Francisco García Paramés

As of 2026-07-18, Francisco García Paramés is chairman and chief investment officer of Cobas Asset Management, a CNMV-registered Spanish management company. Cobas describes him as a self-taught follower of Buffett, Graham, Lynch, and Austrian business-cycle theory; CNMV records list his Cobas board and executive roles from 2017. This file was researched while the A-profile task for this investor was still freshly claimed and absent from main, so it reconstructs the philosophy directly from primary Cobas materials, CNMV records, interviews, book-related sources, and adverse performance/legal coverage (Cobas bio; CNMV Cobas administrators).

Core worldview

Paramés' worldview is orthodox long-term value investing with an Austrian-economics overlay. The practical core is simple: equities are fractional ownership interests in real businesses, market quotations are often unstable opinions, and long-term purchasing power is best protected by owning businesses at prices below conservative estimates of intrinsic value. Cobas states the doctrine as buying below intrinsic value, waiting patiently, and letting business value converge with market price over time (Cobas philosophy).

The Austrian element matters because it shapes what he distrusts. Paramés is skeptical of monetary debasement, credit-cycle precision, and top-down market forecasts. In his 2017 letter launching Cobas, he argued that the philosophy was unchanged from his prior career: buy undervalued assets to preserve and increase purchasing power, avoid short-term market prediction, and focus on long-term performance of well-priced companies that the team understands. He explicitly contrasted this with products driven by asset gathering or short-term prediction (2017 Paramés letter).

The resulting worldview is not "cheap stocks always win." It is closer to business-owner arbitrage: identify companies whose durable earnings power, asset value, or strategic position is worth materially more than the market price, then finance the wait with patient capital. Cobas defines higher value as coming from competitive advantages, entry barriers, or hidden assets, and lower price as coming from inefficiency, unfashionable sectors, or little analyst tracking (Cobas philosophy).

Two boundaries keep the philosophy from becoming generic optimism. First, he does not treat volatility as the same thing as risk; risk is permanent capital loss. Second, he accepts that a quoted stock can remain wrong for years, so the investor's capital base, communication, fees, and temperament must be built for delay. In a 2011 interview, he described Bestinver clients as co-investors and recommended that they invest only money not needed for at least five to seven years (Value Research interview).

The Edge - What Markets Misprice And Why

Paramés' edge begins with the belief that markets overreact to discomfort and under-price dull, complex, or temporarily unpopular businesses. Cobas' own philosophy page names the habitats: old-fashioned companies, low-tracking companies, inefficient markets, businesses with hidden assets, and businesses whose competitive advantages or entry barriers are not reflected in the price (Cobas philosophy).

The most natural geography for that edge has been Europe, especially Spain and other underfollowed listed markets. At Bestinver, the record was built on the idea that small and mid-sized European businesses, family-controlled companies, and companies outside global growth narratives were less efficiently priced than large U.S. compounders. Value Invest London reports that the Bestinver Spanish funds compounded at 15.7% annually from January 1993 through September 2014, and that Bestinver Internacional compounded at 10.58% annually from January 1998 through September 2014 versus 2.9% for MSCI World. Those are secondary conference-profile figures, not a complete audited composite, but they explain why the European inefficiency thesis became central to the Paramés canon (Value Invest London profile).

The mispricing engine is behavioral and institutional. Investors extrapolate recent earnings, panic in drawdowns, avoid illiquidity, penalize cyclicality at the trough, and prefer fashionable growth when low rates reward duration. Cobas' 2020 stress-period materials described market prices as moved by forced selling, panic, passive flows, machine activity, and liquidity needs, while business value still depended on future profit generation (Cobas COVID letter).

The edge is also informational. Cobas' 2025 commentary says its valuation work combines annual accounts with qualitative work from conferences, specialist providers, interviews with customers, competitors, regulators, and former employees. That evidence stack fits the kind of companies the strategy tends to favor: family-controlled firms, cyclicals, commodity-linked businesses, holding companies, retailers, industrials, and companies whose normalized earnings power is obscured by current conditions (Cobas 2H 2025 commentary).

The risk is that the same habitats also contain value traps. A business can look cheap because its industry is structurally impaired, its balance sheet is fragile, or normalized earnings are overstated. The Aryzta and Teekay episodes, the early Cobas drawdown, and the 2017-2020 value-factor drought are important because they show that the edge depends on distinguishing temporary aversion from permanent impairment (AQR paper on Spain's value investors; Cinco Días on Aryzta).

Process: Idea Sourcing -> Research -> Valuation & Entry -> Sizing -> Portfolio Construction -> Sell Discipline

Idea sourcing. The search starts with accumulated company knowledge, screens, prior holdings, value-manager networks, sector reading, and companies neglected by mainstream investors. Cobas' public materials are unusually explicit that daily research work is company-first rather than macro-first. The firm describes value investing as a painstaking, research-heavy process that starts with individual company analysis and then works upward to portfolio construction (Cobas 1H 2024 commentary; Cobas glossary).

Research. The company analysis has two layers. The quantitative layer reconstructs normalized earnings, returns on capital, balance-sheet resilience, debt maturity, cash conversion, and asset value across a full economic cycle. The qualitative layer asks whether the business is understandable, whether the economics are durable, whether management and controlling shareholders are aligned, and whether competitors, customers, regulators, or former employees corroborate the thesis. Cobas' 2025 commentary explicitly describes using annual accounts, company conferences, specialist providers, and interviews across the business ecosystem (Cobas 2H 2025 commentary).

Valuation and entry. Cobas discloses a mixed valuation toolkit: multiples, discounted cash flow where appropriate, normalized profit through a full cycle, and discount rates usually between 6% and 12%. The output is a target price, an estimated upside, and a margin of safety. Cobas reported at year-end 2025 that its International portfolio traded around 6.9x expected 2026 earnings versus 15.9x for the benchmark, with estimated upside of 106% and ROCE of 27%; those are internal estimates, but they show how the process operationalizes "cheap quality" (Cobas 2H 2025 commentary).

Paramés also has an entry-price discipline that is more concrete than generic margin of safety. His 2017 letter discussed avoiding paying high multiples for uncertain growth and framed the initial Cobas portfolio around earnings yield, ROCE, net cash, and family or reference shareholders. At launch, Cobas described a 62-security portfolio, about 20% commodities exposure, around 13.4% earnings yield, roughly 8.4x PER, and 38% of the portfolio in net-cash companies. Those launch metrics matter less as timeless numbers than as evidence of the checklist (2017 Paramés letter).

Sizing. Sizing is driven by upside, business quality, conviction, liquidity, company size, and diversification rules. The strategy is not a ten-stock partnership; Cobas often runs dozens of names while allowing the top positions to matter. At year-end 2024, Cobas' published portfolio summary showed top-ten weights totaling 36.1%; at year-end 2025 the International and Iberian portfolios still showed broad portfolios with large but not dominant top holdings (December 2024 Cobas portfolio commentary; Cobas 2H 2025 commentary).

Portfolio construction. Paramés generally wants equity exposure high because the philosophy sees cash and fixed income as poor long-term stores of purchasing power in a debasing-currency world. The 2017 letter said Cobas would usually be close to the legal equity limit and did not intend to take fixed-income or money-market risk as a substitute for equity ownership. Product documents reinforce the long horizon: Cobas Internacional is classified as value style, can lose capital, and is not recommended for horizons under five years (2017 Paramés letter; Cobas Internacional page).

The portfolio is also capacity-sensitive. In March 2026, El Español reported that Cobas was working with CNMV to soft-close Cobas Selección and Cobas Internacional near EUR1.5 billion each, with the stated aim of preserving the ability to generate long-term value, especially in smaller companies. That is secondary reporting, not a final CNMV filing, but it is consistent with the capacity discipline Paramés discussed at Cobas' launch (El Español soft-close report; 2017 Paramés letter).

Sell discipline. The sell rule is opportunity cost. Cobas sells or trims positions whose safety margins narrow and reallocates to businesses with better risk-adjusted upside. The 2025 commentary gives direct process evidence: after strong returns narrowed safety margins, Cobas sold or reduced companies that had re-rated and reinvested in names with larger discounts. The International portfolio exited 20 stocks representing about 21% of end-2024 weight and bought 20 new companies representing about 25% of end-2025 weight (Cobas 2H 2025 commentary).

Risk Management

Paramés' risk model is not low volatility. It is avoidance of permanent capital loss. Cobas says volatility creates opportunity, while real risk comes from paying too much, misunderstanding the business, owning weak balance sheets, or losing patience before value realization. This is why the process stresses conservative normalization, balance-sheet strength, ROCE, margin of safety, and long holding periods (Cobas philosophy; Cobas glossary).

Risk management has three practical layers. The first is business risk: avoid leverage, accounting fragility, secular decline, poor incentives, and companies whose normalized earnings cannot be reconstructed. The second is price risk: insist on a margin of safety and compare every holding against the best current alternatives. The third is capital-base risk: educate investors, discourage short-horizon money, align fees with longevity, and avoid building asset size beyond the opportunity set. Cobas' fee structure lowers management fees for longer-tenured investors, which directly supports patient capital (Cobas fees).

There is also a regulatory and mandate layer. CNMV records show Cobas Asset Management as a registered Spanish management company and Cobas Internacional FI as a CNMV-registered fund, with the latest prospectus dated 2026-06-22. Product pages warn that equities can lose capital and that the fund is unsuitable for short investment horizons. Those constraints are not philosophy, but they define the legal wrapper in which the philosophy is executed (CNMV Cobas AM register; CNMV Cobas Internacional; Cobas Internacional page).

The adverse evidence keeps the risk model honest. AQR's 2020 study found that Cobas Selección's November 2016-June 2020 record was sharply negative relative to benchmark, with negative Sharpe ratio and high volatility; AQR connected the pain partly to negative value and size premia. Cobas' own later materials show a strong recovery by 2024-2025, but the early drawdown remains evidence that cheap cyclical and underfollowed stocks can punish clients long before value converges (AQR paper on Spain's value investors; Cobas LUX International page; Cobas 2H 2025 commentary).

Temperament & Psychology

The required temperament is patient, independent, and comfortable looking wrong. Paramés' own book-related materials and interviews emphasize reading, self-education, independence from academic formulas, and emotional resilience through losses. Fundación Rafael del Pino's summary of his book presentation framed value investing as common sense, psychological strength, and preference for businesses where incentives and horizons align, especially family-controlled firms (Fundación Rafael del Pino event; Google Books: Investing for the Long Term).

Client temperament matters as much as manager temperament. The Value Research interview stressed communication, education, honesty, co-investment, and the need for clients to have a five-to-seven-year horizon. That is not soft branding; it is part of the strategy's operating system. If clients redeem during a value drought, the fund can be forced to sell exactly when the philosophy says prices are most attractive (Value Research interview).

The darker side of the temperament is thesis inertia. The same independence that lets a manager buy hated businesses can become stubbornness when business deterioration is real. Aryzta and Teekay became public examples of this tension: they fit the contrarian, cheap, crisis-discount pattern, but also exposed the portfolio to leverage, execution, and cyclical assumptions that took time to resolve and damaged early Cobas trust (Cinco Días on Aryzta; 20Minutos/La Información on Aryzta and Teekay).

Evolution Over Career

Paramés' philosophy has evolved less in label than in emphasis. At Bestinver, he built a European value record that media and conference profiles associate with deep independent research, family-controlled companies, and underfollowed European equities. Over time, secondary interview and book-related sources describe a move from cheaper, Graham-style securities toward more Buffett/Fisher-style attention to business quality, management, and long-term reinvestment capacity (Value Research interview; Value Invest London profile).

The Cobas phase added a new institutional problem: rebuilding a process after a celebrated founder-manager exit. Cobas' early materials leaned heavily on continuity with Bestinver principles, but later regulatory and industry coverage shows a shift away from "fondo de autor" framing toward a more explicitly team-based process. Cinco Días reported in 2024 that Cobas funds would no longer be legally associated as "funds of author," with the firm saying the change reflected a collective, structured process while Paramés continued to lead investment management (Cinco Días on fund-author removal).

The process has also absorbed modern constraints. Cobas integrates ESG as a risk/opportunity variable rather than as a simple exclusion list, having hired an ESG provider in 2019 and signed the PRI in January 2021. That does not turn Paramés into an ESG-first investor; it shows that the old value process now has to account for governance, regulatory, social, and environmental risks that can impair cash flows or terminal value (Cobas ESG approach).

Performance evolution is uneven. The Bestinver-era record was exceptional by secondary and independent accounts; the early Cobas era was poor; the 2021-2025 rebound was strong. Cobas LUX International lists annual returns of -31.48% in 2018, -26.39% in 2020, then +33.59% in 2021, +23.51% in 2024, and +25.24% in 2025. That path is important because it shows the philosophy's regime dependence rather than a smooth personal-alpha line (Cobas LUX International page).

What They Explicitly Reject

Paramés rejects short-term market forecasting. The 2017 letter says the team does not try to predict short-term market behavior and instead focuses on the long-term performance of stable, well-priced companies it understands (2017 Paramés letter).

He rejects fixed income and cash as default long-term stores of purchasing power, especially when monetary debasement is the background risk. This is why Cobas tends to remain highly invested in equities and treats temporary cash as frictional rather than strategic market timing (2017 Paramés letter; Cobas Internacional page).

He rejects fashionable valuation narratives that require heroic growth or low discount rates to work. Cobas' 2019 and 2020 materials pushed back against expensive growth expectations during the value drought, although those same letters should be read as partly defensive because the funds were underperforming at the time (Cobas Q4 2019 commentary).

He rejects scale for scale's sake. Capacity limits appear in the 2017 letter and in 2026 reporting on possible soft closures. The logic is that excess assets dilute access to smaller, less efficient companies and force the portfolio toward more efficient large-cap territory (2017 Paramés letter; El Español soft-close report).

Finally, he rejects the idea that investor money is interchangeable. The strategy wants patient, educated clients, not short-horizon performance chasers. Cobas' fee schedule, long-horizon product warnings, and Paramés' own interview comments all reinforce that client selection is a risk-control variable (Cobas fees; Value Research interview).

Regimes Where It Thrives Vs. Struggles

The philosophy thrives when dispersion is high, value is out of favor, panic creates forced or indiscriminate selling, and underfollowed businesses can be bought with clear balance-sheet protection. It also works best when inflation, monetary uncertainty, or currency debasement make ownership of productive assets more attractive than nominal fixed claims. The 2021-2025 Cobas rebound, especially after the early value drought, is a recent example of how returns can arrive in clusters after long waiting periods (Cobas LUX International page; Cobas 2H 2025 commentary).

It struggles in markets led by expensive growth, low interest rates, passive flows into mega-cap benchmarks, and narratives that extend duration far into the future. It also struggles when the "cheap" cohort is cheap for structural reasons: excess debt, weak industry economics, technological disruption, commodity overcapacity, accounting opacity, or shareholder-unfriendly governance. The early Cobas record from 2017 through 2020 is the clearest stress test of this regime problem (AQR paper on Spain's value investors; Cobas 5th annual conference page).

It also struggles when the client's clock is shorter than the thesis clock. A strategy that requires five to seven years of patient capital cannot be judged honestly by a one-year ranking, but a client who cannot finance that horizon experiences the drawdown as real risk. This is why the client-base design is not an accessory; it is part of whether the philosophy is executable (Value Research interview; Cobas Internacional page).

Tensions Between Stated Philosophy And Actual Behavior

The first tension is Bestinver versus Cobas. The Bestinver-era record supports the philosophy, but the early Cobas record challenged it. AQR found severe underperformance for Cobas Selección through June 2020; Spanish press coverage in 2019 and 2021 reported large losses and peer underperformance. Later Cobas data show a strong rebound, but the canon should not collapse the two eras into a single heroic line (AQR paper on Spain's value investors; Europa Press 2019 report; Cobas 2H 2025 commentary).

The second tension is bottom-up investing with a strong macro worldview. Paramés says he does not forecast markets, but his Austrian framework clearly affects asset-class preference, aversion to fixed income, and interpretation of monetary policy. That is not hypocrisy; it is a boundary issue. The philosophy is bottom-up in stock selection, but not neutral about the monetary environment (Cobas bio; 2017 Paramés letter).

The third tension is patience versus stubbornness. The sell rule says rotate out when safety margins narrow or theses break. The adverse cases show how hard that is in practice when a cheap, hated company also has leverage or operating deterioration. Aryzta and Teekay were not just unlucky marks; they were process stress tests for normalized-earnings estimates and the willingness to change one's mind (Cinco Días on Aryzta; 20Minutos/La Información on Aryzta and Teekay).

The fourth tension is key-person aura versus team process. Bestinver's 2014 departure led to CNMV investor complaints about whether investors should be able to redeem without penalties because Paramés had left. Later Cobas documents and 2024-2025 industry coverage moved away from the "fund of author" structure. The doctrine now needs to be evaluated as a team process led by Paramés, not as a purely personal account (CNMV 2014 complaints report; Cinco Días on fund-author removal).

The fifth tension is legal/commercial conflict versus investor-philosophy legend. The Bestinver/Acciona disputes were employment and compensation matters, not securities-fraud findings. Cinco Días reported that Spain's Supreme Court ordered Bestinver/Acciona to pay Paramés EUR13.2 million in variable remuneration while rejecting a larger claim and non-compete compensation. This matters because it should be kept separate from the investment philosophy and from client harm narratives (Cinco Días Supreme Court report).

The final tension is that the philosophy is more transferable as a question set than as a stock list. A student can copy the questions: Is the business understandable? What is normalized earnings power? What would permanently impair capital? Why is the seller mispricing it? How patient is the capital? What is the opportunity cost? But copying Cobas holdings without the evidence stack, client horizon, liquidity tolerance, and sell discipline would miss the actual system.

As of 2026-07-18, the best-supported answer to the "single best" question is Construcciones y Auxiliar de Ferrocarriles (CAF). A surviving recap of Bestinver's 2011 investor conference reports Francisco Garcia Parames calling CAF probably the firm's best investment in 20 years. The original conference-video page still corroborates the event, but its obsolete recording cannot be checked; the designation is therefore high-confidence secondary evidence, not an auditable trade ledger (2011 conference recap; original video page).

This chapter ranks ten documented successful holdings, including completed exits and Bestinver positions whose Parames-era endpoint is his September 2014 departure. It favors the strength of the manager's designation, portfolio importance, reconstructable thesis, adversity endured, and evidence of realization; it does not rank normalized returns. Bestinver and Cobas publish portfolio snapshots, contribution figures, and selected holding returns, but not security-level cash ledgers. Absolute profit, money-weighted return, and maximum cost-basis drawdown are unavailable for every case below. Results belong to Bestinver- or Cobas-managed vehicles and their investment teams, not to a verified Parames personal account. Alvaro Guzman joined Bestinver in 2003 and Fernando Bernad in 2007; Cobas now explicitly presents investment as a team process (CFA Institute, 2013; Cinco Dias, 2024).

The currencies and measures above are those used by each source. A share-price move is not a fund return: weights change, dividends and foreign exchange matter, and different vehicles can own the same security at different sizes.

1. CAF - the Single Best by Parames's Own Qualitative Ranking

Context, thesis, size and discovery. Bestinver publicly named the Spanish rail-equipment manufacturer as a preferred holding by January 2003. Parames then emphasized good managers, low prices, and strong dividend yields rather than a short-term rail-cycle forecast. A February 2006 Europa Press report says Bestinver-managed collective-investment vehicles had accumulated 10.924% of CAF in its body, while its headline says 10.294% [disputed]; the underlying CNMV notice was not located. Either is an issuer ownership percentage across vehicles, not a fund weight or Parames's personal stake (El Pais, 2003; Europa Press/CNMV report, 2006).

Path, exit and P&L. By the 2011 conference Bestinver was reducing CAF as the valuation gap narrowed, although Parames reportedly still considered it an exceptional company and undervalued. Public sources do not disclose weighted entry price, dividends received, maximum drawdown, final sale date, proceeds, or absolute and percentage P&L. The exceptional-result designation is therefore the evidence for rank one; it is not a reconstructed return claim. The surviving conference recap is one step removed from the inaccessible recording, so even the label warrants a provenance warning (2011 conference recap).

What it teaches. The defensible lesson is patient accumulation in an illiquid, well-managed industrial franchise, followed by reduction as price approached value. It does not prove that concentrated ownership alone created the outcome or that the same economics were available to a smaller investor.

2. BMW - Buying a Quality Cyclical on Recession Earnings

Context, thesis and size. BMW was Bestinver's largest position in November 2008. At EUR19.15 for the common shares, Parames estimated value above EUR90 and a stressed-margin value near EUR40 [single-source direct interview]. The thesis combined a premium brand, resilient luxury-market share, family control, capital strength, and manufacturing-efficiency upside. Bestinver also used preferred shares, whose discount added a separate margin of safety. A May 2010 report put combined Bestinver exposure above EUR156 million; Bestinver Internacional alone held EUR73.4 million, or 8.14% of its portfolio [single-source press report] (Value Investor Insight archive, 2008; Cinco Dias, 2010).

