITDA / Metric | Industry | Appropriate Multiple | Implied EV |
|---|---|---|---|---|
| Aerospace Components | $400M EBITDA | Aerospace / Defense | 14x EV/EBITDA | $5,600M |
| Industrial Chemicals | $250M EBITDA | Specialty Chemicals | 10x EV/EBITDA | $2,500M |
| Financial Services | $180M Net Income | Financial Services | 12x P/E | $2,160M |
| Real Estate Holdings | Asset-based | Real Estate | Fair asset value | $800M |
| Total Sum of Parts | $11,060M |
Note the methodology variation by segment. Aerospace and Chemicals use EV/EBITDA because they're industrial businesses with meaningful D&A and capex — EBITDA is the right profitability proxy. Financial Services is valued on P/E because for financial companies, interest income is revenue and debt is a product input, so EBITDA is meaningless; net income is the appropriate earnings metric and P/E is the standard multiple. Real Estate is valued on asset value — NAV analysis or appraised value — because the business is an asset holding structure, not an operating company, and EBITDA understates the value of appreciated property.
The Aerospace division gets the highest multiple for a reason: defense contractors have long-cycle government contracts, high backlog visibility, and pricing power tied to national security spending — the combination of growth predictability and inelastic demand supports a 14x multiple. Compare that to specialty chemicals at 10x: cyclical commodity exposure, capital-intensive, subject to feedstock price volatility. Same conglomerate, 40% difference in the appropriate multiple.
SOTP gross value is $11,060M, but that's not the full picture. IndustrialCo's corporate headquarters incurs $80M per year in unallocated G&A — headquarters staff, legal, audit, treasury, and executive compensation that doesn't belong to any individual division. These costs exist because the company is structured as a conglomerate; if the divisions were independent, they would bear their own overhead, but corporate overhead would be eliminated. To reflect this, you capitalize the overhead burden and deduct it from SOTP value.
The 10x capitalization multiple reflects a blended sector multiple — you're treating corporate overhead as a perpetual cost that reduces enterprise value dollar-for-dollar when capitalized. Some practitioners use 8x–12x depending on how confident they are that overhead would truly be eliminated in a breakup; use 10x as the base case and sensitize.
With SOTP enterprise value and overhead adjustment in hand, bridge to equity value by subtracting net debt and any other enterprise-level adjustments:
| Item | Amount | Notes |
|---|---|---|
| Gross SOTP Enterprise Value | $11,060M | Sum of four division EVs |
| Less: Corporate Overhead PV | ($800M) | $80M/yr × 10x capitalization |
| Adjusted SOTP Enterprise Value | $10,260M | |
| Less: Total Debt | ($2,700M) | Total face value of debt |
| Plus: Cash & Equivalents | $500M | Unrestricted corporate cash |
| Net Debt | ($2,200M) | |
| SOTP Equity Value | $8,060M | |
| Diluted Shares Outstanding | 400M | |
| SOTP Implied Share Price | $20.15 | $8,060M ÷ 400M shares |
The SOTP implies $20.15 per share. If IndustrialCo currently trades at $16.00/share, the stock is trading at a 20.6% discount to intrinsic SOTP value. This is the conglomerate discount quantified.
Conglomerates empirically trade at a 10–20% discount to the sum of their parts, and the discount has been persistent across market cycles. The mechanism is multi-layered: (1) management complexity — running four different businesses in four industries requires generalist oversight that is typically less effective than focused, specialized management; (2) capital allocation inefficiency — headquarters allocates capital across divisions based on internal politics and legacy business weights rather than marginal return on investment, a problem focused companies don't have; (3) lack of focus — investors who want aerospace exposure would rather buy a pure-play aerospace company; forced conglomerate exposure makes it harder for sector specialists to own the stock; (4) index fund selling pressure — passive flows go to sector indices, and conglomerates don't cleanly fit any single sector ETF; (5) transparency deficit — segment-level financial disclosure is usually less detailed than a standalone company's, making deep analysis harder.
In IndustrialCo's case: SOTP value of $20.15/share vs. current trading price of $16.00/share = a $4.15 conglomerate discount, representing 20.6% of SOTP value. This is near the top of the historical 10–20% range, which is exactly the argument an activist investor would make in a public campaign advocating for a breakup or spinoff.
