Deep value is the most extreme branch of value investing: rather than paying a fair price for a good business, the deep-value buyer pays a cheap price for a statistically cheap one — measured against tangible assets, not future earnings. Its purest expression is the net-net, Benjamin Graham's idea of buying a stock for less than the cash and liquid assets that would be left over after paying off every liability, assigning zero value to factories, brand, or future profits. The core tension is that the very thing that makes a net-net cheap — a failing or hated business — is also what makes it dangerous: the asset value is real, but it can erode (the "melting ice cube") before the market re-rates the stock. Deep value bets that, as a basket, the statistical margin of safety wins despite many individual failures.

How it's calculated / formed

Two measures, both from Graham's Security Analysis (1934) and The Intelligent Investor:

• Net Current Asset Value (NCAV) = Current assets − Total liabilities (including preferred stock). This ignores all fixed and intangible assets entirely. (StableBread; GuruFocus glossary)
• Net-Net Working Capital (NNWC), the more conservative variant, haircuts the assets toward likely liquidation recovery: NNWC = Cash & short-term investments + (0.75 × receivables) + (0.50 × inventory) − Total liabilities. (StableBread; Seeking Alpha)

A stock is a net-net when its market capitalization is below NCAV (it trades below liquidation value). Graham's actual buy rule was stricter: buy at no more than two-thirds (≈67%) of NCAV, building in roughly a one-third margin of safety, and hold a diversified basket — Graham ran dozens at a time, not concentrated bets. (StableBread; Nasdaq/GuruFocus testing series)

A related, broader deep-value metric used by modern practitioners is the Acquirer's Multiple (EV / operating earnings) popularized by Tobias Carlisle, which extends the cheap-and-hated logic to companies that are not strictly net-nets.

How it's used in practice

Net-net investing is mechanical and basket-driven by design. The practitioner screens for stocks under NCAV (or, conservatively, under NNWC), removes obvious frauds, businesses burning cash faster than the asset cushion can absorb, and Chinese reverse-mergers and other reporting-integrity red flags, then buys a diversified portfolio — often 20–30 names — and rebalances annually. The exit thesis is one of three: a market re-rating toward asset value, a return of profitability, or a hard catalyst (liquidation, buyout, activist).

Because qualifying companies are almost always micro- and nano-caps, the strategy is structurally inaccessible to large funds — a key reason any edge may persist. Most net-nets today are found outside the US (Japan has historically had many; also Korea, the UK, and other markets), and US net-nets cluster around recessions and bear markets, becoming nearly extinct in bull markets. (Quant Investing; Carlisle/Mohanty/Oxman) Graham himself emphasized diversification over judgment here — the method explicitly substitutes a statistical edge for the ability to analyze any single troubled business correctly.

Adoption, debate & evidence

The historical backtest record is unusually strong for a simple rule:

• Oppenheimer (1986), "Ben Graham's Net Current Asset Values: A Performance Update," found stocks bought at ≤⅔ NCAV returned roughly 29.4% mean annual return over 1970–1983 vs. ~11.5% for the NYSE-AMEX index.
• Carlisle, Mohanty & Oxman (2010), "Ben Graham's Net Nets: Seventy-Five Years Old and Outperforming," reported a mean monthly return of 2.55% for NCAV stocks over Dec-1983 to Dec-2008, vs. 0.85% (NYSE-AMEX) and 1.24% (small-firm index).
• Xiao & Arnold (Salford, 2008) tested London 1980–2005: a one-year NCAV holding returned ~31.2% vs. ~20.5% for the market.
• A 2026 study (Mohanty, Review of Financial Economics) extends the US NCAV record into recent decades and reports that the strategy continues to generate positive excess returns, with the value-weighted portfolio commonly cited around ~1.9% monthly (exact figure as reported in the paper; not independently re-verified here), still beating benchmarks.

The honest caveats matter as much as the headline numbers. These returns are concentrated in tiny, illiquid stocks where bid-ask spreads and market impact can erode much of the paper return — Alpha Architect and others stress that the realizable edge is smaller than the backtest. Sample sizes in any given year are small and survivorship/data-quality issues plague microcap datasets. Crucially, the outperformance is a basket result driven by a few big winners against many duds; individual net-nets fail frequently. It is genuinely debated whether the premium is a free lunch or compensation for illiquidity, distress, and tail risk — i.e. a risk premium, not an anomaly. The strategy is widely respected and lightly practiced; it is not "folklore," but its real-world capacity is severely limited by the microcap universe.

Strengths & limitations

When it works: as a diversified, mechanical, contrarian basket in beaten-down small caps — especially after broad sell-offs when net-nets are plentiful and quality is mispriced indiscriminately.

When it fails: the dominant failure mode is the value trap / melting ice cube — a company whose current assets shrink (cash burn, inventory write-downs, uncollectable receivables) faster than the discount closes. Buffett's central critique applies precisely here: in his 1989 Berkshire letter he likened cigar-butt stocks to a discarded butt with one puff left — the bargain makes "that puff all profit" but it is only one puff — and warned that "time is the friend of the wonderful business, the enemy of the mediocre," so a low-ROE business held long destroys the asset bargain. Other limitations: net-nets are scarce in bull markets; the universe is too small and illiquid for most capital; reported balance-sheet values may overstate true liquidation recovery (hence NNWC haircuts).

The #1 misuse: treating it as a concentration strategy. Buying one or two net-nets on conviction discards the entire statistical premise — the edge lives in the basket, not the stock.

Sources

• StableBread, "How to Apply Benjamin Graham's Net-Net Stock Valuation Strategy" — NCAV/NNWC formulas, two-thirds rule.
• Seeking Alpha, "Exploring Graham's Net-Net Working Capital Strategy" — NNWC haircuts.
• Oppenheimer, H. (1986), "Ben Graham's Net Current Asset Values: A Performance Update," Financial Analysts Journal.
• Carlisle, Mohanty & Oxman (2010), "Ben Graham's Net Nets: Seventy-Five Years Old and Outperforming."
• Xiao & Arnold (Salford Business School), "Testing Benjamin Graham's Net Current Asset Value Strategy in London."
• Mohanty (2026), "Does Ben Graham's net current asset value investing continue to generate excess returns?", Review of Financial Economics (Wiley).
• Alpha Architect, review of Graham's NCAV / net-net strategy — illiquidity and realizability caveats.
• Warren Buffett, Berkshire Hathaway Shareholder Letter (1989) — cigar-butt critique; The Acquirer's Multiple summaries.
• Quant Investing, "Why and how to implement a net-net investment strategy world-wide" — geographic distribution, scarcity.

Dispute flagged: whether net-net outperformance is a true anomaly or compensation for illiquidity/distress risk is unsettled; backtested returns substantially overstate realizable returns after microcap transaction costs.