Path, drawdown and realization. The 2010 CNMV report shows BMW at 8.74% of Bestinver Internacional and the shares up 67% that year [single-source regulatory filing]; by June 2014 a separate filing put the position at 7.28% [single-source regulatory snapshot]. Common and preferred prices cannot be spliced into one return series, and the filing weights are snapshots rather than average capital. BMW was still a top holding when Parames departed in September 2014, so there was no Parames-era final exit or disclosed absolute P&L (CNMV Bestinver Internacional 2010 report; CNMV first-half 2014 report). In 2011 Bernad called BMW one of the team's best investments; Parames later placed it in the admired four-stock terminal portfolio with Exor, Thales, and Wolters Kluwer (Cinco Dias, 2011; Cinco Dias, 2016).

What it teaches. Separate temporary cyclical earnings from franchise value, but also separate a good security call from a completed trade. Later BMW appreciation cannot be credited to a mandate that had ended.

3. Ferrovial - Sum of the Parts in a Financing Panic

Context, thesis and size. Ferrovial was one of Bestinver's two largest ideas during the financial crisis. The thesis valued toll roads, Heathrow/BAA, services, and construction separately, rather than treating consolidated project debt as ordinary parent-company leverage. A reproduced 2010 Bestinver presentation estimated EUR17.10 per share of value against a price near EUR7.20 [single-source presentation reproduction]; conference notes put it at 9.2% of the Iberian portfolio in 2011 [single-source conference recap] (Bestinver presentation reproduction, 2010; 2011 conference recap).

Path, exit and P&L. Parames's later book-based recollection places the shares near EUR3 at the crisis meeting and around EUR20 later [single-source book-derived recollection]. That is an approximate issuer-price path, not a corporate-action-adjusted Bestinver return. Bestinver had reduced the position by 2013 after strong performance; weighted entry, maximum fund-position drawdown, dividends, final exit, and realized profit remain unavailable (CFA Institute, 2013; MOI Global book notes).

What it teaches. Complex infrastructure groups can become mispriced when investors collapse ring-fenced project finance, holding-company debt, and operating assets into one alarming leverage number. Favorable refinancing conditions and asset markets were also necessary; valuation skill did not control those outcomes.

4. Thales - Enduring the Margin Repair

Context, thesis and size. Bestinver built most of its Thales position during 2010 below EUR30 [single-source manager retrospective]. It saw a defensible aerospace-and-defense franchise whose margins had collapsed because of poor project execution, not a vanished moat. Dassault Aviation's strategic holding and project-management expertise offered a plausible catalyst. Thales was 4.47% of Bestinver Internacional at year-end 2010 and 4.99% in June 2014 [single-source regulatory snapshots at each date] (CNMV 2010 report; CFA Institute, 2013; CNMV first-half 2014 report).

Path, drawdown and exit. The shares fell 27% in 2010 [single-source regulatory filing], then dropped below EUR22 in December 2011 [single-source manager retrospective]. Parames later included Thales in the four-stock terminal Bestinver basket whose prior five- or six-year performance he called extraordinary; that is a qualitative manager designation, not a return ledger. A later Cobas retrospective points to a price above EUR110 in 2018 [single-source], but that endpoint is nearly four years after Parames left Bestinver. The position remained large at departure; no mandate-period sale or exact P&L was located (CNMV 2010 report; Cinco Dias, 2016; Cobas second-quarter 2018 commentary).

What it teaches. A margin gap can be repairable when customer relationships and technical barriers persist, but the long recovery also shows why subsequent issuer performance is not proof of the original fund's realized return.

5. Smurfit Kappa - the Clearest Bestinver Round Trip

Context, thesis and structure. Bestinver began buying the packaging producer in 2007 and continued below EUR1 in late 2008. At a sub-EUR2 share price, the team estimated about EUR950 million of normalized EBITDA and EUR300 million of free cash flow against roughly EUR390 million of market capitalization; debt was about EUR3 billion [single-source direct team interview]. The leverage made asset coverage, maturities, and dilution risk as important as normalized earnings. Bernad articulated much of the detailed thesis; this was a team trade (Value Investor Insight archive, 2008).

Path, exit and P&L. Smurfit Kappa fell about 93% from its EUR16.50 March 2007 IPO price to roughly EUR1 in October 2008; Bestinver later sold its final shares at EUR8 [single-source manager retrospective]. That does not establish a 700% portfolio return: earlier lots cost more, and no weight, weighted basis, dividends, absolute profit, or cost-basis drawdown is public (Cobas second-quarter 2018 retrospective).

What it teaches. Averaging down in a highly levered cyclical can succeed only if normalized cash generation and refinancing survival are right. It is also a warning against turning the best late lot into the return for the whole position.

6. Aldeasa - Patient Accumulation, Strategic Liquidity

Context, thesis and size. Bestinver accumulated the airport retailer for five or six years. Parames described a simple business, capable management, an attractive price, and dividends; it was a preferred holding by 2003. He later cited a peak 9% ownership stake [single-source manager recollection], while a Spanish competition decision records a 5.217% indirect Bestinver interest before the takeover [single-source regulatory record]. Those figures likely reflect different dates or aggregation and should not be forced to match (El Pais, 2003; Cinco Dias interview, 2005; Spanish competition decision).

Path, exit and P&L. No weighted entry price or maximum drawdown was found. Autogrill's successful offer closed in April 2005 at EUR36.57 per share [single-source acquirer release], providing a clear corporate exit but not a Bestinver cash ledger. The number of shares tendered, prior sales, distributions, absolute proceeds, and percentage P&L remain unverified (Autogrill offer result, 2005).

What it teaches. Illiquidity can be acceptable when a manager can build patiently and the business compounds, but the strategic buyer was an external source of liquidity and favorable timing, not a repeatable guarantee.

7. Elecnor - Cobas's Strongest Completed Total-Shareholder Outcome

Context, thesis and size. Cobas held Elecnor from its 2017 launch and says the team had followed it for more than 30 years. The thesis separated renewable developer Enerfin, transmission-platform investment Celeo, and engineering operations that the market valued poorly together. Cobas Seleccion reported a 1.08% weight at year-end 2017, while Cobas Iberia had 6.4% in June 2024 [single-source regulatory/manager snapshots at each date]. These are different vehicles and dates, not one continuous size (Cobas Seleccion second-half 2017 report; Cobas first-half 2024 commentary).

Path, catalyst, exit and P&L. A manager chart shows the shares roughly in a EUR10-EUR12 band for much of 2017-2022 [single-source] while asset value developed. Enerfin's sale then surfaced value; Elecnor reports a EUR1.560 billion transaction price received and EUR1.0149 billion of proceeds after associated tax [single-source issuer disclosure]. A Cobas Seleccion chart reported +111% through April 2024 [single-source; source share-class label/ISIN inconsistent]. Cobas later fully liquidated Elecnor in 2025 after what it described as an almost fourfold price increase including nearly EUR10 per share of dividends [single-source manager calculation]. That is a stock-plus-dividend narrative, not audited fund P&L or a fourfold NAV contribution (Elecnor AGM document, 2026; Cobas first-half 2024 commentary; Cobas second-half 2025 commentary).

What it teaches. A flat price can coexist with rising asset value, and disposals plus dividends can close a conglomerate discount. Six years of dead money was the main observable adversity; exact maximum drawdown is unavailable.

8. Exmar - the Highest Published Completed Cobas Seleccion Chart Return

Context, thesis and size. Cobas introduced Exmar in the fourth quarter of 2017. Cobas Seleccion held 0.57% at year-end 2017 [single-source fund filing]; by April 2023 Cobas-managed vehicles and entities held about 5.94% of Exmar's shares [single-source manager statement]. Those are a fund weight and aggregate issuer ownership, not comparable sizes. The thesis centered on discounted LNG and shipping assets, family control, and corporate actions that could reveal value. The path was punishing: a secondary price series shows annual declines in 2018 and 2019, a roughly 43% fall in 2020, then gains in 2021-2023 [single-source secondary price history]. Those calendar moves are not the fund's cost-basis drawdown (Cobas Seleccion second-half 2017 report; Cobas bid statement, 2023; CompaniesMarketCap price history).

Exit and P&L. Saverex offered EUR12.10 before the dividend in 2023. Cobas said that price was a good reflection of fair value and that it would sell its approximately 5.94% stake on those or similar terms [single-source manager statement]; the prospectus independently confirms the offer price and conditions. By April 2024 Cobas had exited. A Cobas Seleccion chart reported +220% through the time in the portfolio or 30 April 2024 [single-source; source share-class label/ISIN inconsistent]. It does not state dividend, currency, or cash-flow-weighting conventions and supplies no absolute profit (Cobas bid statement, 2023; Saverex prospectus, 2023; Cobas first-half 2024 commentary).

What it teaches. A family-controlled, asset-heavy security can require years of adverse prices before a corporate event unlocks value. Cobas's willingness to tender when the offer reflected its estimate is evidence of valuation discipline, not activism for a higher price.

9. International Seaways - Shipping Consolidation After a Deep Drawdown

Context, thesis and size. Cobas replaced Euronav with International Seaways (INSW) in the fourth quarter of 2017 because it judged INSW more attractive. Cobas Seleccion held 2.69% at year-end 2017 [single-source fund filing], while an early-2018 newsletter said Cobas vehicles together owned 10.3% of the issuer [single-source manager report] - again, fund weight and issuer ownership are different measures. The thesis combined tanker assets at a discount, moderate leverage, governance, and eventual industry normalization (Cobas fourth-quarter 2017 commentary; Cobas January 2018 newsletter).

Path, merger, exit and P&L. INSW contributed 3.9 percentage points to Cobas Internacional in 2019 [single-source regulatory filing], then its share price suffered a roughly 57.5% peak-to-trough fall from a 2019 high to a 2020 low [single-source secondary adjusted-price series, not a cost-basis drawdown]. In 2021 Cobas supported the all-stock Diamond S merger, citing fleet complementarity, scale, expected synergies, no additional leverage, and INSW management. It later exited; a Cobas Seleccion chart reported +143% through April 2024 [single-source; source share-class label/ISIN inconsistent]. The 2019 Cobas Internacional contribution is not the lifecycle return, and neither figure yields absolute P&L (CNMV 2019 report; Macrotrends price history; Cobas first-quarter 2021 commentary; Cobas first-half 2024 commentary).

What it teaches. Asset discounts can converge through consolidation, liquidity, and a favorable cycle. Extreme volatility, however, makes a shipping outcome especially dependent on balance-sheet survival and cycle timing.

10. Babcock International - Re-underwriting Through a Greater-Than-50% Fall

Context, thesis and size. Cobas owned Babcock from its 2017 launch after the shares had fallen from about GBP13 in 2014 to below GBP10 [single-source manager case study]. It viewed the company as a high-barrier defense-services operator with long contracts and specialized assets, misclassified with troubled generalist outsourcers. Cobas Seleccion reported 3.52% at year-end 2017 [single-source fund filing]. After the stock fell by more than 50% in 2020 while Cobas cut its value estimate by only about 10%, the team increased ownership above seven million shares [single-source manager case study] (Cobas Seleccion second-half 2017 report; Cobas first-half 2025 commentary).

Adverse evidence, exit and P&L. The drawdown was not mere sentiment: Babcock's 2021 contract review produced charges and impairments of about GBP1.7 billion, including an approximately GBP1 billion impairment [single-source issuer update]. Cobas continued only after re-underwriting the business, then began selling as disposals, targets, and valuation improved. A manager case study says the shares rose about 100% in the first half of 2025; Cobas sold the remainder in June at roughly 14 times earnings versus about seven times for its portfolio, and Babcock contributed 2.2 percentage points to Cobas Internacional in that half [single-source manager calculations]. Exact inception return and absolute P&L remain unavailable (Babcock business update, 2021; Cobas first-half 2025 commentary).

What it teaches. Averaging down was defensible only because the team revised its business-value estimate and treated the impairment as evidence, not noise. The long recovery also benefited from a stronger defense-spending environment.

Cross-Case Lessons and Selection Limits

Five patterns recur. First, the largest gains began with a valuation gap that could be explained asset by asset or earnings power by earnings power. Second, family or strategic owners mattered at BMW, Thales, Aldeasa, Elecnor, and Exmar, but only when their incentives produced action. Third, Parames's teams often endured long periods of adverse prices; survival depended on balance sheets, not temperament alone. Fourth, exits were valuation-led or event-led rather than macro forecasts. Fifth, collaboration mattered: later Bestinver and all Cobas cases cannot responsibly be narrated as one man's trades.

The list is retrospectively selected and says nothing about hit rate. It excludes open successes such as Golar LNG, Atalaya Mining, and Tecnicas Reunidas because their final economics remain unresolved. It also excludes Maire Tecnimont and Avio only to keep the chapter within ten cases, despite useful completed outcomes. Aryzta, Teekay, CIR/Cofide, and Debenhams are survivorship controls: the same patience, cyclicality, concentration, and family-company preferences sometimes produced losses or opportunity costs and belong in mistakes-and-losses.md.

The clean conclusion is deliberately narrow. CAF is the strongest surviving answer to Parames's own "single best" designation, Smurfit Kappa is the clearest Bestinver round trip, Elecnor is Cobas's strongest documented completed total-shareholder result, and Exmar has the highest published completed figure in the Cobas Seleccion chart. None has a public transaction-level ledger sufficient to calculate exact absolute profit, money-weighted return, or fund-position maximum drawdown.

Executive finding and evidence boundary

The most defensible adverse record is not a personal trading ledger. It is a sequence of regulated-fund drawdowns, manager-reported position contributions, issuer failures, and retrospective admissions spanning two firms. No public personal-account statements, complete Bestinver or Cobas lot history, or lifetime list of Francisco García Paramés's errors was located. Fund results are team and vehicle results; an issuer's price fall is not the fund's percentage loss; a one-year contribution is not lifecycle profit and loss; and an unrealized drawdown is not a realized loss for an investor who stayed. Those distinctions materially narrow what can be claimed.

Within that boundary, the evidence is unusually candid. An early concentrated Nissan Motor Ibérica position is the clearest reported personal loss. Bestinver's International portfolio later suffered four substantial drawdowns before Paramés left in September 2014. At Cobas, Aryzta became the clearest security-selection and sizing failure; DIA exposed premature entry before operating evidence had stabilized; Teekay showed the danger of correlated exposure to a leveraged cycle; and Valaris was a small exposure to an issuer that entered bankruptcy after Cobas reported exiting. The most severe measured portfolio path was Cobas Selección Class C's 61.52% peak-to-trough fall from 23 January 2018 to 23 March 2020 [single-source official NAV calculation], reconstructed from the fund's official daily NAV export. The same series did not regain its old high until 13 March 2024. This was a six-year recovery from peak, not proof that the management company was near insolvency or that every holder realized the loss. (Cobas Selección daily NAV export)

The strongest process lesson is also the least comfortable: several failures were not violations of Paramés's philosophy but extensions of it. Patience became thesis inertia; normalized earnings obscured balance-sheet and transition risk; trust in a familiar executive substituted for aligned ownership; and conviction allowed one mistake to dominate a diversified portfolio. Cobas did articulate changes—greater caution with manager-run companies, more emphasis on executive ownership, continuous thesis review, and an admission that Aryzta's excessive weight was itself the largest error. Public evidence does not establish a binding postmortem checklist, a new hard position limit, or a complete error scorecard.

The loss record that can actually be measured

Cobas's fourth-quarter 2018 commentary reconstructed four earlier Bestinver International drawdowns under Paramés: 17.5% in 1998, 38.4% in 2002, 61.8% from 2007 to 2009, and 22.9% in 2011 [single-source manager retrospective]. It also reported gains of 78.4%, 64.9%, 153.6%, and 66.0%, respectively, over the two years after each trough. A 61.8% fall followed by 153.6% appreciation still leaves about 3.1% to recover if the endpoints are exactly sequential. These are manager-selected portfolio episodes, not personal losses or independently calculated maximum drawdowns. The document itself identifies the series as Bestinver International and notes that Paramés left on 23 September 2014. (Cobas fourth-quarter 2018 commentary)

Current official factsheets independently anchor the calendar-year damage, a different measure. Bestinfond returned -42.14% in 2008 and Bestinver International returned -44.71% [single-source official factsheet for each vehicle], followed by +60.20% and +71.85% in 2009. Those annual returns corroborate crisis severity but do not reproduce the later 61.8% peak-to-trough figure. (Bestinfond official factsheet) (Bestinver International official factsheet)

The same source makes the 2011 case more informative than a bare drawdown. Cobas described a roughly six-month 22% fall and identified CIR/Cofide—then about 7% of the International portfolio—as a very important valuation error [single-source manager figures]. That is a direct team postmortem, but the disclosure does not provide purchases, sales, dividends, currency effects, or a position-level realized loss. It supports a large valuation mistake inside a bad period, not a clean trade return. (Cobas fourth-quarter 2018 commentary)

The Cobas launch period was worse than the 2018 calendar figure alone suggests. In 2018, Cobas reported -31.4% for the International portfolio against -10.6% for its benchmark and -14.1% for the Iberian portfolio. Its own contribution table attributes about -11.7 percentage points in International to Aryzta and a combined -5.7 points to Teekay Corporation and Teekay LNG. In Iberia, Duro Felguera and DIA cost roughly -2.8 and -2.0 points, respectively [single-source manager figures]. The newsletter also reports issuer-price moves of -86.9% for Aryzta, -64.2% for Teekay Corporation, -45.3% for Teekay LNG, -97.0% for Duro Felguera, and -89.1% for DIA during 2018. The contribution figures are fund-period effects; the price changes need not match Cobas's holding periods or cost bases. (Cobas February 2019 newsletter)

The later Class C NAV path supplies a consistent vehicle-level denominator. Calculated directly from the official daily series, NAV fell from 114.265235 on 23 January 2018 to 43.966360 on 23 March 2020, or 61.52%. It first exceeded the old peak on 13 March 2024, at 114.575434. These are reproducible calculations from a mutable primary spreadsheet, not figures printed by Cobas in a narrative report. They exclude investor cash-flow timing and do not represent firmwide assets, another share class, a personal account, or the maximum loss of every client. (Cobas Selección daily NAV export)

Nissan Motor Ibérica: the formative personal concentration loss

Paramés's book describes an early investment in Nissan Motor Ibérica that grew to about 40% of his personal portfolio before the security lost roughly 40% [single-source book-derived figures]. The surviving public evidence is book-derived rather than an account statement: Wiley confirms the authorized work, while an MOI compilation produced with extensive Cobas collaboration summarizes the episode. It does not disclose every purchase, the portfolio's full starting value, or a broker-confirmed realized euro loss. The defensible conclusion is that a highly concentrated personal position suffered an approximately 40% security loss, not that his entire account fell 40%. (Wiley, Investing for the Long Term) (MOI Global book compilation)

This is the clearest loss-to-rule chain in his record. Paramés presents the episode as the lesson that one apparently compelling security must not dominate the account. Later fund concentration remained meaningful, so “diversification” did not become low-conviction indexing. The safeguard was to prevent one position from controlling personal survival; Aryzta later showed that a 9%-plus weight could still control a professional fund's annual outcome.