The highest-value unlock in an SOTP analysis is usually to spin off the highest-multiple division as a pure-play, then allow the remaining entity to re-rate at a more focused multiple. For IndustrialCo, the Aerospace division at 14x is the obvious candidate — it's pulling the blended group multiple up, but the market is discounting the group as a messy conglomerate instead of giving Aerospace full credit.
| Entity | EBITDA | Multiple | Implied EV | Notes |
|---|---|---|---|---|
| Spun-off Aerospace Co | $400M | 14x (full pure-play) | $5,600M | Trades at defense sector multiple |
| RemainCo — Chemicals | $250M | 10x | $2,500M | Focused chemicals pure-play |
| RemainCo — Fin. Services | $180M NI | 12x P/E | $2,160M | Potential further separation |
| RemainCo — Real Estate | Asset value | NAV | $800M | Monetizable through REIT spinoff |
| Less: Reduced Corp Overhead | ($400M) | Overhead halved post-Aerospace spin | ||
| Total Post-Spin EV | $10,660M | |||
| Less: Net Debt (allocated) | ($2,200M) | Allocated between SpinCo and RemainCo | ||
| Post-Spin Equity Value | $8,460M | $21.15/share — 32% premium to current |
The spinoff analysis shows that separating Aerospace unlocks $400M in additional equity value ($21.15/share vs. $20.15/share SOTP), primarily because corporate overhead decreases and the remaining entity is a cleaner story. The real value creation, however, is in eliminating the conglomerate discount: if the Aerospace spinco trades at its full 14x pure-play multiple instead of being dragged down by the group's blended discount, that's the $4.15/share gap between SOTP and current trading price being recovered.
SOTP can overstate value in several situations. First, deep operational integration: if the Aerospace and Chemicals divisions share a manufacturing campus, procurement contracts, and key personnel, separating them isn't free — there are stranded costs and dis-synergies that the analysis needs to capture. A breakup isn't just addition of individual EVs; it has a cost. Second, when EBITDA allocation between divisions is arbitrary: many conglomerates allocate corporate overhead across divisions using revenue-based formulas that have little economic logic; if Aerospace is absorbing $60M in allocated overhead that would actually disappear in a breakup, the segment EBITDA is understated and the separation value is cleaner than it appears. Third, when corporate overhead truly can't be eliminated: if the CEO, CFO, and legal team are genuinely running all four businesses with no redundancy, the overhead stays in RemainCo and the SOTP calculation is accurate as modeled. Fourth, illiquid non-operating assets: private equity stakes and real estate portfolios are often marked at book or cost, and realizing their fair market value requires a transaction with timing uncertainty and discount for illiquidity. Always note these constraints in your SOTP presentation.
Interviewers asking about SOTP want to know three things: (1) when you'd use it, (2) how you mechanically build it, and (3) what the output tells you. Here's how to answer "walk me through a SOTP analysis" in 60 seconds:
"SOTP is appropriate when a company has divisions operating in distinct industries where a single blended multiple would distort value — classic case is a conglomerate. You start by identifying each segment and its appropriate valuation methodology: EV/EBITDA for industrial or operating businesses, P/E for financial services, NAV for real estate or asset-holding entities. You apply the relevant sector comps multiple to each segment's financials to get segment-level implied EVs. Then you sum those EVs and subtract two things: the PV of unallocated corporate overhead — capitalized at a representative multiple — and net debt. That gives you SOTP equity value, which you divide by diluted shares to get implied share price. Most conglomerates trade at a 10–20% discount to SOTP because of management complexity, capital allocation inefficiency, and investor preference for pure-plays — and that discount is the opportunity an activist would highlight in arguing for a spinoff or breakup. The SOTP tells you what the pieces are worth individually; the gap to trading price tells you how much value a structural change could unlock."
Multiple compounding factors: management generalism is less effective than specialization, capital allocation is less efficient internally than through market mechanisms, investors who want focused sector exposure can't get it cleanly, and transparency is lower. The discount averages 10–20% empirically. The catalyst to close it is usually an activist campaign, a strategic review, or a voluntary spinoff announcement — all of which force the market to re-price individual divisions at full sector multiples.
Standard comps apply one set of multiples to one set of consolidated financials, appropriate when the company is a focused business with a single industry peer set. SOTP applies different multiples from different peer sets to different segment financials, appropriate when the company is a collection of distinct businesses. SOTP is more work but more accurate when the segments would legitimately trade at different multiples as standalone entities.
You can't apply an EV/EBITDA multiple to a negative number, so you need an alternative approach: revenue multiples (EV/Revenue) if the division has meaningful sales and a path to profitability, DCF if you can project the path to positive cash flow, or liquidation/book value if the division is being wound down. In some cases, a loss-making division has negative value in the SOTP — it's a drag on consolidated results and would cost money to exit (severance, lease termination, restructuring charges). A complete SOTP assigns realistic positive or negative values to every segment; leaving negative-EBITDA divisions out of the analysis overstates aggregate value.