Aryzta: the central error was governance plus weight

Aryzta is the best-documented Paramés mistake because Cobas disclosed both diagnosis and portfolio damage. The initial thesis treated a global bakery franchise as a temporarily impaired but valuable business. What broke was not merely the share price. Cobas's third-quarter 2018 review said the first error was trusting a management team with little ownership. Past experience with chairman Gary McGann had been treated as a substitute for a family or owner-operator shareholder; Cobas concluded that it was not. Debt, initially judged manageable, became dangerous when operating costs, a delayed asset sale, and a leadership transition arrived together. (Cobas third-quarter 2018 commentary)

The second error was concentration. By year-end 2018, Aryzta represented 8.6% of Cobas Selección, 9.5% of International, 9.0% of Large Cap, and 16.9% of Concentrated Opportunities [single-source manager figures]. Cobas explicitly accepted that Aryzta was an investment error and the principal reason for underperformance, while still arguing that the operating business could recover. A direct interview that November records Paramés's sharper retrospective: the largest mistake was the excessive weight, and the failure was more in analyzing management than in analyzing the business. That is manager testimony, not an independent causal finding. (Cobas fourth-quarter 2018 commentary) (El Economista interview, 2018)

Cobas opposed Aryzta's proposed €800 million capital increase and advocated an alternative. That activism may have protected the thesis from dilution in the manager's view, but it did not remove the underlying leverage or execution problem. The February 2019 newsletter's -86.9% issuer move and double-digit contribution damage in major funds show how a governance judgment became a portfolio-construction failure. A cheap security can remain analytically attractive while its weight makes one forecast error intolerable. (Cobas third-quarter 2018 commentary) (Cobas February 2019 newsletter)

The exit record prevents a simple “total loss” label. Cobas reduced Aryzta in 2019. In the first half of 2023 it reported trimming the holding after a strong period and fully exiting Aryzta in the Large Cap portfolio, not necessarily every vehicle. The public materials located do not supply a consolidated, dividend- and currency-adjusted lifecycle P&L across all Cobas funds. Aryzta was unquestionably a devastating early detractor and an admitted mistake; the evidence does not prove that every Aryzta position was sold at a total or permanent loss. (Cobas second-quarter 2019 commentary) (Cobas first-half 2023 report)

DIA: buying before the operating evidence changed

DIA supplies the cleanest entry-timing admission. In the fourth-quarter 2018 commentary, Cobas said it entered too early. The team had not waited for either a reduction in management's margin target or a change in management; it believed a roughly 60% year-to-date price fall already incorporated those risks, and later acknowledged that it did not. The annual newsletter reports an 89.1% issuer-price decline and about a 2.0-point negative contribution to Cobas Iberia. (Cobas fourth-quarter 2018 commentary) (Cobas February 2019 newsletter)

This was not simply bad market timing. The process error was allowing price damage to stand in for confirmation that the earnings base, incentives, and leadership assumptions had reset. The lower quotation increased apparent upside while the value input remained unstable. DIA therefore challenges a common value-investing shortcut: a large fall can create a margin of safety only after the denominator—sustainable cash earning power—is credible.

Duro Felguera and Mota-Engil belong in the loss ledger, but only Duro has a documented analytical narrative. Cobas said it supported Duro's recapitalization after consultant and competitor diligence; one month later, work by a new board exposed losses that prior management had underestimated and not disclosed. Cobas exited in 2019. The newsletter attributes about -2.8 Iberian percentage points to Duro and -2.2 to Mota-Engil in 2018. No direct Paramés admission, full transaction history, or lasting Duro-specific rule was located, and no equally clear diagnosis was found for Mota-Engil. (Cobas third-quarter 2018 commentary) (Cobas third-quarter 2019 commentary) (Cobas February 2019 newsletter)

Teekay: correlated cyclical exposure and an incomplete loss story

Teekay shows why the economic unit of risk can be larger than a ticker. In 2018, the parent and Teekay LNG together cost Cobas International about 5.7 percentage points. Teekay Corporation fell 50% in the fourth quarter alone as a January 2020 refinancing approached. Cobas argued that the parent's stakes in listed subsidiaries supported the value and that refinancing risk was manageable. That was the contemporaneous thesis, not proof that the market was wrong. (Cobas fourth-quarter 2018 commentary)

The root risk was correlated: parent discount, subsidiary values, energy-shipping rates, refinancing access, and market risk appetite could deteriorate together. Separate line items did not supply independent diversification. Unlike Aryzta, however, Teekay should not be described as a demonstrated total loss. Cobas reported fully exiting Teekay Corporation in the first half of 2023, after the structure and subsidiary value had evolved, but disclosed no complete cost-basis return. A severe 2018 contribution and a later exit are not enough to reconstruct the lifecycle. (Cobas first-half 2023 report)

Valaris: Cobas exited before the issuer restructured

Valaris is a useful counterexample to the idea that every collapse was temporary volatility. In March 2020, Cobas grouped Valaris with Petrofac, Subsea 7, and Saipem as oil-service holdings totaling about 6% of the international portfolio [single-source manager figure] and argued that permanent business damage was mitigated. A later CNMV half-year filing reported that Cobas International exited Valaris, KT, and Petra Diamonds, whose combined prior weight was below 1%. (Cobas March 2020 announcement) (CNMV Cobas International first-half 2020 report)

Valaris filed for Chapter 11 on 19 August 2020 under an agreement that contemplated canceling existing equity, with limited warrants in specified circumstances, and later said its reorganization eliminated $7.1 billion of debt [single-source issuer figure]. The filing validates permanent impairment for old issuer equity, but Cobas had already reported exiting by the first-half cutoff; it does not reveal the fund's exact sale date, cost, proceeds, or stand-alone loss. Because the three exited holdings together had been under 1% [single-source regulatory filing], the case was small at the reported portfolio level. It still matters: a low weight limited exposure before a cyclical recovery thesis encountered an overleveraged balance sheet and legal restructuring. (Valaris restructuring agreement, 2020) (Valaris restructuring completion, 2021)

McClatchy: structural change arrived faster than the thesis

In a 2009 interview, Paramés acknowledged that he had not foreseen the speed at which newspapers would move from print to the internet in McClatchy. That is a direct structural-disruption admission. McClatchy entered Chapter 11 in February 2020, which validates the danger to the business model but not a Bestinver wipeout: no evidence located shows that the fund retained its old shares until bankruptcy, and no weighted purchase-and-sale ledger supplies its realized loss. (El Confidencial interview, 2009) (McClatchy SEC filing, 2020)

McClatchy is analytically different from an ordinary cyclical drawdown. A return to normalized print economics was not assured because the distribution technology and advertising market were changing. The episode supports a higher burden of proof when “cheap” depends on a legacy profit pool surviving substitution.

In the same 2018 interview used for his Aryzta postmortem, Paramés recalled losing the entire investment in Escada. The vehicle, dates, weight, and absolute amount were not disclosed, so this remains a single-source total-position-loss recollection rather than a measurable portfolio event. (El Economista interview, 2018)

Error of omission: Inditex

Paramés's clearest named omission is Inditex. In a 2016 interview he said not owning it had been an error. He recalled coming close in 2009 after the price fell from about €40 to €25, but other opportunities and limited time displaced the work. This is first-person evidence of a missed investment, not enough information to compute a foregone return: the source gives no intended allocation, exact decision date, hypothetical purchase price, holding period, sale rule, or dividend treatment. (Cinco Días interview, 2016)

The omission reveals a different process constraint from Aryzta. The error was not excessive confidence in completed analysis but failure to allocate scarce research attention to a high-quality business when price improved. It supports a capacity lesson: opportunity cost arises not only from holding the wrong security but from letting the existing pipeline crowd out a potentially superior one. A famous investor's rejected ideas are usually invisible, so one volunteered omission should not be mistaken for a complete counterfactual scorecard.

Was Cobas near death?

No verified corporate or fund near-death event was found. The investment experience was severe: the Class C NAV drawdown exceeded 61%, early relative returns were poor, and trust in a new manager was tested. Yet Cobas's fourth-quarter 2018 report showed approximately €1.856 billion of assets and €264 million of net subscriptions during the year, while its fifth annual conference said March 2020 brought net inflows and that asset continuity through 2020 was close to 97% [single-source manager figures]. These disclosures contradict a claim that a run forced liquidation, closure, or rescue. (Cobas fourth-quarter 2018 commentary) (Cobas fifth annual investor conference)

The accurate category is a portfolio and reputational crisis with liquidity risk for clients who could not wait—not insolvency of Paramés, Cobas, or its funds. The open-end structure meant withdrawals could still have amplified forced sales, and Cobas's March 2020 letter discussed that pressure. But no suspension, rescue financing, liquidation, personal bankruptcy, or inability to meet redemptions was located. (Paramés March 2020 letter)

Behavioral root causes

1. Governance by analogy

Aryzta shows an explicit substitution error: good prior experience with an executive was treated as equivalent to aligned owner oversight. Paramés's general preference for family companies was not cosmetic; it was a risk control against agency problems. Departing from it required more, not less, evidence about incentives, capital allocation, and transition behavior. Cobas's postmortem recognizes this asymmetry. (Cobas third-quarter 2018 commentary)

2. Normalization while the balance sheet moved

Aryzta, DIA, Teekay, and Valaris differ, but all expose the same valuation hazard. A normalized earnings estimate can be slow-moving while leverage, refinancing access, governance, or industry structure changes quickly. A large price discount does not protect capital if the normalized value is being revised down faster than the quotation. The adverse cases were therefore partly denominator errors, not just patience tested by volatility.

3. Concentration before disconfirmation

Aryzta's 9%-plus weights in major portfolios and 16.9% in the concentrated vehicle made incomplete governance analysis a portfolio-level event. Teekay's linked securities made nominal name count overstate economic diversification. Valaris, by contrast, illustrates the benefit of small exposure when the fund exits before the issuer later enters legal restructuring; Cobas's realized loss remains unknown. The comparison supports a robust lesson even without a published hard rule: position size should reflect the probability that the valuation model itself is wrong, not only the upside if it is right.

4. A philosophy that can explain away price evidence

Separating price from value is indispensable to contrarian investing, but it can make adverse prices psychologically easy to dismiss. The 2018 commentary paired new losses with a table of past drawdowns and two-year recoveries. That history provided perspective, yet it could not establish that Aryzta or DIA would repeat an earlier portfolio rebound. Each holding still required a fresh base-rate, balance-sheet, and governance test. (Cobas fourth-quarter 2018 commentary)

Independent factor analysis provides another control. AQR estimated that from November 2016 through June 2020 Cobas Selección had a -13.6% average annual excess return, 22.7% volatility, and -0.60 Sharpe ratio; its selected factor model explained 83% of returns and produced a statistically insignificant -4.9% alpha estimate [single-source study]. AQR cautioned that factor choice is vulnerable to hindsight and implementation frictions. The finding neither proves no skill nor excuses security-specific errors. It suggests that an adverse value-and-small-cap regime and concentrated mistakes operated at the same time. (AQR, “Spain's Value Investors,” 2020)

What changed—and what is not proved

The personal Nissan episode produced the clearest stated sizing rule: diversify enough that one idea cannot dominate survival. Cobas's clearest later change after Aryzta was analytical: it said companies run by professional executives without meaningful ownership would receive greater caution, while its preference for family-controlled businesses was reinforced. The team also reduced Aryzta after the crisis. The direct interview adds a sizing admission. These are real postmortem signals, but reduction can reflect price, flows, or valuation as well as a new rule; the public evidence does not disclose a binding lower position cap. (Wiley, Investing for the Long Term) (MOI Global book compilation) (Cobas third-quarter 2018 commentary) (Cobas second-quarter 2019 commentary) (El Economista interview, 2018)

Babcock's 2021 issuer update disclosed large charges and impairments but does not document Cobas's response. Cobas's 2025 case study says it kept increasing the position through much of 2023, began selling after operating milestones and share-price appreciation, and sold its remaining shares in June 2025 as estimated upside fell. This documents active resizing and eventual exit, not an explicit re-underwriting procedure or an Aryzta-caused process reform. (Babcock business update, 2021) (Cobas first-half 2025 commentary)

No public evidence was found for a formal error taxonomy, mandatory premortem, leverage ceiling at the investee level, automatic reduction after a target-value cut, or disclosed portfolio-wide correlation budget. Cobas said in 2018 that it had counted as many as 45 errors across more than 500 investments. That self-reported count shows an effort to normalize error, but without the names, criteria, denominator construction, or outcomes it is not an auditable hit rate. (Cobas fourth-quarter 2018 commentary)

Legal and governance boundary

The adverse legal record should not be inflated into an investment loss. Norway's financial regulator imposed a NOK 200,000 penalty [single-source regulator figure] on Cobas Asset Management, not Paramés personally, for negligent late notification of a Höegh LNG ownership threshold and omission of holdings through underlying funds. The decision is a compliance failure, not a fraud finding or evidence of client-asset loss. (Finanstilsynet decision, 2021)

A separate Spanish civil dispute concerned Paramés's departure from Bestinver. In June 2025, Spain's Supreme Court reversed the appellate court's finding that Paramés had breached a reasonable pre-notice duty and reinstated dismissal of Acciona's claim. It was a civil company/shareholder contract dispute—not an employment judgment, securities-law sanction, or adjudication of investment performance. (Spanish Supreme Court civil ruling summary, 2025)

Bottom line

Paramés survived losses large enough to test the practical meaning of a long horizon. The best evidence does not support a personal-bust legend. It supports a 61.52% measured Cobas Selección drawdown, a six-year recovery to the old high, an admitted oversized Aryzta error, a premature DIA entry, concentrated cyclical stress in Teekay, small exposure to Valaris before that issuer's bankruptcy, and a volunteered Inditex omission.

The deeper conclusion is not that patience failed. It is that patience is not a risk control by itself. It must be paired with aligned governance, balance-sheet survivability, explicit disconfirming evidence, economic rather than ticker-level concentration, and sizing that assumes the appraisal may be wrong. Paramés and Cobas publicly diagnosed several of those failures. The record is still too incomplete to prove that the resulting safeguards are formal, binding, or sufficient to prevent recurrence.

As of 2026-07-18, Francisco García Paramés is living and remains the chairman, CEO, director, and public investment leader of Cobas Asset Management. This file privileges signed Paramés letters, his own book or publisher samples, official Cobas/Bestinver materials, direct interviews, and conference or podcast pages. Spanish originals are kept where the evidence is Spanish; English excerpts are used only when the source itself is English or an official English translation.

Provenance Rules

• Direct personal voice means a signed letter, book excerpt, direct interview, or event page presenting Paramés as the speaker.
• Cobas/Bestinver voice means official investment commentary from his firm or team. It is useful for the living doctrine, but it is not treated as a private personal diary.
• Reported interview/conference voice means a reputable outlet quoting or paraphrasing remarks at a conference or interview. It is included, but lower-ranked than signed letters and transcripts.
• Quote-list pages, motivational roundups, and attractive English paraphrases without a traceable original are excluded from the quote list.

Quotes By Theme

Formation, Study, And Temperament

1. "el enfoque adecuado del estudio de la economía y de la inversión debe ser el estudio del hombre" - direct book excerpt, 2016. (Deusto/Planeta sample, 2016)
2. "No. We focus on very basic things." - direct Bestinver-team interview, 2008. (Value Investor Insight via Scribd, 2008)
3. "we'd like our time horizon to be forever." - direct Bestinver-team interview, explicitly framed as following Buffett's lead, 2008. (Value Investor Insight via Scribd, 2008)
4. "It was 50 per cent but that was pure luck." - direct interview on Arcelor/Mittal, 2011. (Value Research, 2011)
5. "Nadie sabe con certeza que pasará y nosotros desde luego no tenemos la bola de cristal." - Paramés-authored Cobas blog, 2022. (Cobas, 2022)
6. "para invertir bien a largo plazo no es necesario saberlo." - Cobas/team commentary on macro humility, 2022. (Cobas 4Q commentary, 2022)

Owners, Clients, And Time Horizon

7. "clients and potential clients as co-investors." - direct interview on client alignment, 2011. (Value Research, 2011)
8. "five to seven years." - direct interview on the minimum recommended equity horizon, 2011. (Value Research, 2011)
9. "We don't fight management." - direct interview on shareholder-meeting activism, 2011. (Value Research, 2011)
10. "when a better opportunity presents itself." - signed Cobas launch letter on sell discipline, 2017. (Paramés letter, 2017)
11. "la gestión pasiva tiene su papel en las carteras de los inversores" - reported Value School launch remarks, 2017. (Funds Society, 2017)
12. "los inversores tengan donde elegir" - reported Value School education mission, 2017. (Funds Society, 2017)
13. "Without it, our work would be meaningless." - signed fee-structure letter; the pronoun refers to investor confidence, 2020. (Paramés fee letter, 2020)

Price, Value, And The Research Process

14. "buying companies with high and sustainable returns... tends to work a little bit better." - published English interview excerpt, 2013. (Wiley excerpt, 2013)
15. "Each company... must be a very safe business by its nature." - direct Bestinver interview reported by CFA Institute, 2013. (CFA Institute, 2013)
16. "el circulo de competencia de Bestinver es la valoración de empresas" - Bestinver/team voice in a CNMV-filed report, 2014. (CNMV Bestinver report, 2014)
17. "invertir en compañías que valen más de lo que dice el mercado" - Cobas/team definition of value investing, 2020. (Cobas 3Q commentary, 2020)
18. "al final la generación de caja es el único factor que cuenta" - Cobas/team commentary on long-run stock prices, 2021. (Cobas 2Q commentary, 2021)
19. "no tratamos de predecir lo que va a hacer la bolsa" - Cobas/team commentary, 2023. (Cobas 2H commentary, 2023)
20. "Sólo hace falta paciencia y mentalidad empresarial a largo plazo." - Cobas/team commentary, 2023. (Cobas 2H commentary, 2023)
21. "un trabajo minucioso y casi artesanal" - Cobas/team description of the process, 2024. (Cobas 1H commentary, 2024)
22. "determine the normalised profit that is sustainable over a complete economic cycle." - Cobas/team valuation note, 2025. (Cobas 2H commentary, 2025)

Volatility, Mistakes, And Crisis Behavior

23. "Los errores son parte del proceso." - Cobas/team post-2018 drawdown commentary, 2018. (Cobas 4Q commentary, 2018)
24. "Comprar cuando nadie quiere comprar es una buena idea a largo plazo." - reported conference remarks, 2018. (Funds Society, 2018)
25. "si pones dinero donde todo el mundo lo pone ya sabemos como acaba" - reported conference remarks, 2018. (Funds Society, 2018)
26. "Hemos cometido 43 errores de inversión en 16 años" - reported annual-conference remarks, 2019. (Funds Society, 2019)
27. "el ciclo bajista ya está aquí, ya lo hemos vivido" - reported annual-conference remarks, 2019. (Funds Society, 2019)
28. "el mercado puede ser irracional a corto plazo, pero no lo es a largo plazo" - Cobas/team commentary, 2019. (Cobas 2Q commentary, 2019)
29. "típico comportamiento errático de mercado que debemos aprovechar." - Cobas/team commentary, 2019. (Cobas 3Q commentary, 2019)
30. "This does not affect our valuations." - signed COVID-period letter, 2020. (Paramés COVID letter, 2020)
31. "The risk of failure of any of the 100 contracts is very low." - signed COVID follow-up letter, 2020. (Paramés COVID follow-up, 2020)
32. "short-term suffering is compensated by greater long-term value." - signed COVID follow-up letter, 2020. (Paramés COVID follow-up, 2020)

Education, ESG, And Current Opportunity Set

33. "Invertimos a largo plazo" - reported annual-conference remarks on ESG and long-term ownership, 2021. (Funds Society, 2021)
34. "ESG no es una palabra, es un hecho" - reported annual-conference remarks, 2021. (Funds Society, 2021)
35. "es muy atractiva y está muy penalizada" - reported Expansión interview on Spanish equities, 2024. (Cobas repost of Expansión interview, 2024)
36. "Comprar buenas compañías... y venderlas cuando están de moda" - Cobas/team process summary, 2025. (Cobas 1H commentary, 2025)

Annotated Index Of Primary Materials

Books And Book-Adjacent Materials

1. Invirtiendo a largo plazo / Investing for the Long Term, 2016-2018. The Spanish book and later English edition are the core source for Paramés's self-narrative, economic worldview, Graham-Buffett-Fisher synthesis, and reading-heavy temperament; public samples are partial, so the file quotes only from accessible excerpts. (Deusto sample, 2016; Wiley edition, 2018)
2. Fundación Rafael del Pino book dialogue, 2016. Official event page for Paramés's dialogue with Carlos Rodríguez Braun at the book launch; useful for his public framing of value investing as common sense, psychology, Austrian economics, and long-term ownership. (Fundación Rafael del Pino, 2016)
3. Europa Press / Deusto quote digest, 2016. Secondary book-excerpt article listing compact maxims from the book. It is a lead source, not used as a primary substitute when the public book sample is available. (Europa Press, 2016)

Letters And Official Investor Commentaries

4. Bestinver Internacional semiannual report, 1H 2014. CNMV-filed Bestinver-era primary document showing team language on circle of competence, valuation work, rotation, and portfolio economics shortly before Paramés's departure. (CNMV, 2014)
5. Letter to investors from Francisco García Paramés, 7 March 2017. Signed founding Cobas letter on unchanged philosophy, high equity exposure, sell discipline, team, capacity limits, communication, and family co-investment. (Paramés letter, 2017)
6. Cobas fourth-quarter commentary, 2018. Primary firm commentary after the Aryzta/Teekay drawdown; strongest official text on error-counting, recovery analogies, patience, and process under stress. (Cobas, 2018)
7. Cobas second-quarter commentary, 2019. Primary defensive letter during early Cobas underperformance, with explicit market-efficiency, time, value-accumulation, and quality/valuation language. (Cobas, 2019)
8. Cobas third-quarter commentary, 2019. Primary firm commentary on market erratic behavior, shipping, automakers, and contrarian re-underwriting during a weak style period. (Cobas, 2019)
9. COVID letter, 16 March 2020. Signed stress-period letter separating market price from business value, emphasizing recession effects, panic/passive selling, and ongoing communication. (Paramés letter, 2020)
10. COVID follow-up letter, 23 March 2020. Signed portfolio review by sector, especially LNG contracts, financials, sales loss, and the short-term-suffering-versus-long-term-value framework. (Paramés letter, 2020)
11. New fee-structure letter, December 2020. Signed client-alignment letter on lower fees for long-tenured investors, seniority preservation, and trust. (Paramés letter, 2020)
12. Cobas third-quarter commentary, 2020. Primary philosophy restatement defining value investing as buying companies worth more than the market price, not merely low-multiple or declining businesses. (Cobas, 2020)
13. Cobas second-quarter commentary, 2021. Primary commentary on cash generation as the long-run driver of share prices and on style recovery after growth dominance. (Cobas, 2021)
14. Paramés-authored Cobas blog, April 2022. Direct blog post on value investing's moment, macro uncertainty, value-versus-growth history, and the limits of forecasting. (Cobas, 2022)
15. Cobas fourth-quarter commentary, 2022. Primary text on economic-prediction humility, preparation versus forecasting, owner mindset, and balance-sheet quality. (Cobas, 2022)
16. Cobas second-half commentary, 2023. Primary semiannual commentary with a section explicitly by Paramés on Asia, plus broader text on non-prediction, quality businesses, patience, and enterprise mentality. (Cobas, 2023)
17. Cobas first-half commentary, 2024. Primary detailed process source: idea generation, preliminary study, complete model, team debate, investment committee vote, sizing, and portfolio rotation. (Cobas, 2024)
18. Cobas first-half commentary, 2025. Primary case-study source using Babcock to explain buying unloved companies, increasing conviction through adverse price action, and selling when valuation normalizes. (Cobas, 2025)
19. Cobas second-half commentary, 2025. Latest annual-process source found in English; useful for valuation mechanics, normalized profit, multiples/DCF, discount-rate ranges, target values, and risk disclaimers. (Cobas, 2025)

Interviews, Conferences, Speeches, And Podcasts

20. Value Investor Insight interview, 2008. Direct Bestinver-team interview covering long horizon, basic research questions, geographic circle of competence, holding companies, and post-crisis opportunities; available via a third-party scan. (Value Investor Insight via Scribd, 2008)
21. Value Research interview, 2011. Direct interview on client expectations, co-investor framing, five-to-seven-year money, education, Austrian economics, Arcelor luck, and non-activism. (Value Research, 2011)
22. CFA Institute profile/interview, 2013. Contemporaneous reporting with direct access to the Bestinver team; useful for concentration, safety requirements, family businesses, European inefficiency, and portfolio evolution. (CFA Institute, 2013)
23. Value School launch, 2017. Reported remarks on investor education, personal biases, passive management's role, and the goal of broadening the Spanish investing ecosystem. (Funds Society, 2017)
24. Second Cobas annual investor conference, 2018. Reported remarks on fixed-income optionality, expensive quality stocks, shipping, UK contrarian exposure, research costs, and sticking with good ideas. (Funds Society, 2018)
25. Third Cobas annual investor conference, 2019. Reported remarks on early Cobas losses, 43 investment errors, Aryzta/Teekay attribution, revaluation potential, and the claim that the bear cycle had already arrived. (Funds Society, 2019)
26. Value Investing with Legends podcast, 2020. Columbia/Heilbrunn podcast page for an English-language interview on his self-taught route, investment experience, and long-term value discipline. Exact audio transcript was not available in the opened page, so it is indexed but not quoted. (Apple Podcasts, 2020)
27. Fifth Cobas annual conference, 2021. Official Cobas page and press coverage for fee changes, ESG/impact language, BrainVestor, pension plans, and post-2020 recovery. (Cobas, 2021; Funds Society, 2021)
28. Expansión interview reposted by Cobas, 2024. Direct recent interview source on Spanish equities, passive/index flows, opportunities, and the persistence of his style after the early Cobas drawdown. (Cobas, 2024)
29. Rankia Markets Experience session, 2025. Official Cobas video page for "40 años invirtiendo con convicción y sin modas," covering career, errors, quality/catalysts, passive-management opportunity, real assets, and teaching children to invest; no transcript was available in the opened page. (Cobas, 2025)
30. BrainVestor workshop interview, 2026. Official Cobas page for an interview on investor psychology, biases, doubts, conviction, and decision-making after 30-plus years; indexed as current testimony but not quoted without a transcript. (Cobas, 2026)
31. Tenth Cobas annual conference video excerpts, 2026. Official pages on investment philosophy, revaluation potential, and annual-conference conclusions; useful current materials but quoted only where transcript text is visible in page or PDF sources. (Cobas philosophy video, 2026; Cobas revaluation video, 2026; Cobas conclusions video, 2026)

Quote Boundaries And Watchlist

• Personal versus institutional voice. The signed 2017 and 2020 letters are direct Paramés voice. Cobas and Bestinver commentaries are institutional/team voice unless explicitly marked as a Paramés section. The quote list labels that distinction.
• Language and translation. Spanish wording is retained where the source is Spanish. English quotes come from English interviews, English publisher excerpts, or official English translations.
• Video and podcast limits. Several recent Cobas, Ivey/Ben Graham Centre, Value School, and Columbia podcast/video pages are strong source leads, but exact quotes were not used unless the opened page supplied text that could be checked.
• Quote aggregators. Lines such as "patience is an investor's biggest asset," "risk comes from not knowing what you're doing," or polished English sayings about market oscillation were excluded because they are either Buffett-origin lines, paraphrases, or untraced quote-list material.
• Adverse context. Own-words sources around Aryzta, Teekay, Babcock, early Cobas underperformance, the Bestinver/Acciona litigation, and the 2021 Norwegian Cobas entity-level disclosure penalty should not be laundered into clean aphorisms. They are context for the later mistakes and mental-models files, not evidence that every stated principle worked as intended.

Bibliographic verdict

Francisco García Paramés has one verified sole-authored book-length work, not a shelf of separate investment books. Invirtiendo a largo plazo: Mi experiencia como inversor appeared in Spanish in 2016; Wiley published the English translation, Investing for the Long Term: My Experience as an Investor, in 2018. The English edition is a translation and later edition of the same work. Publisher records show nine substantive chapters divided between career history and investment theory, followed by conclusions, a 26-idea appendix, reading material, and references. (Planeta/Deusto, 2016) (authorized Planeta excerpt, 2016) (Wiley, 2018)

Beyond the book, the defensible written corpus includes three substantive individually bylined essays, four individually signed Cobas letters, a signed educational appeal, and several forewords. The official Cobas author archive is a useful ceiling check, but it shows only the book and two Cobas-hosted bylines; the third essay and forewords require their own publisher records. Interviews and recorded appearances are important direct testimony but are not authored essays, while Bestinver team interviews and most Cobas commentaries are collective. A named speaker, a signature, a byline, and sole authorship are different kinds of provenance. (Cobas author archive)

Works by Paramés

1. Investing for the Long Term / Invirtiendo a largo plazo (2016; English 2018)

Central thesis. Durable investment judgment is learned rather than inherited. It combines historical experience, Austrian-economics ideas about uncertainty and capital, business-owner analysis, valuation discipline, and emotional independence. The book's first half explains how Paramés says he acquired that method; the second turns it into an argument for long-horizon ownership of undervalued productive businesses.

Ten key ideas, paraphrased:

1. Build a personal framework through study. Reading great investors, economic history, and business cases is not ornament; it is how an investor forms independent judgment and avoids outsourcing conviction.
2. Treat a share as ownership. The analytical object is a productive enterprise and its long-run cash-generating capacity, not a ticker, chart, or near-term market forecast.
3. Accept radical uncertainty. Economic actors cannot forecast a complex future precisely. Scenario thinking, balance-sheet resilience, and a valuation discount are more defensible than point forecasts.
4. Prefer productive real assets over monetary promises over long horizons. Equities can protect purchasing power because sound businesses can adapt prices and reinvest, although the claim is not a guarantee for every starting valuation or period.
5. Use passive investing when no active edge exists. Active management is justified only when skill, temperament, time, and costs plausibly allow the investor to exploit mispricing.
6. Look for quality before applying a cheap multiple. Competitive position, reinvestment opportunity, capital employed, growth economics, financial strength, and normalized earning power determine whether apparent cheapness is real.
7. Search where aversion creates neglect. Cyclical, temporarily troubled, unfashionable, or underfollowed companies can offer opportunity, but only if the business and financing can survive the path.
8. Demand aligned stewardship. Owner-managers and family shareholders may think longer term, yet alignment must be tested rather than inferred from ownership alone.
9. Separate price from value and sell on opportunity cost. A falling quote is not automatically new information about intrinsic value; a rising quote is not success if better alternatives exist. Capital should move when the relative discount changes.
10. Design for one's own psychology. Patience, independence, diversification, and awareness of cognitive bias are operating safeguards. Conviction without sizing and falsification controls can become fragility.

The publisher-authorized excerpt supports the structure and intellectual intent, but it is only 28 pages; it does not provide public access to all case details or appendices. The book is also autobiography. Its Bestinver recollections are selected by the protagonist and often describe a team record, not an audited personal account.

Best chapters and reading path.

• Chapter 2, “Going Solo (1991–2002),” and Chapter 3, “Investing as a Team (2003–2014)” are the best historical pair. Read them to see the process evolve from self-education and early mistakes into a team practice, while keeping hindsight and collaborator attribution in view.
• Chapter 4, “Austrian School of Economics,” supplies the intellectual premises: subjective value, entrepreneurship, capital, cycles, and the limits of forecasting. It explains the vocabulary used later, though readers need not accept every Austrian claim to use the investment method.
• Chapters 5 and 6, “Investment” and “Passive and/or Active Management,” place stock selection inside the prior questions of saving, real versus monetary assets, asset allocation, costs, and whether active management is appropriate.
• Chapters 7 and 8, “Investing in Stocks I” and “Investing in Stocks II,” are the practical core. They move from Graham-derived foundations to opportunities, valuation, business quality, management, normalized profits, and reinvestment.
• Chapter 9, “The Irrational Investor Lurking Within Us All,” is the necessary control on the preceding chapters. It is most useful when read alongside the later Aryzta and early-Cobas record, because behavior is easiest to discuss before a cherished thesis is under pressure.
• The concluding 26 ideas and reading appendix are the best rereading tool, not a substitute for the argument or cases that precede them.

Two reviews improve the reading. Enrique García Sáez praises the European market history but argues that the theoretical half is loosely organized, light on detailed cases and mistakes, and unevenly edited. Jesús Huerta de Soto, despite intellectual proximity to Paramés, questions the universality of the equity-superiority claim, the selected historical intervals, fee incentives, and star-manager instability. These critiques make the book more useful: it is a coherent statement of one practitioner's development, not a neutral textbook or complete validation of his results. (inBestia review, 2016) (Procesos de Mercado review, 2019)

2. Letter to investors (7 March 2017)

This signed letter is the founding written manifesto of Cobas. It is not another book chapter: it states what Paramés intended to carry unchanged into the new firm and how the organization and products were supposed to embody it. (Cobas signed letter, 2017)

Central thesis. A valuation-led equity discipline can survive a change of institution if the team, client horizon, capacity, co-investment, communication, and fees are built to support it.

Five key ideas, paraphrased:

1. The method remains bottom-up purchase of securities below assessed value rather than prediction of market levels.
2. Long-run purchasing-power preservation requires accepting equity volatility; short-horizon capital is therefore a poor fit.
3. High equity exposure is the normal posture, but the manager retains room to act when securities are unattractive.
4. Selling is an opportunity-cost decision: replace a holding when its discount narrows or a superior risk-adjusted opportunity appears.
5. Alignment is institutional as well as rhetorical—reference shareholders, family capital, capacity limits, team structure, and plain client communication affect whether the method can endure.

Best sections. Read the opening philosophy statement, the portfolio-and-selling discussion, and the closing sections on people, ownership, capacity, and client relationship. Product descriptions and launch mechanics are dated, but they are useful ex-ante evidence against which later Cobas practice can be judged.

3. COVID-19 letter (16 March 2020)

This signed crisis letter is the strongest contemporaneous test of the doctrine during the early pandemic selloff. Portfolio diagnosis likely incorporated the investment team's work, so the signature establishes personal communication, not sole origination of every company estimate. (Cobas signed crisis letter, 2020)

Central thesis. A violent price decline should be evaluated through the permanent effect on each business's normalized cash generation and financing, not through the decline itself.

Five key ideas, paraphrased:

1. Market price and business value can diverge sharply when holders need liquidity or shorten their horizon.
2. The right question is how much earning power is permanently impaired, not how severe one recession quarter becomes.
3. Sector labels are inadequate; contract structures, balance sheets, liquidity, and demand persistence differ by company.
4. Redemptions and market liquidity are portfolio risks distinct from the operating risks of the underlying businesses.
5. Long-term conviction must rest on updated company work, not on a generic claim that every fall is temporary.

Best sections. The company-group discussion is more valuable than the market rhetoric because it exposes the assumptions being made in real time. The liquidity and investor-behavior discussion is the second essential section. Read both beside actual later outcomes; an ex-ante letter records reasoning, not proof that the reasoning was correct.

4. COVID-19 follow-up letter (23 March 2020)

The second signed letter, one week later, is unusually useful because it shows a process updating while uncertainty was still expanding. It should be read as a companion work rather than a repetition. (Cobas signed follow-up, 2020)

Central thesis. Rapid contact with portfolio companies can refine near-term damage estimates without replacing the longer-horizon valuation framework.

Five key ideas, paraphrased:

1. New information should change company-level assumptions even when the overarching philosophy is stable.
2. Contract duration, refinancing needs, buybacks, regulation, and customer solvency can matter more than the headline sector.
3. Temporary earnings loss and permanent capital impairment require different valuation responses.
4. A portfolio should be understood as distinct blocks of exposure rather than one undifferentiated “value” bet.
5. Confidence stated during crisis needs later falsification; speed of updating is not the same as accuracy.

Best sections. Focus on the dated company-contact updates and changed assumptions. The closing defense of long-term value is context, but the analytical delta from the first letter is the distinctive contribution.

5. Three individually bylined essays

“Cobas AM lanza una colección de libros sobre el ahorro y la inversión” (2017)

This is more than a collection announcement: it is a substantive Spanish review of Peter Lynch's method. Its central thesis is that a repeatable business-analysis process and financial education can replace emotional timing and macro pessimism. Five useful ideas are to study companies systematically, distrust recurring end-of-the-world narratives, distinguish true active management from closet indexing, combine field observation with accounts, and test whether a cyclical can finance the trough. The best sections discuss pessimism, manager selection, observation, and cyclical survivability. The text substantially overlaps Paramés's foreword to the Deusto edition of Batiendo a Wall Street, so the two contexts count as one intellectual work, not independent essays. (Cobas bylined essay, 2017) (Planeta/Deusto Lynch edition, 2017)

“La Escuela Austriaca de Economía: Un marco conceptual sólido de base” (2018)

This Value School byline condenses the book's Chapter 4 for a public audience. Its central thesis is that Austrian economics offers an investor a coherent compass for human action and changing capital structures without enabling precise macro forecasts. Five ideas organize it: investors need a conceptual base; markets reflect purposeful but fallible actors; mathematical precision cannot capture all economic choice; dispersed knowledge makes central prediction fragile; and entrepreneurs operate in continuing disequilibrium. The opening case for investor relevance and final “compass” synthesis are the best sections. It is Paramés's chosen lens, not economic consensus, and it is an adaptation rather than a second original theory. (Value School bylined essay, 2018)

“El momento de la inversión en valor” (2022)

This is the most substantial later bylined investment essay. Its central thesis is that historically wide value-versus-growth valuation dispersion and monetary normalization created a favorable starting setup for value, without making reversal timing knowable. Five ideas are that value's relative slump was unusually deep; starting valuation matters; long-duration growth cash flows are more discount-rate-sensitive; nearer-term cash generation can be less sensitive; and scenario outcomes must not be confused with forecasts. The dispersion evidence, scenario construction, and explicit uncertainty paragraph are the best sections. It remains dated manager advocacy and must be judged against subsequent fund results. (Cobas bylined essay, 2022)

6. Shorter signed and foreword contributions

The December 2020 fee letter is a signed operational work rather than an investment essay. Its central idea is that tenure-based fee reductions can align a manager with patient clients; the useful sections explain seniority preservation and the fee table. Product terms can change, so the letter belongs in the bibliography but not in the philosophical core. (Cobas signed fee letter, 2020)

Value School's institutional “Quiénes somos” page ends with an appeal signed by Paramés. Its central thesis is that broad financial education and patient investing can strengthen individual freedom; it is a signed institutional statement, not investment scholarship. (Value School signed appeal)

Three foreword chains are defensible. The Batiendo a Wall Street foreword is the other publication context for the 2017 Lynch essay and should not be double-counted. A Marcial Pons catalog explicitly credits Paramés with the foreword to the third edition of Luis Allué Bellosta's Alicia en Wall Street, but no publisher-licensed full foreword was retrieved, so its ideas should not be reconstructed. The official Planeta record and excerpt verify another Paramés foreword in Jaime Alonso Stuyck's Inversión de patrimonios, but the public excerpt omits the foreword itself, so its ideas and best sections are likewise not reconstructed. (Marcial Pons catalog, 2013) (Planeta/Deusto, Inversión de patrimonios, 2022) (authorized publisher excerpt, 2022)

An endorsement on a cover is not authorship, and the collection announcement is not evidence that Paramés wrote every later foreword.

Other direct primary material

Paramés's direct interviews form a valuable chronological supplement rather than additional authored works. The 2003 El País interview is the clearest early account of appraisal value, cash, bottom-up research, and rejection of technical analysis. The 2011 Value Research interview adds co-investment, five-to-seven-year client capital, communication, and the shift toward business quality. The 2016 Cinco Días retrospective bridges Bestinver and Cobas and discusses errors as well as continuity. (El País interview, 2003) (Value Research interview, 2011) (Cinco Días interview, 2016)

The 2008 Value Investor Insight interview is important crisis evidence but names Paramés, Álvaro Guzmán, and Fernando Bernad; it is team-direct, not a solo work. The 2016 Fundación Rafael del Pino dialogue is the best official companion event to the book. The December 2024 Expansión interview is the best recent process update, explaining the seven-member, one-person/one-vote decision structure and delegation. The 2025 Rankia and 2026 BrainVestor pages extend the chronology into late-career reflection and investor psychology, but their edited descriptions are not substitutes for transcripts. (Value Investor Insight archive, 2008) (Fundación Rafael del Pino, 2016) (Expansión interview PDF, 2024) (Rankia event page, 2025) (BrainVestor event page, 2026)

Best works about Paramés, ranked

No independent book-length biography was located. The best external literature is analytical and fragmented; each item answers a different question.

1. AQR, “More Superstar Investors: Francisco Garcia Paramés” (2019). Best independent quantitative reconstruction of the Bestinfond era. It tests how much of the apparent record is associated with value, low-risk, and size factors and leaves only a statistically insignificant residual alpha in its selected model. Read it for disciplined skepticism, while remembering its hindsight in factor selection, gross implementation assumptions, USD excess-cash convention, and focus on one fund rather than personal wealth. (AQR, 2019)
2. AQR, “More Superstar Investors: Spain's Value Investors” (2020). Best adverse analysis of early Cobas and strongest control against treating “Paramés” as a continuous personal return series. It ends in June 2020, before the later recovery, and explicitly confronts short samples and team attribution. (AQR, 2020)
3. CFA Institute, “Bestinver: How to Deal with the Departure of Your Star Manager” (2015). Best case study of platform versus star attribution, redemptions, client communication, inherited holdings, and succession after Paramés left. It is practitioner analysis, not peer-reviewed scholarship, and its contemporary suggestion about a notice obligation was overtaken by later litigation. (CFA Institute, 2015)
4. Jesús Huerta de Soto's Procesos de Mercado review (2019). Best published intellectual critique of the book. Its author is sympathetic to the Austrian tradition, yet presses on historical-period selection, capital income, fees, equity exceptionalism, and star-manager team instability. It is a critique of the argument, not independent performance verification.
5. Enrique García Sáez's inBestia review (2016). Best practical review of the Spanish edition's usability. It identifies the book's historical value and its organizational, case-detail, error-analysis, and editing weaknesses. It appeared before Cobas provided an out-of-sample test.
6. Morningstar Spain, “¿Es García Paramés un buen gestor?” (2019). Best concise example of evaluation discipline during underperformance: the article records a poor early rating but argues that three years is insufficient for a skill verdict. The rating is backward-looking and the article predates both the pandemic trough and recovery. (Morningstar Spain, 2019)
7. Strictly Value review (2018). Best early English-language warning that some “cheap” holdings combined leverage with deteriorating economics, making it a useful bridge from the book's doctrine to Aryzta-era risk. It is an anonymous practitioner blog, not formal research. (Strictly Value, 2018)
8. Bo Börtemark, InvestingByTheBooks (2019). Best compact English reading guide to the chapters, normalized earnings, cyclicals, rebalancing, behavior, and bibliography. It is favorable and mostly summarizes Paramés; use it for navigation, not corroboration. (InvestingByTheBooks, 2019)

What to read first

For the shortest serious sequence, read book Chapters 2–4 for formation, Chapters 7–9 for practice and psychology, then the 2017 letter for the institutional design promised at Cobas. Read both March 2020 letters next as an ex-ante stress test. Finish with the two AQR papers and CFA succession case so that factor exposure, team attribution, regime dependence, and platform risk qualify the first-person story.

The source hierarchy should remain explicit. Publisher pages establish editions and contents; the authorized excerpt supports only the pages it exposes; signed letters establish Paramés's communicated view; interviews establish direct testimony; team documents establish team positions; reviews interpret; and factor studies test fund returns under specific models. Translations, retailer blurbs, quotation compilations, and Cobas-edited video descriptions do not become independent corroboration merely because they repeat the same claim.

Bottom line

Paramés's canon is small enough to read completely. The book is the durable synthesis, the bylined essays are compact extensions or adaptations, the 2017 letter is the institutional manifesto, and the March 2020 pair is the live crisis laboratory. Together they show a consistent owner-oriented, valuation-driven philosophy—but consistency is not validation. The strongest reading practice is therefore adversarial: pair the autobiography with team-attribution controls, the launch promise with later governance evidence, and the crisis letters with realized outcomes. A reader who does that gets something more useful than a list of maxims: a method, its intended institutional supports, and the evidence needed to test where conviction becomes error.

Evidence and attribution boundary

Paramés does not publish a formally numbered “mental models” system. The defensible reconstruction combines four evidence classes: ideas he states in his book, signed letters, and bylined essays; concepts adapted from Graham, Buffett, Fisher, Lynch, and Austrian economics; Bestinver or Cobas team practices; and safeguards reconstructed by the Investing Canon from successes and failures. Those classes are not interchangeable. The Spanish and English editions of Investing for the Long Term are the sole-authored synthesis, while later portfolio mechanics usually belong to the institution and its changing team. (Wiley, 2018) (authorized Planeta excerpt, 2016)

The distinction is especially important after 2020. Paramés said in December 2024 that Cobas had moved to a seven-person collective process in which his vote counted as one and the team sometimes acted without him. An individual idea can take weeks, months, or a year of modeling and expert/company contact before a vote. This describes current Cobas governance, not a timeless solo-Paramés rule. (Expansión interview, 2024) (Cobas team-process communication, 2024)

Named, adapted, and reconstructed models

The operational decision checklist

This checklist reconstructs the public method. Items labeled “required” are Canon safeguards where public evidence does not prove a formal Cobas rule.

1. Capital-fit gate. Keep emergency and near-term spending outside the strategy. Bestinver-era guidance asked clients to invest only capital unnecessary for at least five to seven years; Cobas's current International product warns that it may be unsuitable below five years. This is a horizon screen, not a promise that five years always suffices. (Value Research interview, 2011) (Cobas International product page)
2. Generate ideas without outsourcing judgment. The current team says it reads reports, results, and news, follows competitors and prior holdings, and does not rely on consensus-data screens except to notice recent losers. Treat that as a 2023 Cobas practice, not proof that Paramés never used screens historically. (Cobas valuation-method article, 2023)
3. Explain the business. Map products, customers, suppliers, industry structure, value-chain position, competitive advantage, disruption risk, and the few variables that drive cash. If the thesis needs opaque mathematics to hide weak economics, reject it.
4. Reconstruct a cycle. Current Cobas material says models normally include at least ten years of financial history [single-source team disclosure]. Normalize revenue, margins, working capital, maintenance capital expenditure, taxes, cash conversion, and share count. Write why the next cycle should resemble—or differ from—the last one. (Cobas valuation-method article, 2023)
5. Underwrite stewardship. Examine insider economics, capital allocation, pay, acquisitions, leverage, dilution, minority treatment, and behavior under stress. Interview management but corroborate it through customers, competitors, former employees, specialists, regulators, and filings. The expanded corroboration list is a Canon implementation of the public qualitative process, not a disclosed mandatory protocol. (Cobas second-half commentary, 2024)
6. Make survival the gate. Before upside, test cash, debt maturity, covenants, refinancing, pension/lease liabilities, working-capital needs, counterparty exposure, and dilution capacity. This solvency-first ordering is a Canon repair made necessary by Aryzta, Teekay, and Valaris; it is not a publicly disclosed binding Paramés checklist.
7. Value by a range, with version control. A historical Cobas article described normalized profit times a 15× P/E and rejected WACC/beta. Current team documents permit business-specific multiples or DCF and say the explicit or implicit discount rate is usually 6%–12% [single-source current range]. Preserve the date and method used; do not blend them into a timeless formula. (historical Cobas work-method article) (Cobas second-half commentary, 2024)
8. Write “why cheap” and the kill criteria. Specify which holders are constrained or mistaken, what operating evidence should close the discount, and what would show that distress is structural. Required Canon repair: precommit an exit or review trigger for leverage, governance, competitive erosion, and target-value cuts; no formal public Cobas kill-criteria policy was found.
9. Enter through evidence, not urgency. Current Cobas says a majority vote puts a company into the portfolio at a small weight, then the team can increase it as conviction, price, and the rest of the portfolio evolve. “Small” has no public universal percentage. (Cobas first-half commentary, 2024)
10. Size the shared scenario. Combine upside, quality, confidence, liquidity, financing risk, and opportunity cost, then aggregate economic exposures across names. Three shipping or levered cyclical holdings are not three independent risks. Required Canon repair: cap portfolio loss under one shared adverse scenario; no public Cobas correlation budget was located.
11. Monitor the delta. Update the model after results and new qualitative evidence. Separate price change from target-value change. A lower price with stable verified value can justify adding; a lower value can require reducing even if the share price is already down.
12. Sell for one of four reasons. Sell or reduce when price closes the discount, the original analysis was wrong, fundamentals reduce target value, or a better alternative dominates. A 2023 Cobas article says more than 90% of sales occur for price/target convergence [single-source unaudited team claim]; it is not a transaction-ledger statistic. (Cobas valuation-method article, 2023)
13. Run a postmortem. Required Canon repair: record forecast versus outcome, contribution versus realized P&L, and which evidence was knowable at the time. Keep omissions and exited failures in the sample. A rising fund after a mistake does not validate the failed thesis; no binding public Cobas postmortem template was found.

Disclosed rules—and the rules not found

Case tests: where the models worked and failed

CAF—patient ownership and reallocation, narrowly evidenced. A surviving conference recap calls CAF probably the team's best investment, says it remained exceptional but was less undervalued, and records that the position was reduced. It supports patient ownership followed by valuation-led reallocation. It does not establish a complete trade chronology, personal attribution, owner-economics causation, or the source of the return. (Bestinver conference recap, 2011)

Inditex—valuation and research scarcity caused omission. Paramés later identified not buying Inditex as an error, explaining that the team sought lower multiples and lacked time to analyze every company deeply. The case tests opportunity-cost allocation and the risk that a cheapness filter misses exceptional reinvestment; it does not prove that Inditex lay outside the team's competence. The Canon repair is a watchlist with an explicit “what price or fact would change my mind?” field. (Cinco Días interview, 2016)

Aryzta—alignment, normalization, and sizing failed together. Cobas acknowledged that Aryzta lacked its preferred reference-owner alignment and that it had relied too heavily on a professional manager with little ownership. Debt, a dilutive financing fight, and portfolio weight made the path part of the thesis. Paramés later said the weight had been too large [single-source; access-controlled]. The lesson is not “never average down”; it is that governance evidence, solvency, and scenario loss must overrule the attraction of a widening modeled discount. (Cobas third-quarter commentary, 2018) (Cobas fourth-quarter commentary, 2018) (El Economista interview, 2018) (Cinco Días adverse report, 2018)

McClatchy, DIA, and Valaris—cycle versus structure. Paramés admitted that McClatchy's print-to-internet change arrived faster than expected. DIA's operating evidence and governance worsened while price weakness encouraged more confidence. The fourth-quarter 2018 commentary records DIA's target-value cut and partial sale. Cobas's first-half 2020 regulatory report records the Valaris exit before the issuer's August bankruptcy filing, so the failure validates the balance-sheet danger but not a claim that the fund held through bankruptcy. Each case asks the same question: what observable fact proves that “normalized” profit is obsolete? (El Confidencial interview, 2009) (CNMV Cobas report, first half 2020) (Valaris restructuring release, 2020)

Maire Tecnimont—the reallocation loop worked. Cobas's 2024 case shows a small initial position after a vote, additions as price fell while estimated value stayed stable, and progressive sales when price rose faster than target value. It is the clearest public operational example, but the figures and counterfactual target values are manager-reported, not audited transaction P&L. (Cobas first-half commentary, 2024)

Babcock—the team updated and exited, but no formal repair is proved. The 2025 case study says Cobas continued buying through much of 2023, sold after operating milestones and appreciation, and exited the remainder in June 2025 as upside fell. It supports active resizing and opportunity-cost selling; it does not document a mandatory re-underwriting protocol created by Aryzta. (Cobas first-half commentary, 2025)

The 2018–20 portfolio—the factor and rationalization test. AQR found significant early-Cobas value and size exposure, negative realized premia, and a statistically insignificant negative alpha estimate in its selected model. Its factor choice, hindsight, gross implementation, and short sample limit the conclusion. Still, bottom-up names shared common regime exposure, and the signed March 2020 letter is ex-ante price/value testimony to test against later vehicle-specific outcomes—not proof of eventual correctness. (AQR, 2020) (Paramés COVID letter, 2020)

Failure modes of the model

1. Cyclical normalization of a secular decline. Historical margins become false anchors when technology, regulation, customer behavior, or industry structure changes.
2. Leverage outruns patience. Debt maturity and dilution can transfer the upside before normalized earnings return.
3. Alignment halo. Family control or insider ownership can coexist with entrenchment, poor incentives, and weak minority treatment.
4. Confirmation by falling price. A lower quote mechanically increases modeled upside and can reward the analyst psychologically for not changing the denominator.
5. Bottom-up names, top-down factor. Many individually researched holdings can share value, size, cyclicality, liquidity, commodity, or European-demand exposure. AQR's Bestinver analysis similarly associates much of the earlier record with selected systematic styles, without proving zero stock-selection skill. (AQR, 2019)
6. Simple model, omitted tail. Parsimony exposes assumptions, but a fixed multiple or uncomplicated cash-flow model may miss covenants, feedback loops, options, pensions, or nonlinear demand.
7. Patience becomes identity. Client education and a long record can make changing one's mind feel like betrayal of the philosophy.
8. Founder consensus without visible dissent mechanics. One-person/one-vote reduces key-person dependence, but unanimity of worldview can still suppress an independent bear case. The 2026 BrainVestor page identifies confirmation bias and overconfidence as discussion topics, but its edited description is not a transcript or proof of a formal control. (BrainVestor event page, 2026)
9. Capacity erodes the habitat. More capital makes neglected small/mid-cap positions harder to enter, size, and exit; a proposed soft close recognizes the problem but does not solve it until effective.

Transferability

What an individual can copy

• Match equity exposure to actual cash needs and write a minimum holding-horizon assumption.
• Keep a one-page model with ten-year history where available, normalized earnings, a valuation range, financing runway, and share-count reconciliation.
• State the reason for cheapness and prewrite the operating, governance, and balance-sheet facts that would disprove it.
• Read the entire value chain, then use public calls, filings, competitors, customers, trade journals, and regulators to challenge management's version.
• Start small, stage only after new evidence, and cap the combined loss from correlated scenarios.
• Compare price change with target-value change; make every add, trim, and exit legible in a decision journal.
• Retain failed theses and omissions, and compare results with investable style-aware alternatives.

What an individual cannot copy reliably

• Seven analysts' accumulated sector memory, language coverage, continuous monitoring, specialist databases, conferences, and access to management, customers, competitors, and former employees.
• Institutional block liquidity, issuer influence, dealing, legal, compliance, distribution, and cross-vehicle portfolio infrastructure.
• Cobas's collective voting process without recreating genuinely independent review; a social investing group is not automatically a dissent mechanism.
• Headline Bestinver or Cobas outcomes without their original vehicle, team, benchmark, currency, fees, flows, and endpoint.
• Paramés's temperament, tacit pattern recognition, family co-investment, or ability to withstand public multi-year tracking error.

Individuals do have one compensating advantage: smaller capital can enter neglected securities without moving the price and can simply pass on ideas whose complexity, liquidity, governance, or disclosure exceeds their resources. The transferable essence is therefore not “buy what Paramés owns.” It is a falsifiable sequence: understand, normalize, test survival, value, size the shared scenario, update, and sell on evidence or opportunity cost.

Bottom line

The Paramés model is strongest as a disciplined way to separate a business from its quotation and weakest when that separation becomes unfalsifiable. Its durable components are owner thinking, normalized cash economics, simplicity, margin of safety, patient capital, and opportunity-cost reallocation. Its necessary repairs are equally important: solvency before upside, explicit kill criteria, portfolio-level scenario aggregation, independent dissent, and version-controlled valuation. Those repairs do not reject the philosophy; they make patience conditional on facts.

Evidence cutoff: 18 July 2026. This synthesis distinguishes Paramés's decisions from those of the Bestinver and Cobas teams, fund results from personal wealth, and manager calculations from independently reconstructed returns.

Executive Brief

Francisco García Paramés is best understood as a concentrated, long-horizon European value investor whose edge comes from combining business-owner economics with a willingness to look where benchmarked capital is uncomfortable. The label “Spanish Buffett” is less useful than the operating description: start with the company, normalize its earnings across a cycle, value the business with several methods, demand a discount, and wait. His preferred hunting ground has repeatedly included family-controlled companies, neglected industrials, capital-intensive cyclicals, restructurings, and smaller firms whose accounting or industry structure requires more work than a screen can provide. Austrian business-cycle ideas shape his preference for productive real assets and suspicion of monetary forecasting, but the portfolio process is presented as bottom-up rather than a macro-timing system (Cobas, 2026; Cobas, 2017).

The record has two distinct chapters. Bestinver was an exceptional, long-lived public-fund outcome, although published summaries use different dates, currencies, benchmarks, and attribution conventions. CFA Institute reported a roughly 16% annualized Bestinfond result through April 2013 [single-source contemporaneous report], while AQR's April 1994–August 2014 reconstruction found strong benchmark-relative performance but also attributed much of the statistical alpha to systematic style exposures (CFA Institute, 2013; AQR, 2019). Cobas was a harsher live test: its flagship suffered a 61.52% peak-to-trough NAV drawdown [single-source raw reconstruction] from the official Class C series, then recovered strongly. That path does not erase the recovery or prove the method failed; it proves that “volatility is not risk” is incomplete for clients whose behavior, liquidity needs, or mandate cannot survive the interval (Cobas NAV export, 2026; Cobas, 2020).

The trade archive supports the process but not a precise personal P&L ledger. CAF, Elecnor, and Babcock illustrate different combinations of normalized earnings, asset value, owner alignment, and recovery (Bestinver conference recap, 2011; Cobas first-half commentary, 2024; Cobas Babcock review, 2025). Yet several return figures are manager charts, snapshot weights, or recollections, not independently audited lifecycle returns. Aryzta and the early Cobas drawdown are therefore central evidence, not footnotes: the former shows how leverage, deteriorating economics, and a capital raise can turn patience into thesis inertia, while the latter shows the client-level cost of a long adverse path (Cinco Días, 2018; Cobas NAV export, 2026).

What is transferable is a research and behavior system: define normalized economics, write the disconfirming case, separate solvency from quotation risk, size for a worse path than the base case, and cultivate capital that can wait. What is not transferable is the institutional research network, access, fee-and-client architecture, capacity to influence an issuer, or the benefit of a large team. Cobas's 2024 removal of “author fund” language formalized a collegiate process while Paramés remained CIO, making team attribution and succession—not only stock selection—the decisive open questions (Cobas, 2024; Cinco Días, 2024).

Selling completes the method. A position should leave when the thesis breaks, the price reaches a reasonable estimate of value, or a better risk-adjusted opportunity displaces it—not merely because it rose. This is a demanding value discipline, not a shortcut from low multiples to excess returns (Cobas, 2017; Cobas, 2026).

10 Transferable Lessons, Ranked

1. Normalize the business before valuing the security

A low current multiple can be false comfort at a cyclical peak, while a high reported multiple can hide trough earnings. Paramés's better cases begin with unit economics, mid-cycle margins, balance-sheet survival, and the asset base, then triangulate value through earnings multiples, discounted cash flow, replacement value, or sum of the parts. Cobas explicitly uses discount rates ranging from 6% to 12%, but that range is a judgment framework—not a precision engine (Cobas, 2026). An individual can copy the discipline by publishing a base, adverse, and normalized case before looking at expected upside.

2. Make solvency the gate before upside

Time benefits the investor only if the company can reach normalization without a destructive refinancing, dilution, covenant event, or loss of customer trust. Aryzta's capital raise and Babcock's impairment shock show that asset value and franchise value can change while the investor waits (Cinco Días, 2018; Cobas, 2025). The first page of every thesis should therefore contain liquidity, maturities, covenants, pension or off-balance-sheet obligations, working-capital sensitivity, and an explicit “survival without friendly markets” test. Margin of safety is partly a balance-sheet property, not merely the gap between price and target.

3. Write the kill criteria when conviction is still cheap

Patience is useful only when paired with falsifiability. Before buying, state which facts would show that normalized earnings, management integrity, capital allocation, or industry structure is wrong. Then distinguish a price decline caused by forced selling from a decline accompanied by adverse operating evidence. Babcock shows the value of re-underwriting after a shock; Aryzta shows the danger of allowing the thesis to absorb every disappointment (Cobas, 2025; Cinco Días, 2018). A dated disconfirmation log is more transferable than Paramés's temperament.

4. Fish where institutional constraints create an analytical toll

Small and mid-cap companies, family-controlled firms, cyclicals, holding-company discounts, and asset-heavy restructurings often repel capital because they create benchmark, liquidity, governance, or modeling discomfort. The edge is not neglect itself; it is being able to pay the analytical toll and reject the bad neglected businesses. CAF and Elecnor illustrate different variants of this pattern, while the failures warn that obscurity is not value (Bestinver conference recap, 2011; Cobas first-half commentary, 2024). Individuals have a capacity advantage here, provided they accept wider spreads and slower exits.

5. Treat owner alignment as evidence, never absolution

Reference shareholders and family control can lengthen decision horizons and encourage rational reinvestment. They can also entrench weak governance, related-party behavior, or poor capital allocation. Paramés's documented preference is best copied as a checklist: voting control, insider economics, dilution history, remuneration, treatment of minorities, and reinvestment record (CFA Institute, 2013). “Family-owned” should change the questions, not settle the answer.

6. Size for the adverse path, not the elegance of the thesis

Expected upside, quality, conviction, liquidity, and correlation all belong in sizing. A position that appears independent may share the same hidden exposure to freight rates, European industrial demand, leverage, commodity prices, or value-factor crowding. Individually researched names can still share value, size, cyclical, and liquidity exposures; AQR's Bestinver factor evidence demonstrates the portfolio-level caution (AQR, 2019). An individual should cap correlated scenario loss across the portfolio and assume that an apparently liquid small-cap position becomes expensive to exit during stress.

7. Let price volatility create work, not dictate truth

A falling quotation is neither automatic evidence of impairment nor an automatic bargain. The correct response is a fresh evidence cycle: update liquidity, operating metrics, incentives, competitive position, and valuation; compare the new facts with the prewritten kill criteria; then hold, add, reduce, or exit. This keeps Paramés's useful separation of price from value while avoiding the slogan's behavioral hazard. The March 2020 letter states the former; the 61.52% drawdown [single-source raw reconstruction] is the empirical reason to add the latter (Cobas, 2020; Cobas NAV export, 2026).

8. Design the capital base as part of the strategy

Long-duration value cannot be executed with short-duration liabilities. Client education, clear warnings, compatible fees, liquidity buffers, and honest drawdown expectations are portfolio controls because they reduce forced selling at the worst moment. Paramés has long emphasized a five-to-seven-year client horizon and Cobas rewards longer tenure through its fee schedule (Value Research, 2011; Cobas, 2026). An individual should match equity exposure to actual spending needs rather than an aspirational tolerance questionnaire.

9. Separate repeatable process from favorable regime exposure

Bestinver's result combined security selection, value and small-cap exposures, favorable recoveries, team work, and luck; AQR found that selected systematic styles explained nearly 80% of estimated CAPM alpha, leaving residual alpha statistically insignificant at the 95% level [single-source quantitative study] (AQR, 2019). That is not proof of no skill. It is a warning against paying twice for a factor exposure wrapped in a star narrative. Evaluate the manager against investable style-aware alternatives, across full cycles, while separately judging research quality and behavior.

10. Institutionalize dissent before the founder becomes the process

The record belongs partly to teams: Álvaro Guzmán de Lázaro, Fernando Bernad, and other Bestinver colleagues; a different research group at Cobas (CFA Institute, 2013). Trade commentary itself sometimes attributes a thesis to a colleague. Cobas's 2024 move from legally identified “author funds” toward collegiate decisions is constructive, but it also makes decision rights, escalation, succession, and analyst attribution material (Cobas, 2024). Any investment partnership should specify who can veto, who owns the bear case, and how a position survives its original sponsor.

Style Taxonomy Tags

The tags synthesize the stated doctrine, launch constraints, observed holdings, and factor evidence rather than repeating Cobas marketing language (Cobas, 2017; AQR, 2019).

• Core: long-only public equities; intrinsic-value investing; concentrated European/global value; bottom-up fundamental research; long holding periods; low forecast dependence.
• Security and business preferences: family/reference shareholders; owner-operator economics; small/mid-cap inefficiency; industrials and capital-intensive cyclicals; hidden assets; replacement value; sum-of-the-parts; restructuring and normalization.
• Portfolio behavior: high equity exposure; conviction sizing; low turnover; tolerance for tracking error; liquidity- and capacity-aware; valuation-led selling.
• Risk doctrine: permanent-capital-loss framing; balance-sheet survival; margin of safety; patient-client alignment; no routine market timing.
• Contextual influences: Graham/Buffett/Lynch lineage; Austrian business-cycle worldview; productive-real-asset preference; team-based public-fund platform.
• Required caveats: value-factor and small-cap loading; deep drawdown; star-manager and team attribution; vehicle/benchmark/currency discontinuity; capacity; survivorship; manager-reported trade returns.
• Not the style: net-net-only liquidation investing, systematic factor indexing, short-term macro trading, price-momentum confirmation, market-neutral hedging, or a pure quality-at-any-price compounder strategy.

Regime Dependence

These are conditional inferences from the Bestinver/Cobas split, the trade archive, and adverse cases—not forecasts that any macro regime guarantees performance (AQR, 2019; Cobas, 2024).

Closest and Most-Opposite Investors Already in the Repo

Closest

1. Peter Cundill — closest philosophical ancestor. Both search globally for neglected asset value, do work that screens cannot, accept long recognition periods, and use sum-of-the-parts reasoning. Cundill is more classically Graham, distressed, net-net, and catalyst-oriented; Paramés gives more weight to normalized earnings, owner alignment, and business quality.
2. Anthony Bolton — closest public-fund operating analogue. Both built European contrarian portfolios from company research, hidden assets, recovery potential, staged conviction, and a deep institutional team. Bolton explicitly used charts as a timing and cross-check input; Paramés is more willing to act before price confirmation and endure a longer adverse path.
3. Joel Tillinghast — closest quality-conscious value comparator. Both emphasize cash economics, honest management, unpopular smaller companies, valuation, patience, and capacity. Tillinghast subtracts leverage, fraud risk, and fragile economics more aggressively and diversifies much more broadly; that contrast is especially instructive after Aryzta and the early Cobas drawdown.

Most opposite

1. William J. O'Neil — the cleanest opposite. CAN SLIM demands accelerating growth, price-volume leadership, market confirmation, and rapid loss cutting. Paramés buys valuation gaps before confirmation, accepts weak near-term news, and treats falling prices as prompts for fundamental re-underwriting.
2. Ed Seykota — the process opposite. Seykota makes price, systematic rules, portfolio heat, and small precommitted losses sovereign; Paramés makes business value and patient discretionary judgment sovereign. Both respect survival and behavioral fit, but reach them through almost inverse evidence hierarchies.

Skill, Luck, Team, and Attribution

The durable skill evidence is recurrence across different securities and cycles: normalized-earnings analysis, willingness to own unpopular assets, sustained tracking error, and several cases in which asset value or operating recovery was eventually recognized. The strongest trade archive also shows variation—CAF's long industrial growth, Elecnor's asset realization, and Babcock's impairment-and-recovery path—rather than one lucky template (Bestinver conference recap, 2011; Cobas first-half commentary, 2024; Cobas Babcock review, 2025).

Luck and regime exposure remain material. Bestinver's published record benefited from value/small-cap exposure and recoveries that an investor could select but not schedule. The departure date prevents a clean observation of how the same Bestinver portfolio would have evolved under Paramés, while the Cobas NAV path shows a severe early drawdown followed by recovery under a similar stated philosophy (AQR, 2019; Cobas, 2017; Cobas NAV export, 2026). Endpoint choice changes the story.

Team attribution is not a disclaimer to append after celebrating a star; it changes the unit of analysis. Bestinver theses and interviews were explicitly collaborative, and some celebrated cases were presented by colleagues (CFA Institute, 2013). Cobas now formally describes individual idea generation followed by collegiate portfolio decisions. Paramés says the seven-member investment team votes with one vote each and sometimes acts without him; Gonzalo Recarte runs business strategy, while Paramés remains chairman, CEO, and CIO (Expansión interview, 2024; Cobas, 2024). Public evidence still does not disclose quorum, tie-break, veto, incapacity rules, or a named CIO successor. The cautious conclusion is that Paramés supplied leadership, philosophy, and important security judgments inside two different institutional systems; public evidence cannot allocate security-level profit or research credit precisely enough to construct a personal track record.

Current Adverse and Open-Case Controls

Norway's Finanstilsynet imposed a NOK 200,000 administrative violation penalty on Cobas Asset Management on 5 October 2021 after finding that the manager negligently notified Oslo Børs late when delegated voting rights in Höegh LNG crossed below 15%; the notice also omitted the underlying funds (Finanstilsynet, 2021). This was an entity-level major-shareholding disclosure breach, not a personal sanction on Paramés or a fraud finding. No public appeal or final-payment record was located in the bounded search. It prevents a CNMV-only negative search from becoming a general claim that Cobas had no regulatory penalties.

The latest located official portfolio note, dated 30 June 2026, still showed Golar among the main positions; separate April issuer evidence showed that Atalaya had been reduced but not fully exited, and Técnicas Reunidas remained an open Iberian holding (Cobas, 2026; Atalaya Mining, 2026). These remain open or partially harvested cases, not completed-trade evidence. Aggregate manager voting rights are neither one fund's weight nor Paramés's personal ownership.

What an Individual Can Copy

• Use a one-page thesis with normalized economics, valuation range, financing runway, disconfirming evidence, and sell rules.
• Read annual reports across a full cycle and reconcile cash generation, working capital, capitalized costs, and share count.
• Search neglected small/mid-cap and family-controlled firms without making neglect or ownership a buy signal.
• Track aggregate scenario exposures—freight, commodities, leverage, European cyclicality—not only company labels.
• Maintain a decision journal and a written bear case owned by someone other than the idea sponsor, or by a deliberately separated second review.
• Keep near-term spending and emergency capital outside a strategy that may suffer a multi-year drawdown.
• Compare results with style-aware, investable alternatives and retain failed theses in the sample.

What an Individual Cannot Copy Reliably

• A professional team's site visits, supplier and competitor calls, specialist databases, language coverage, and continuous monitoring.
• Institutional access to management, block liquidity, engagement, or transaction processes.
• Cobas's client-communication, distribution, legal, compliance, and dealing infrastructure.
• The diversification available across multiple vehicles and the bargaining power that can accompany a significant issuer stake.
• Paramés's personal temperament, decades of tacit pattern recognition, or ability to withstand public tracking error.
• Headline historical returns without their original vehicle, fees, benchmark, currency, team, and endpoint. The trade archive likewise does not provide a complete personal cost-basis and cash-flow ledger.

Unresolved Questions

1. Can the Cobas recovery compound across a second full cycle without recreating the leverage and value-trap concentration of its first years?
2. Beyond the disclosed one-member/one-vote process, what are the quorum, tie-break, veto, removal, emergency, and incapacity rules, and who is the named CIO successor?
3. How much of the Bestinver result remains after investable value, size, country, sector, liquidity, and currency controls using consistent euro, net-of-fee data?
4. Can a public, inception-to-date, cash-flow-aware record bridge the different Bestinfond, Bestinver Internacional, and Cobas vehicles without splicing unrelated mandates?
5. Which predeclared operating metric would now force an exit from a deeply discounted holding even if management and the team still estimate high upside?
6. Has the reported capacity/soft-close plan become legally effective, for which vehicles and classes, and at what asset threshold?
7. How will founder control, economics, and research leadership transition, and can Cobas retain the process without either key-person dependence or committee dilution?
8. What are the complete realized, weighted, net-of-fee contribution histories for the celebrated trades and the omitted failures? Public snapshots and manager charts cannot answer this.
9. Was the Norwegian penalty appealed or paid, what procedures changed afterward, and do other foreign-regulator records reveal additional disclosure-control issues?
10. Do current large ownership notifications reflect one coherent cross-vehicle thesis and risk budget, or different mandates aggregated only for regulatory reporting?
11. Which lessons from modules D–G materially alter this A–C evidence base once those separately claimed research modules are completed and independently audited?

Task A - Profile (T0487)

As of 2026-07-18. Sources are ranked by evidentiary value for identity, appointments, vehicle structure, bounded track records, current status, adverse outcomes, and legal dispositions. Issuer sources establish what Cobas or Bestinver reported; they are not independent endorsement of manager skill.

Annotated source map

1. CNMV - Cobas Asset Management manager register and BORME, 9 December 2025 - Primary regulatory records for Cobas's registration, Paramés's president/CEO/director appointment dates and 2025 renewals, and Gonzalo Recarte's later director-general appointment.
2. CNMV - current Cobas Selección fund register and CNMV public-sanctions registry - Primary records for fund number 5075, registration, current prospectus/report history, and the bounded adverse-regulatory search described below.
3. Cobas - Francisco García Paramés biography - Current issuer biography for birth year/place, education, living status, and public chairman/CIO role; promotional “top of rankings” language is not used as independent performance proof.
4. Cobas Selección product page - Primary live source for mandate, classes, fees, benchmark, annual returns, NAV, risk warning, and reported Class C annualized return.
5. Cobas live vehicle table and Class C historical NAV export - Mutable primary sources for 16 July 2026 class/product assets and the full daily NAV series. The €3.494 billion figure sums listed Spanish funds, SICAV, and pensions—not established firmwide AUM; the €1.489 billion Selección total, 7.56% CAGR, and 61.52% drawdown are raw reconstructions.
6. Cobas Selección second-half 2017 report - Contemporaneous primary report for 2017 return, benchmark comparison, concentration, Aryzta weight, portfolio claims, and initial conditions.
7. CNMV/Cobas Selección first-half 2018 report and Q1 2020 report - Regulator-hosted primary reports for policy, risk, fees, contemporaneous holdings, and the -43.0% quarter/-50.4% since-inception trough.
8. Bestinver - Bestinfond factsheet and Bestinver Internacional factsheet - Primary vehicle sources for launch dates, annual history including the 2008 calendar losses, later benchmark/policy changes, and the warning that current inception figures include successor management.
9. CFA Institute - “What Is the Warren Buffett of Europe Up To?”, 2013 - Contemporaneous independent profile with direct access to the Bestinver team; supplies the April 2013 Bestinfond record, mandate evolution, team dates, portfolio data, and career narrative.
10. AQR - “More Superstar Investors: Francisco Garcia Paramés,” 2019 - Independent quantitative reconstruction of April 1994-August 2014 Bestinfond returns, volatility, benchmark outperformance, and factor exposures. Its selected styles explain nearly 80% of CAPM alpha and leave residual alpha statistically insignificant at 95%; the study uses USD excess returns and is not a replacement for the euro fund CAGR.
11. Bloomberg - Bestinver departure, 2014 and manager conference biography - The contemporaneous report gives departure timing, tenure, more than €7.5 billion managed, and the Bestinver-sourced 16% figure; the promoter biography supplies the exact 2,279.2% fund/410% benchmark endpoint claim. The latter is manager-supplied marketing, independently recomputed rather than treated as neutral validation.
12. Cobas - final CNMV approval and four-fund registration - Primary issuer chronology for Cobas's authorization, direct assumption of Selección management, and initial product set; corroborated by CNMV registers.
13. Deusto press sheet, Planeta/Deusto hardback, and Wiley English edition - Publisher records for the first 2016 Spanish/e-book date, later hardback, English edition, education, and career outline; descriptions are controlled biography/marketing copy, not independent validation.
14. Universidad Francisco Marroquín - honorary doctorate, 2016 - Institutional primary source for the award date and ceremony.
15. EL PAÍS - aircraft crash, 2006 - Contemporaneous reporting for the fatal crash, casualties, and Paramés's survival; the profile makes no unsupported claim about its investment impact.
16. Cinco Días - Cobas exceeds €3 billion, 2025 - Independent current-scale and adverse-history source for assets, 2025 inflows, earlier flagship recovery, Aryzta, and the firm's structure.
17. CNMV 2014 complaints report, pp. 52-53, official Cobas team-process communication, and Cinco Días, 2024 - Primary complaint history and issuer communication paired with independent reporting on exit-fee disputes, the later author-fund designation change, team-process rationale, separation right, and Paramés's continuing leadership.
18. El Economista - Aryzta and Dia admission, 2018 and El Español - Aryzta follow-through, 2019 - Contemporaneous adverse sources for manager admissions, continued conviction, Aryzta's 2018 collapse, and the portfolio-relative damage.
19. Iberley commercial reproduction of Supreme Court judgment 941/2025 - Non-official republication of the judicial text for the Acciona civil notice/damages dispute, distinction between labor and shareholder obligations, procedural history, and final disposition; no matching public CENDOJ link was located, so provenance is explicit.
20. Cinco Días - Supreme Court labor award, 2018 - Independent report that the Supreme Court confirmed €13.209 million of variable compensation and rejected the distinct, larger non-compete compensation request.
21. CNMV - exact Cobas Selección prospectus dated 2026-06-22 and El Español - proposed soft close, 2026 - Dated primary prospectus paired with independent reporting on the prospective €500 million capacity threshold; the profile does not call the funds closed without a final effective notice.
22. Cobas - 2025 Rankia interview and Cobas - 2026 BrainVestor interview - Recent direct testimony on philosophy, errors, process evolution, asset preferences, and living/current status; edited issuer pages are used as testimony, not neutral validation.
23. Funds Society - Cobas year-end 2025 scale - Reports manager-supplied €3.454 billion AUM, €420 million 2025 net inflows, and investor count; useful current scale, but not an audited firmwide statement.
24. Meliá Hotels International 2025 governance report filed with CNMV - Issuer governance filing that relays Paramés's declared 50% Esmelle interest and the Esmelle-to-Santa Comba-to-Cobas control chain; it is not a Cobas ownership filing, and control attribution is not personal ownership of fund assets.
25. Acciona H1 2014 results - Controlling-company primary disclosure of €10.198 billion Bestinver AUM on 30 June 2014; this is manager-wide company scale, not one fund or Paramés's personal capital.

Evidence limitations

• Bestinfond's reported 16% figure, CFA's 16.1%, and AQR's 13.5% are not contradictory replicas: they use different periods, currencies, compounding/average-return definitions, and benchmarks.
• Bestinver's historical benchmark did not include net dividends until 2016. The promoter's 410% comparison is therefore not an investor-grade total-return alternative for the Paramés period.
• Bestinfond is a continuing fund. Current since-inception performance includes successor management after September 2014 and cannot be assigned entirely to Paramés.
• Cobas classes share a portfolio but began in different years and bear different fees. Class C is used for the full Cobas-era sequence; later, cheaper classes would create survivorship-like start-date bias.
• Firm AUM reports differ in date and scope. Acciona's €10.198 billion Bestinver total on 30 June 2014 and Bloomberg's more-than-€7.5 billion departure figure are not merged; Cobas's €3.494 billion live-table sum excludes or separately presents Luxembourg/mandate data whose additivity was not established.
• Portfolio outcomes reflect colleagues, factor exposures, markets, currency, flows, and fees as well as Paramés. No public personal-account IRR was found.
• Cobas's internal target values and “upside” estimates are management opinions, not realized returns or independent appraisals.
• Litigation surrounding the 2014 departure involved separate labor and civil claims. Neither establishes securities fraud, regulatory misconduct, or an investment-process violation.
• Targeted searches of the CNMV public-sanctions registry and BOE through 2026-07-18 for exact-name variants of Paramés and Cobas found no matching public sanction. This is a bounded negative search, not proof that no complaint, private claim, non-public inquiry, or differently indexed event ever existed.

Task B - Investment Philosophy (T0488)

Guiding questions

1. What does Paramés say his core value-investing philosophy is, and how much continuity is there between Bestinver and Cobas?
2. What types of businesses and market conditions does he believe public markets misprice?
3. How does Cobas describe idea sourcing, research, valuation, sizing, portfolio construction, and sell discipline?
4. What is risk in this system: volatility, permanent capital loss, client mismatch, capacity, leverage, or thesis error?
5. How did the philosophy evolve from Bestinver through Cobas, and what tensions do adverse performance/legal records reveal?

Annotated source map

1. Cobas official bio - Francisco García Paramés - Primary/current. Confirms chairman/CIO role, value-school identity, Graham/Buffett/Lynch influence, Austrian business-cycle interest, and book authorship.
2. CNMV Cobas AM administrators - Primary regulatory. Confirms Cobas AM registration context and Paramés' board/executive roles from 2017.
3. Cobas Philosophy and Values - Primary doctrine. Defines value investing, intrinsic value, mispricing sources, patience, volatility as opportunity, and permanent capital loss as risk.
4. 2017 Paramés letter to investors - Primary. Foundational Cobas launch letter on unchanged philosophy, purchasing-power preservation, no short-term prediction, high equity exposure, launch portfolio metrics, family/reference shareholders, capacity, and sell discipline.
5. Cobas 2H 2025 commentary - Primary/current portfolio process. Best source for valuation mechanics, normalized-cycle profits, multiple/DCF toolkit, 6%-12% discount rates, target price/upside, 2025 sell/reinvestment examples, PER/ROCE/upside metrics, and turnover.
6. Cobas 1H 2024 commentary - Primary process. Supports company-first, painstaking, bottom-up analysis and continuity over 30 years.
7. Cobas glossary: Value Speak for the Uninitiated - Primary definitions. Useful for intrinsic value, margin of safety, upside, ROCE, expected yield, circle of competence, and sizing language.
8. Cobas March 2020 COVID letter - Primary stress-test source. Supports panic/passive/machine selling as mispricing sources and the distinction between market price movement and business value.
9. Cobas Q4 2019 commentary - Primary but defensive. Useful for what Cobas rejected during the growth-led value drought: fads, expensive expectations, and short-term impatience.
10. Cobas fees - Primary. Supports long-term investor alignment through lower fees for longer-tenured clients.
11. Cobas ESG approach - Primary. Shows ESG integration as risk/opportunity analysis, not exclusion-first doctrine; notes third-party provider in 2019 and PRI signing in 2021.
12. Cobas Internacional FI page - Primary/current product page. Supports value style, equity risk, and unsuitable-under-five-years warning.
13. CNMV Cobas Internacional FI registry - Primary regulatory. Confirms fund registration, manager, depositary, and latest prospectus date.
14. Cobas LUX SICAV International page - Primary/current product data. Supports annual-return path, high equity exposure, geography, sector exposure, and rebound after early Cobas drawdowns.
15. December 2024 Cobas portfolio commentary - Primary portfolio evidence. Supports 2024 returns and top-ten concentration around 36.1%.
16. Value Research 2011 interview - Strong secondary/direct interview. Supports client co-investor framing, communication/education/honesty, five-to-seven-year capital horizon, and Bestinver-era record checkpoints.
17. Value Invest London profile - Secondary conference profile. Supports Bestinver-era performance figures, Bestinver Internacional record, Cobas team context, and European inefficiency framing; not a primary audited composite.
18. Google Books: Investing for the Long Term - Publisher/bibliographic. Supports book provenance, English title, publisher/date context, and broad description of Paramés' long-run record; not granular operational evidence.
19. Fundación Rafael del Pino book-event page - Near-primary event summary. Supports common-sense value investing, psychology, real assets, quality, and family-company preference; not a transcript.
20. El Español March 2026 soft-close report - Secondary/current. Supports reported soft-close capacity discussions with CNMV; cite only as reporting, not final regulatory action.
21. AQR paper: More Superstar Investors? Spain's Value Investors - Independent quantitative/adverse. Supports Bestinver versus Cobas contrast, early Cobas underperformance, style/factor headwinds, and caution against hagiographic performance attribution.
22. Europa Press 2019 report - Secondary adverse/current-at-the-time. Supports 2019 investor-patience appeal and Cobas Internacional drawdown versus benchmark.
23. Cinco Días on Aryzta - Secondary adverse. Supports Cobas' Aryzta exposure, dilutive capital-raise controversy, and value-trap/stubbornness risk.
24. 20Minutos/La Información on Aryzta and Teekay - Secondary adverse. Supports Aryzta/Teekay stress-test framing; use cautiously because it relies partly on Cobas investor-relations messaging.
25. CNMV 2014 complaints report - Primary regulatory. Supports investor complaints after Paramés left Bestinver and CNMV's view that free-redemption triggers were not then covered by fund documents.
26. Cinco Días on fund-author removal - Secondary regulatory/business coverage. Supports 2024 move away from "fondo de autor" framing toward a team process, with Paramés still leading investments.
27. Cinco Días Supreme Court bonus report - Secondary legal coverage. Supports separation of Bestinver/Acciona employment-compensation litigation from investment-philosophy or client-misconduct claims.

Evidence limitations and cautions

• Bestinver performance figures vary by start date, end date, currency, benchmark, and whether they are fund-level, excess-of-cash, or conference-profile figures. This file uses them only to contextualize the philosophy, not as a verified audited composite.
• Cobas upside, PER, ROCE, and target-price figures are management estimates. They are evidence of process, not independent proof of value.
• No CNMV sanction against Paramés or Cobas was found in this pass, but absence from targeted searches is not a legal certificate. The Bestinver/Acciona litigation found here was employment/commercial compensation litigation, not securities fraud.
• Podcast pages, quote compilations, and book excerpts were used only for orientation unless independently supported by primary Cobas/CNMV/interview sources.

Task C - Greatest Trades (T0489)

Guiding questions

1. Which investment has the strongest surviving evidence for Parames's own "single best" designation?
2. Which Bestinver- and Cobas-era cases have reconstructable theses, sizing, adverse paths, and exits?
3. Which published figures are portfolio weights, issuer ownership, contributions, share-price moves, or holding returns, and which can legitimately be called realized P&L?
4. How much of each outcome belongs to Parames, the wider investment team, company management, capital structure, or favorable external conditions?
5. Which apparently successful holdings remain open, and which losses or opportunity costs prevent a survivorship-only account?

Annotated source map

1. Bestinver 2011 conference recap - Detailed secondary/contemporaneous-style recap. Strongest surviving evidence that Parames called CAF probably Bestinver's best investment in 20 years; also supports 2011 sizing and sell-down context. The linked recording is no longer usable.
2. Original 2011 conference video page - Event provenance. Corroborates that the tenth conference recording existed, but does not preserve an accessible transcript or playable video.
3. El Pais 2003 Parames interview - Direct early interview. Names CAF and Aldeasa among preferred holdings and documents the management, price, and dividend framework.
4. Europa Press on Bestinver's CAF stake - Contemporaneous CNMV-based report. Its body says 10.924% aggregate issuer ownership while its headline says 10.294% [disputed]; the underlying CNMV notice was not located. Neither number is a single-fund weight.
5. Value Investor Insight interview archive, 2008 - Direct Bestinver team interview. Best source for the crisis theses and valuation work on BMW and Smurfit Kappa; the latter was principally explained by Fernando Bernad.
6. Cinco Dias on Bestinver's BMW stake, 2010 - Contemporaneous holdings report. Supports aggregate euro exposure, issuer ownership, and fund-specific BMW weight while keeping those measures distinct.
7. CNMV Bestinver Internacional 2010 report - Primary regulatory filing. Supports BMW, Thales, Exor/IFI, and Wolters Kluwer snapshot weights and 2010 issuer-price moves, including Thales's negative path.
8. CNMV Bestinver Internacional first-half 2014 report - Primary, closest filing before Parames's departure. Establishes that BMW and Thales remained large holdings; it is not an exit or cost-basis ledger.
9. Cinco Dias 2016 Parames interview - Direct retrospective. Supports his pride in the terminal BMW/Exor/Thales/Wolters Kluwer cluster and a near-40% recollection [single-source manager recollection], which differs in perimeter/date from the 23.42% four-stock sum in the June 2014 CNMV snapshot [single-source regulatory snapshot].
10. CFA Institute Bestinver profile, part one - Strong secondary reporting. Establishes team dates, concentration, turnover, family-company preferences, and the 2008-09 portfolio drawdown context.
11. CFA Institute Bestinver profile, part two - Strong secondary reporting. Supports Ferrovial reduction, the Thales margin-repair thesis, and contemporaneous discussion of CIR as an error.
12. Cinco Dias 2005 Parames interview - Direct interview. Supports Aldeasa's multi-year accumulation and Parames's peak ownership recollection.
13. Spanish competition decision on Aldeasa - Primary legal/regulatory record. Supplies a 5.217% indirect Bestinver interest before the offer [single-source], preserved separately from Parames's 9% peak recollection [single-source].
14. Autogrill Aldeasa offer result - Primary acquirer release. Confirms successful completion and EUR36.57 offer price [single-source]; does not establish Bestinver's tendered shares or profit.
15. Cobas second-quarter 2018 commentary - Primary manager retrospective. Supplies the Smurfit Kappa entry/exit narrative and Thales's long issuer-price recovery; both require mandate-period and weighted-basis caveats.
16. Cobas Seleccion second-half 2017 report - Primary fund report. Establishes early Elecnor, Exmar, International Seaways, and Babcock position weights and makes clear that these are vehicle snapshots.
17. Cobas fourth-quarter 2017 commentary - Primary manager report. Dates the Exmar and International Seaways entries and explains the switch from Euronav.
18. Cobas January 2018 newsletter - Primary manager research. Details the International Seaways asset, leverage, governance, and shipping-cycle thesis and reports 10.3% aggregate issuer ownership [single-source].
19. Cobas first-quarter 2021 commentary - Primary manager report. Explains the Diamond S/International Seaways all-stock merger, expected synergies, fleet fit, and absence of added leverage.
20. CNMV Cobas Internacional 2019 report - Primary regulatory filing. Supports International Seaways's 3.9 percentage-point 2019 contribution [single-source], a period measure rather than lifecycle P&L.
21. Cobas first-half 2024 commentary - Primary manager report and central numerical source. Its Cobas Seleccion chart reports +220% for Exmar, +143% for International Seaways, and +111% for Elecnor [single-source manager chart] during their time in the portfolio or through 30 April 2024; methodology is not fully disclosed and the stated share-class label conflicts with the displayed ISIN.
22. Cobas second-half 2025 commentary - Primary/current manager report. Confirms Elecnor's full exit and almost-fourfold dividend-inclusive narrative and distinguishes sold positions from still-open Golar and Atalaya holdings.
23. Elecnor 2026 AGM document - Primary issuer corroboration. Verifies the Enerfin transaction's EUR1.560 billion price received and EUR1.0149 billion proceeds after associated tax [single-source issuer figures].
24. Cobas statement on the Exmar bid - Primary manager sale record. Says EUR12.10 before the dividend was a good reflection of fair value and that Cobas would sell its approximately 5.94% stake on those or similar terms [single-source manager statement].
25. Saverex Exmar offer prospectus - Primary transaction document. Confirms the EUR12.10 offer and conditions against which the manager's position can be evaluated.
26. Cobas first-half 2025 commentary - Primary manager case study. Best source for Babcock's original thesis, sizing, greater-than-50% decline, re-underwriting, contributions, valuation-led sale, and June 2025 exit.
27. Babcock 2021 business update - Primary issuer announcement mirror. Adverse evidence for approximately GBP1.7 billion of charges and impairments, including roughly GBP1 billion of impairment [single-source issuer update].
28. CompaniesMarketCap Exmar price history - Secondary market-data series. Used only to establish broad calendar-price adversity; not a Cobas cost-basis or total-return record.
29. MOI Global Parames book notes - Book-derived secondary notes. Used cautiously for Ferrovial's approximate remembered price path and rejected where a primary or contemporaneous source was available.
30. Cinco Dias on Cobas team structure, 2024 - Current secondary reporting. Supports the formal move away from "author fund" framing while Parames remained investment leader.
31. Macrotrends International Seaways price history - Secondary adjusted-price series. Used only for the approximate 57.5% 2019-high-to-2020-low path [single-source]; it is not Cobas's cost basis or a position return.
32. Cinco Dias Bestinver team interview, 2011 - Direct team interview. Fernando Bernad describes BMW as one of Bestinver's best investments; qualitative corroboration, not a return ledger.
33. Reproduction of Bestinver's 2010 presentation - Secondary hosted reproduction. Supports Ferrovial's sum-of-the-parts estimate [single-source reproduction]; used with the conference recap and clearly separated from realized performance.

Evidence limitations and selection controls

• CAF's number-one rank rests on a reported qualitative designation whose original recording is inaccessible, not on comparable transaction economics. BMW is the strongest alternative supported by repeated direct/team praise and position evidence.
• CNMV holdings are point-in-time market values, not average invested capital or lot ledgers. They cannot be multiplied by later stock returns to infer fund profit.
• Cobas's 2024 Cobas Seleccion table describes holding-period returns but does not fully disclose treatment of dividends, foreign exchange, or cash-flow weighting. Each chapter figure is labeled single-source manager data.
• The first-half 2024 commentary labels that return table "Class C" while displaying ISIN ES0124037013, identified elsewhere in the report as Class D; every affected chapter figure flags the inconsistency.
• Bestinver and Cobas cases are team and regulated-vehicle outcomes. No public Parames personal-account P&L or position ledger was found.
• Golar LNG, Atalaya Mining, and Tecnicas Reunidas were excluded from the completed ranking because they remained open at the research date. Maire Tecnimont and Avio were completed but omitted to respect the ten-case cap.
• Aryzta, Teekay, CIR/Cofide, and Debenhams are explicit survivorship controls. Their adverse outcomes or opportunity costs test the same claims about patience, concentration, cyclicality, and business quality.

Task H - Synthesis (T0494)

Guiding questions

• Which parts of the Paramés process survive the split between the Bestinver and Cobas records?
• What can an individual reproduce without a professional research team, open-end-fund platform, or influential issuer stake?
• Which regimes reward normalized-value underwriting, and which turn patience into thesis inertia?
• Who in the completed Canon offers the closest operating comparison and the cleanest methodological opposite?
• What remains unknown because vehicle, team, factor, benchmark, currency, and transaction-level evidence cannot be unified?

Annotated source map

1. Cobas Philosophy and Values - Primary doctrine for intrinsic value, behavioral mispricing, patience, volatility, and permanent-capital-loss risk. Used as the baseline, not as independent evidence of implementation.
2. Paramés's 2017 investor letter - Primary launch statement for company-first value investing, high equity exposure, productive assets, capacity, long horizons, and valuation-led selling.
3. CFA Institute Bestinver profile - Contemporaneous independent reporting with direct team access; supports the pre-departure record, team evolution, concentration, and family-company preference.
4. AQR, “More Superstar Investors: Francisco Garcia Paramés” - Independent reconstruction of the Bestinver period and central adverse evidence on factor attribution. Its USD excess-cash design is not interchangeable with a euro fund CAGR or personal GIPS composite.
5. Cobas Selección Class C daily NAV export - Primary mutable data used for the raw 61.52% drawdown reconstruction. It is one fund class, not firmwide or personal performance.
6. Cobas March 2020 letter - Primary stress-period testimony on forced selling and price versus business value; read alongside, not instead of, the realized drawdown and failed-thesis evidence.
7. Bestinver 2011 conference recap - Best surviving detailed source for the CAF “best investment” designation. The original recording is no longer accessible, so the evidence remains a secondary recap rather than a verified transcript.
8. Cobas first-half 2024 commentary - Primary manager source for the research process and selected holding-return chart. The chart's class label and displayed ISIN conflict, and its methodology is incomplete.
9. Cobas first-half 2025 commentary - Primary manager postmortem for Babcock's thesis, impairment shock, re-underwriting, recovery, and valuation-led exit; not an audited trade ledger.
10. Cinco Días on Aryzta - Contemporaneous adverse reporting on exposure, the disputed capital raise, and the leverage/dilution risk that tests patience.
11. Cobas second-half 2025 commentary - Primary/current process source for normalized-cycle profits, valuation tools, discount rates, target prices, upside, and sell/reinvestment decisions.
12. Value Research 2011 interview - Direct interview for client alignment, communication, and the five-to-seven-year capital horizon.
13. Cobas fee schedule - Primary current schedule showing tenure-related fee structure; supports alignment design, not performance causation.
14. Cobas December 2024 team-process communication - Primary statement that individual idea generation feeds collegiate portfolio decisions and that Paramés remains investment leader.
15. Cinco Días on removal of “author fund” status - Independent reporting on the same governance change, its regulatory context, and the continuing association with the founder.
16. El Español on proposed 2026 soft closes - Current secondary reporting on capacity planning. It does not establish that a closure became effective.
17. CNMV current Cobas Selección register - Primary regulatory checkpoint for the vehicle, current prospectus history, classes, and disclosures through the evidence cutoff.
18. Cobas December 2024 interview - Direct testimony on the seven-member one-vote-each investment process, team decisions without Paramés, his continuing role, and longer-run delegation intent.
19. Cobas appointment of Gonzalo Recarte - Primary issuer source distinguishing Recarte's business-strategy role from Paramés's continuing investment leadership.
20. Finanstilsynet decision regarding Cobas Asset Management - Primary Norwegian regulator decision imposing a NOK 200,000 entity-level penalty for negligent late Höegh LNG threshold disclosure and omission of underlying funds. It is not a personal Paramés sanction or fraud finding.
21. Cobas June 2026 portfolio commentary - Latest official portfolio evidence located; used to keep Golar and Técnicas Reunidas classified as open cases. The note is manager reporting, not an audited holdings ledger.
22. Atalaya April 2026 voting-rights notice - Dated issuer-market notice showing Cobas reduced but retained aggregate voting rights; not one fund's weight or Paramés's personal ownership.

Evidence limitations

• The synthesis uses completed modules A-C. Tasks D-G were already claimed by separate agents but had not produced mergeable files at the cutoff; the unresolved-question list preserves that boundary rather than inventing their conclusions.
• Bestinver and Cobas are different teams, firms, vehicles, class series, benchmarks, and market regimes. Their endpoints cannot be spliced into one audited personal record.
• Trade evidence does not disclose complete weighted purchases, sales, dividends, fees, currency effects, or cash flows. Manager charts and retrospective recollections are labeled rather than treated as audited P&L.
• AQR's factor result depends on the selected style model. It materially qualifies personal-alpha claims but does not prove that security selection, factor choice, implementation, or behavioral persistence had no skill component.
• “Author fund” removal describes formal process language and the interview provides a high-level voting protocol; neither discloses quorum, tie-break, veto, incapacity rules, or succession economics.
• The soft-close report describes prospective capacity management. No final effective closure notice was located in the bounded current check through 18 July 2026.
• A bounded CNMV sanctions search found no personal Paramés match, but it is jurisdiction-limited. The Norwegian entity-level violation penalty prevents any broader clean-record claim.

Task D - Mistakes and Losses (T0490)

Guiding questions

• Which losses belong to Paramés personally, and which belong to Bestinver or Cobas teams and regulated vehicles?
• Which figures are security-price moves, annual contribution, realized position P&L, or NAV drawdowns?
• What did Paramés or the investment team actually admit, and what is only inferred from a later issuer outcome?
• Did the 2008 or 2018-20 crises threaten firm survival, or only capital, liquidity, and reputation?
• Which process changes are documented, and which plausible safeguards were not found publicly?

Annotated source map

1. Cobas Selección Class C daily NAV export - Primary mutable series used to reconstruct the 61.52% maximum drawdown, exact peak/trough, and March 2024 recovery. It is one share class, not personal or firmwide performance.
2. Cobas fourth-quarter 2018 commentary - Primary manager retrospective for historical Bestinver drawdowns, the CIR/Cofide valuation error, 2018 portfolio results, Aryzta/DIA diagnoses, position weights, AUM, inflows, and the unaudited 45-error count.
3. Cobas February 2019 newsletter - Primary manager contribution table separating annual issuer-price moves from percentage-point effects in each Cobas fund. It is not a lifecycle trade ledger.
4. Wiley, Investing for the Long Term - Official publisher record for Paramés's authorized book; provenance anchor for the formative Nissan narrative, not an account statement.
5. MOI Global Paramés book compilation - Book-derived secondary compilation produced with extensive Cobas collaboration. Supplies the public Nissan summary and diversification lesson; exact personal transactions remain unavailable.
6. Cobas third-quarter 2018 commentary - Primary team postmortem for Aryzta's management-alignment, debt, execution, and transition failures, and for the Duro Felguera recapitalization and later-discovered losses.
7. El Economista interview with Paramés - Direct 2018 interview supporting the Aryzta excessive-weight admission and single-source recollection of a total Escada position loss.
8. Cobas second-quarter 2019 commentary - Primary portfolio update documenting an Aryzta reduction after the 2018 crisis. It does not prove a new hard sizing rule or final exit.
9. Cobas first-half 2023 report - Primary later record of the Teekay Corporation exit, Aryzta reduction, and Aryzta exit from Large Cap only. No complete lifecycle P&L is disclosed.
10. Cobas third-quarter 2019 commentary - Primary evidence that Cobas fully exited Duro Felguera; no weighted realized loss or personal Paramés postmortem is supplied.
11. Cobas March 2020 announcement - Primary contemporaneous defense of oil-service holdings including Valaris. Used as an ex-ante claim tested against the later exit and restructuring.
12. CNMV Cobas International first-half 2020 report - Primary regulatory filing reporting the Valaris/KT/Petra Diamonds exits and their combined prior weight below 1%; it does not isolate Valaris's loss.
13. Valaris restructuring agreement - Primary issuer announcement for the 19 August 2020 Chapter 11 filing and contemplated old-equity treatment. Issuer impairment is not proof Cobas held through filing.
14. Valaris restructuring completion - Primary issuer confirmation that reorganization eliminated $7.1 billion of debt; corroborates balance-sheet failure, not Cobas's exact realized P&L.
15. El Confidencial interview with Paramés - Direct 2009 admission that he missed the speed of McClatchy's print-to-internet transition. No Bestinver position return is given.
16. McClatchy SEC filing - Primary regulatory evidence of McClatchy's 2020 Chapter 11. It validates structural danger but not continued Bestinver ownership or a Paramés bankruptcy loss.
17. Cinco Días Paramés interview - Direct retrospective identifying Inditex as an error of omission and describing the near-entry in 2009. It cannot support a numeric counterfactual return.
18. Cobas fifth annual investor conference - Primary manager report of positive March 2020 inflows and approximately 97% asset continuity. Relevant to the near-death test, but not independent solvency verification.
19. Paramés March 2020 letter - First-person stress-period account of client selling and price-versus-value reasoning; paired with the actual NAV loss rather than treated as proof of recovery.
20. AQR, “More Superstar Investors: Spain's Value Investors” - Independent September 2020 analysis of early Cobas returns, including the November 2016-June 2020 regression used in the chapter. Its selected factors, hindsight, and gross-implementation assumptions limit any no-skill conclusion.
21. Babcock 2021 business update - Primary issuer-announcement mirror for major charges and impairments. It documents the company shock, not Cobas's response.
22. Cobas first-half 2025 commentary - Primary manager case study documenting continued purchases through much of 2023, later sales, and the June 2025 exit. It shows active resizing and eventual exit, not a formal re-underwriting procedure or a proven Aryzta-caused reform.
23. Finanstilsynet decision - Primary Norwegian regulatory decision imposing a NOK 200,000 entity-level penalty for negligent late and incomplete Höegh LNG threshold disclosure. It is not a personal Paramés sanction or fraud finding.
24. Spanish Supreme Court civil ruling summary - Reproduction/summary of the 2025 civil ruling reversing the pre-notice-duty finding and reinstating dismissal in a company/shareholder contract dispute. It was not an employment judgment or securities-misconduct case.
25. Bestinfond official factsheet - Primary current factsheet reporting Bestinfond's -42.14% calendar-2008 return and +60.20% 2009 return [single-source official series]. These are annual returns, not maximum drawdown.
26. Bestinver International official factsheet - Primary current factsheet reporting Bestinver International's -44.71% calendar-2008 return and +71.85% 2009 return [single-source official series]. It independently anchors crisis severity but not the later retrospective peak-to-trough calculation.

Evidence limitations and classification controls

• Nissan is the only quantified personal loss discussed, and even it comes from an authorized book narrative rather than statements. Bestinver and Cobas results belong to teams, vehicles, and specific fund classes.
• Manager-reported price changes, fund contribution points, NAV returns, maximum drawdowns, and realized transaction losses are not interchangeable. No complete weighted cost, sale, dividend, foreign-exchange, and fee ledger was found for the named public-fund positions.
• The 2007-09 Bestinver drawdown table is a Cobas retrospective citing Bloomberg. The reported 61.8% decline and subsequent 153.6% rise do not quite restore the prior peak if their endpoints are exactly sequential.
• Aryzta was an admitted and severe 2018 mistake, but later reductions and a Large Cap exit do not establish every Cobas vehicle's aggregate lifecycle result. Teekay Corporation and Teekay LNG are distinct securities and the -5.7-point figure combines them only in Cobas International for 2018.
• Valaris and McClatchy bankruptcies validate issuer-level risks, not continued fund ownership through bankruptcy. Cobas reported exiting Valaris before the issuer's August 2020 filing; its realized fund loss remains unknown. Duro's annual contribution and exit do not disclose a realized lifetime return.
• No fund or firm insolvency, redemption gate, rescue, forced liquidation, or personal financial ruin was located. A 61.52% share-class drawdown is a severe capital and reputation event, not by itself a near-death finding.
• The Norwegian penalty applies to Cobas Asset Management as an entity. The Bestinver dispute was civil. Bounded regulator searches cannot establish a universal clean record.
• Cobas's 45-error count is a manager-defined, unaudited self-assessment. It does not establish a 90% hit rate, and equal-weight error counts obscure the disproportionate effect of one oversized position.

Task E - In Their Own Words (T0491)

Guiding questions

1. Which words can be traced to Paramés himself, rather than to later admirer summaries or unattributed quote pages?
2. Which official Cobas/Bestinver texts are personal voice, and which are institutional/team voice?
3. How does his language change across Bestinver, Cobas launch, early Cobas drawdown, Covid-19, recovery, and the 2024-2026 team-process period?
4. Which quotes are useful but potentially misleading without adverse context around Aryzta, Teekay, Babcock, early Cobas underperformance, litigation, and capacity?
5. Which video, podcast, or book leads should later agents revisit for exact transcripts or page-numbered citations?

Annotated source map

1. Deusto/Planeta public sample of Invirtiendo a largo plazo - Primary excerpt. Source for personal-history and reading/psychology language; partial sample only, so not treated as full book access.
2. Wiley English edition record and Wiley excerpt - Publisher/provenance plus English interview excerpt for quality-at-reasonable-price language; useful but not a full substitute for the book.
3. Fundación Rafael del Pino book dialogue - Official event page for the 2016 public dialogue with Carlos Rodríguez Braun; strong context, but no complete transcript was available in the opened page.
4. Europa Press book-quote article - Secondary quote-lead article based on the book; useful for locating maxims, but not a final authority where the public book sample is silent.
5. Value Investor Insight scan via Scribd - Direct Bestinver-team interview, third-party hosted; used for time-horizon, basic-process, and circle-of-competence language with explicit provenance caveat.
6. Value Research interview - Direct 2011 interview. Best source for co-investor framing, five-to-seven-year client horizon, education/communication, Austrian economics, Arcelor luck, and "we don't fight management."
7. CFA Institute profile/interview, part one - Contemporaneous direct-access profile for concentration, business safety, family-company preference, turnover, and Bestinver-era portfolio context.
8. CFA Institute profile/interview, part two - Companion source for mistakes/process context, including Pescanova and Spanish-market discrimination claims; mainly contextual rather than quote-list material.
9. CNMV Bestinver Internacional 1H 2014 report - Primary regulatory filing. Best Bestinver-era official wording on circle of competence, valuation work, portfolio rotation, and price/value.
10. Paramés letter to investors, 7 March 2017 - Signed primary Cobas launch letter. Core source for philosophy continuity, sell discipline, capacity, team, communication, and co-investment.
11. Funds Society on Value School launch - Reported direct remarks on investor education, personal biases, passive management, and ecosystem-building.
12. Funds Society on second Cobas annual conference - Reported direct remarks on fixed-income optionality, expensive quality, shipping, UK exposure, and contrarian buying.
13. Cobas fourth-quarter commentary, 2018 - Primary drawdown commentary after Aryzta and Teekay losses; useful official wording on errors, patience, valuation confidence, and behavioral client risk.
14. Funds Society on third Cobas annual conference - Reported direct remarks on 43 investment errors, bear-cycle interpretation, and recovery expectation after early Cobas underperformance.
15. Cobas second-quarter commentary, 2019 - Primary official commentary with personal-letter framing and useful lines on short-term irrationality, time, and value recognition.
16. Cobas third-quarter commentary, 2019 - Primary official commentary on erratic market behavior, sector stress, and contrarian use of volatility.
17. Paramés COVID letter, 16 March 2020 - Signed primary stress letter separating business value from market price and explaining sector-by-sector Covid implications.
18. Paramés COVID follow-up, 23 March 2020 - Signed primary follow-up on portfolio expectations, LNG contracts, financials, buybacks, and short-term suffering versus long-term value.
19. Paramés fee-structure letter, December 2020 - Signed primary client-alignment source on lower fees, seniority preservation, and investor trust.
20. Cobas third-quarter commentary, 2020 - Primary official philosophy restatement clarifying that value investing is not merely low multiples or declining industries.
21. Cobas fifth annual investor conference and Funds Society 2021 conference coverage - Official page plus reported remarks on fees, ESG/impact, BrainVestor, pension plans, and Asia research capability.
22. Cobas second-quarter commentary, 2021 - Primary official wording on cash generation as the long-run driver of equity value.
23. Paramés-authored Cobas blog, April 2022 - Direct author source on uncertainty, value-versus-growth history, and macro humility.
24. Cobas fourth-quarter commentary, 2022 - Primary official commentary on preparation instead of recession timing and not needing macro certainty.
25. Cobas second-half commentary, 2023 - Primary commentary with a section explicitly by Paramés on Asia; also source for non-prediction, patience, and enterprise mentality.
26. Cobas first-half commentary, 2024 - Primary process source for idea study, full models, team debate, committee voting, sizing, and valuation-led rotation.
27. Cobas repost of Expansión interview - Direct recent interview source on Spanish equities being attractive/penalized and passive/index flow opportunity; the embedded full interview PDF should be preserved in future F/G tasks.
28. Cobas first-half commentary, 2025 - Primary case-study source for Babcock and the concise current process summary: buy good companies when rejected and sell when fashionable.
29. Cobas second-half commentary, 2025 - Primary/current English process source on normalised-cycle profit, multiples/DCF, discount-rate ranges, target values, and disclaimers.
30. Cobas Rankia interview page, 2025 and Cobas BrainVestor interview page, 2026 - Official recent video pages. Indexed for current testimony on psychology, errors, passive management, real assets, and teaching investing, but not quoted without transcripts.

Evidence limitations and exclusion notes

• Several official video and podcast pages were located but not transcribed in accessible text. They are indexed as primary materials, not mined for unverifiable quotes.
• Quote aggregators and polished English paraphrase pages were excluded unless the wording could be traced to a book excerpt, direct interview, official letter, or official commentary.
• The 2017 Cobas launch letter and later 2020 signed letters are direct personal voice. Quarterly and semiannual Cobas materials are official firm/team voice unless the document marks a section as Paramés's own commentary.
• The Value Investor Insight source was available through a Scribd-hosted scan rather than the original VII archive. It is used because the text is visible and consistent with other profiles, but the hosting provenance is weaker than Cobas, CNMV, or the original outlet.
• Adverse context should travel with attractive maxims: Aryzta, Teekay, Babcock, the 2018-2020 Cobas drawdown, the Bestinver/Acciona litigation, and the 2021 Norwegian Cobas entity-level late-disclosure penalty prevent clean aphorism-only treatment.
• Bounded current/adverse checks through 2026-07-18 found no public personal Paramés sanction in opened CNMV/legal sources, but did find an entity-level Cobas penalty in Norway. That distinction should be preserved.

Task F - Key Writings (T0492)

Guiding questions

• What did Paramés verifiably author alone, sign personally, say in a direct interview, or issue through a team?
• Which Spanish and English records are editions of one book rather than distinct works?
• What is each work's central thesis, key ideas, and best reading path?
• Which external works best test the autobiography, factor attribution, team dependence, and later Cobas record?
• Which derivative, promotional, translated, or inaccessible materials should not be counted as independent corroboration?

Annotated source map

1. Planeta/Deusto Spanish book record - Official Spanish publisher record for Invirtiendo a largo plazo. Establishes the original work and edition chain; publisher claims are marketing, not independent performance evidence.
2. Authorized Planeta excerpt - Official 28-page Spanish sample containing the copyright page, full contents, and opening argument. It does not provide public access to all cases or appendices.
3. Wiley English edition - Official English publisher record, chapter contents, dates, and identifiers. This is the 2018 English edition of the same sole-authored book, not a second distinct work.
4. inBestia review - Named practitioner's detailed review of the Spanish edition, useful for organizational, case-detail, mistake-analysis, and editing criticisms; not peer-reviewed.
5. Procesos de Mercado review - Substantial published review by Jesús Huerta de Soto. Intellectually sympathetic yet adverse on historical selection, equity claims, fees, and star-manager instability; not performance verification.
6. Paramés letter to investors, 7 March 2017 - Signed primary founding manifesto for Cobas, covering philosophy continuity, equity exposure, selling, capacity, ownership, team, communication, and alignment.
7. Paramés COVID-19 letter, 16 March 2020 - Signed primary crisis letter. Records real-time price/value, permanent-impairment, company, liquidity, and redemption reasoning; later outcomes are required to judge it.
8. Paramés COVID-19 follow-up, 23 March 2020 - Signed one-week update based on further company contact. Strong evidence of changing assumptions, but not proof those revisions were accurate.
9. El País interview - Direct early interview on appraisal value, cash, entry/exit discipline, bottom-up research, and rejection of technical analysis.
10. Value Research interview - Direct 2011 interview on co-investment, long-horizon clients, communication, Austrian economics, and the move toward business quality.
11. Cinco Días interview - Direct 2016 retrospective bridging Bestinver and Cobas and identifying errors as well as continuity.
12. Value Investor Insight archive - Original archive record for a 2008 interview with Paramés, Álvaro Guzmán, and Fernando Bernad. Team-direct and partly access-controlled; not solo authorship.
13. Fundación Rafael del Pino book dialogue - Official 2016 event page and near-primary companion to the book; summary rather than complete transcript.
14. Expansión interview PDF - Direct recent interview reproduced by Cobas, useful for the seven-member, one-person/one-vote process and delegation; not an authored treatise.
15. Rankia event page - Cobas-hosted 2025 direct-feature page for late-career reflection. Its edited description is not a transcript.
16. BrainVestor event page - Cobas-hosted April 2026 direct feature on doubt, bias, confidence, and emotional control; description is firm-edited.
17. AQR, “More Superstar Investors: Francisco Garcia Paramés” - Best independent quantitative reconstruction of Bestinfond. Its factor selection, gross implementation, USD convention, and single-fund scope limit inference about personal alpha.
18. AQR, “More Superstar Investors: Spain's Value Investors” - Best adverse analysis of early Cobas and explicit team-attribution control. Ends in June 2020 and therefore omits the later recovery.
19. CFA Institute succession case - Best external case study of star-manager dependence, redemptions, communication, team continuity, and inherited holdings; practitioner analysis rather than a formal report.
20. Morningstar Spain manager assessment - Balanced 2019 assessment emphasizing that early underperformance and a backward-looking rating did not yet establish a full-cycle skill verdict.
21. Strictly Value English book review - Anonymous practitioner review that usefully connects book doctrine to leverage and deteriorating-business risk; not formal research.
22. InvestingByTheBooks review - Named reviewer's clear English chapter map and reading guide. Favorable and mainly derivative of Paramés's claims.
23. Cobas Paramés author archive - Official firm-controlled ceiling check showing the book and two Cobas-hosted bylines. It does not capture the separate Value School byline or publisher-recorded forewords.
24. Cobas Peter Lynch/book-collection essay - Individually bylined 2017 post containing a substantive Lynch review. It substantially overlaps the Batiendo a Wall Street foreword and is counted once intellectually.
25. Planeta/Deusto Batiendo a Wall Street record - Official publisher record for the edition containing Paramés's foreword; establishes the second publication context, not an independent text.
26. Value School Austrian-economics essay - Individually bylined 2018 condensation of the book's Chapter 4. Primary for Paramés's chosen framework, not proof of economic consensus.
27. Cobas value-investing essay - Individually bylined April 2022 essay on value-growth dispersion, rates, scenarios, and timing uncertainty; dated manager advocacy rather than independent forecast validation.
28. Paramés fee-structure letter - Signed December 2020 operational letter on tenure-linked fees and investor seniority. Bibliographic material, but not a major investment essay.
29. Marcial Pons Alicia en Wall Street catalog - Strong catalog evidence that the third edition contains a Paramés foreword. No licensed full foreword was retrieved, so its detailed ideas are not reconstructed.
30. Planeta/Deusto Inversión de patrimonios record - Official publisher record for Jaime Alonso Stuyck's 2022 book and its Paramés foreword.
31. Authorized Inversión de patrimonios excerpt - Publisher-authorized front matter verifying the contents-page attribution to Paramés; the public excerpt omits the foreword itself.
32. Value School “Quiénes somos” - Institutional page ending with a Paramés-signed appeal for financial education and patient investing. It is a signed statement, not investment research.

Evidence limitations and exclusion notes

• One Spanish original and its English and other translations count as one distinct book. Retailer blurbs and foreign-edition descriptions share publisher provenance and are not independent corroboration.
• Three individually bylined essays, four signed Cobas letters, a signed educational appeal, and several forewords were found beyond the book. Overlapping essays/forewords and revised editions are consolidated rather than multiplied.
• The book is autobiography and selected manager testimony. Bestinver and Cobas outcomes belong to teams, firms, vehicles, classes, and regimes; none is an audited personal composite.
• A signature establishes Paramés's communicated first-person view, not necessarily sole origination of every security analysis. Institutional commentaries and conference materials remain team evidence unless individually bylined.
• The AQR studies are the strongest quantitative counterweight, not final skill verdicts. Model choice, hindsight, implementation, currency, fees, sample length, endpoint, and fund-versus-person distinctions all matter.
• No independent book-length biography was located. The best “about” literature is a combination of factor studies, a succession case, reviews, and dated press assessments.
• MOI compilations, Goodreads, quotation pages, promotional extracts, retailer biographies, and lightly sourced wealth profiles were excluded as derivative, copyright-sensitive, or insufficiently verified.
• Direct video/audio pages without accessible transcripts are indexed for chronology but do not support exact quotations. All book and review ideas in the chapter are paraphrased.

Task G - Mental Models (T0493)

Guiding questions

• Which models are Paramés's direct statements, adaptations from earlier investors/economists, Cobas or Bestinver team practices, or Canon reconstructions?
• How does the public method move from capital fit and idea generation through research, valuation, entry, sizing, monitoring, and sale?
• Which numerical rules are dated or vehicle-specific, and which important risk rules are not publicly disclosed?
• Where did price/value, normalization, owner alignment, patience, and bottom-up diversification work or fail?
• What can an individual reproduce without Cobas's team, access, infrastructure, capital base, and institutional governance?

Annotated source map

1. Wiley, Investing for the Long Term - Official English record for the sole-authored synthesis; provenance anchor for Paramés's career, Austrian lens, equity process, and psychology.
2. Authorized Planeta excerpt - Official Spanish sample and full contents. Partial access only; chapter ideas are paraphrased rather than reconstructed as quotations.
3. Paramés Austrian-economics essay - Individually bylined direct source framing Austrian economics as a conceptual compass rather than an exact forecasting engine.
4. Cobas glossary - Institutional definitions for shares as ownership, intrinsic value, upside, margin of safety, expected yield, long term, and circle of competence.
5. Cobas work-method article - Historical institutional process source for simplicity, ROCE, management/owner checks, normalized profit, the 15x P/E convention, structural-change test, and sell reasons. Not a sole Paramés byline or current universal formula.
6. Cobas valuation-method article - 2023 institutional explanation of current idea sourcing, no-consensus-screen practice, minimum ten-year history, normalization, qualitative checks, margin of safety, and sell discipline.
7. Cobas philosophy - Primary institutional doctrine for behavioral mispricing, intrinsic value, patience, and permanent capital loss; not independent implementation evidence.
8. CFA Institute Bestinver profile - Contemporaneous team-access reporting on quality evolution, ROCE, concentration, owner alignment, and Austrian influence.
9. Value Research interview - Direct 2011 interview for co-investor framing and five-to-seven-year capital guidance; explicitly discusses the three-manager Bestinver team.
10. Expansión interview PDF - Direct current evidence on seven-person voting, idea development, daily team decisions, founder delegation, and capacity intentions.
11. Cobas team-process communication - Primary institutional confirmation that individual idea generation feeds a collegiate process while Paramés remains investment leader.
12. Cobas International product page - Current product-level risk and suitability control, including the below-five-year warning; not a universal recovery guarantee.
13. Cobas second-half 2024 commentary - Current team valuation source for normalized-cycle profit, multiples/DCF, mostly 6%-12% discount rates, model inputs, and VaR definitions.
14. Cobas first-half 2024 commentary - Primary team process and Maire case: majority vote, small initial weight, staged sizing, monitoring, target-value reallocation, and internal-estimate limits.
15. Paramés launch letter - Signed direct source for high equity posture, purchasing-power objective, valuation-led selling, capacity, co-investment, and communication.
16. El Español soft-close report - Current secondary report that Cobas was working with CNMV on soft closes. Proposal/process evidence, not proof of effective closure.
17. CNMV current Cobas Selección register - Primary regulatory checkpoint for current fund documents and classes; used to bound the soft-close claim.
18. Cobas fees - Primary current tenure-linked fee schedule. Evidence of capital-base design, not security-level risk reduction.
19. Bestinver conference recap - Secondary surviving evidence for CAF's qualitative “best investment” designation; original recording remains unavailable.
20. Cinco Días Paramés interview - Direct retrospective identifying Inditex as an error of omission; no calculable counterfactual return.
21. Cobas fourth-quarter 2018 commentary - Primary team postmortem and portfolio context for Aryzta, DIA, historical drawdowns, and error reasoning; manager claims are not audited trade ledgers.
22. El Economista Paramés interview - Direct 2018 admission that Aryzta's portfolio weight was excessive; access-controlled in automated checks.
23. Cinco Días on Aryzta - Contemporaneous adverse report on exposure, leverage, dilution, and the disputed capital increase.
24. El Confidencial interview - Direct 2009 admission that McClatchy's structural print-to-internet change happened faster than expected.
25. Cobas first-half 2025 commentary - Primary manager case study supporting continued Babcock purchases, later sales, and final exit; not proof of a formal re-underwriting rule or audited lifecycle P&L.
26. AQR Spain value-investor study - Independent adverse factor analysis of early Cobas. Short sample, factor choice, hindsight, and gross implementation limit any skill verdict.
27. Paramés March 2020 letter - Signed stress-period price/value reasoning. Ex-ante testimony to test against outcomes, not proof of recovery.
28. AQR Bestinver study - Independent reconstruction associating much of the Bestinver record with selected factor exposures; model design prevents a no-skill conclusion.
29. BrainVestor event page - Official 2026 event page identifying confirmation bias, overconfidence, conviction, and team-error control as topics. Edited description, not a transcript or control audit.
30. GuruFocus Paramés interview - Direct 2019 interview on business understanding, ROCE, free-cash-flow yield, leverage, ownership, and comparison with portfolio alternatives; secondary transcript hosting.
31. Cobas third-quarter 2018 commentary - Contemporaneous primary team account of Aryzta sizing, debt, normalized-margin assumptions, governance, and the proposed capital increase; manager estimates rather than audited attribution.
32. CNMV Cobas report, first half 2020 - Regulatory vehicle report recording the Valaris exit during the first half of 2020; it bounds the bankruptcy chronology but is not a complete transaction ledger.
33. Valaris restructuring release - Issuer primary source announcing the August 2020 Chapter 11 filing and debt-reduction agreement; company testimony, not evidence of Cobas ownership at filing.

Evidence limitations and classification controls

• Paramés did not publish a numbered mental-model canon. Descriptive labels are classified as direct, adapted, institutional/team, or Canon-reconstructed instead of being placed in his mouth.
• Historical Bestinver and current Cobas processes belong to different teams and institutions. Cobas's 15x normalized-P/E explanation and later multiples/DCF with 6%-12% rates are dated practices, not simultaneous universal formulas.
• Manager target values, upside, ROCE, normalized profit, the >90% sell-reason claim, and case charts are internal estimates or self-reports. They demonstrate process but do not validate valuation accuracy or transaction P&L.
• Public evidence does not disclose a universal main-fund position cap, correlated-exposure budget, investee leverage ceiling, stop loss, mandatory premortem, formal kill criteria, or complete committee rules. Absence from bounded public research is not proof no internal control exists.
• Price change, target-value change, fund contribution, position return, realized P&L, and NAV drawdown remain distinct. Cases are used to test models, not splice a personal return series.
• Current soft-close planning is prospective until an effective document is found. Current videos without transcripts support topics and chronology, not exact operational claims.