Seth Klarman on Mistakes & Learning

The 13F record also reveals Klarman's willingness to hold cash even within the equity portfolio. Across multiple cycles, the disclosed long book has consistently represented a fraction of the firm's total assets under management, with the remainder held in cash, distressed debt, private positions, and real estate that do not appear in the public filing. He has argued that the 13F is therefore an incomplete view, and that drawing conclusions about the firm's market timing from the equity disclosures alone is misleading. The firm's true exposure to any risk factor is the sum of all asset classes, not the long-equity slice visible to the public. That said, the visible pattern is consistent with the broader philosophy: the equity book is increased during market dislocations and trimmed as valuations become stretched. The 13F snapshots during the post-2008 recovery and again during the 2020 dislocation show Baupost adding to positions while many peers were reducing exposure - the same contrarian disposition that characterizes the firm's distressed-debt work showing up, on a lag, in the public equity record.

Because if you’re too exposed, if you’re getting margin calls, if you’re getting massively redeemed because you took the wrong clients and they’re short-term, then you’re going to be out of commission on that day. So to be around on that day and be able to do what we do — we just did the same thing we do every day, you did it in a little bigger size. [35:45] BARRY RITHOLTZ: So I’m kind of fascinated by the dynamic tension between fundamental bottoms-up research on a credit-by-credit or equity-by-equity basis versus the top-down. You’ve said that you really don’t think about markets or investing from a top-down perspective, but it seems that everybody who panicked, everybody who helped create those distressed assets, was either responding or over-responding to the top-down environment. How do you look at that sort of environment? [36:21] SETH KLARMAN: There are several layers to that. First of all, people were responding to all kinds of things. They were responding to redemption requests by their mutual fund shareholders. They were responding to credit downgrades, so it wasn’t just nervousness that things are going to be bad — this bond is no longer investment grade, and maybe my mandate is I can only own investment-grade bonds; or this bond has defaulted and I can no longer hold it. So you have forced selling all over the place.

[38:36] BARRY RITHOLTZ: So let’s talk a little bit about cash. I think a lot of investors look at cash as a drag on their performance — the net return is usually zero or close to zero relative to inflation. How do you think of cash? It’s always been such a historically important part of your toolkit. What sort of optionality does it create, versus the career pressure of staying fully invested at all times? [39:06] SETH KLARMAN: You’re nailing it with your question. You’ve covered all the parts of holding cash. Cash can be valuable optionality. Just imagine you have a reasonably concentrated portfolio, and a large position or two comes off the books. Should you put it to work in a nanosecond? Or can you wait until something really interesting comes along? That’s the origin of us holding cash — positions would come off and we’d hold some cash until something great came along. But not just a couple of percent. With concentrated positions, we have 5% and 10% positions in the portfolio. When two or three of them come off, cash goes from next to nothing to 15% or 20%. So that’s the origin, that’s how we got started with the idea that we would hold some cash from time to time. That said, I would accept that I almost certainly made a mistake in holding cash to that extent. There were times when we were 30% cash and even higher, and I viewed it as valuable optionality.

We really don’t know what asset class is going to do, because we think that’s very time-specific and very valuation-dependent. Rather, we see what’s available right this second. By looking bottom-up, opportunity after opportunity, I think we can paint a really clear picture. So right this second, real estate’s been in tough shape since COVID, especially commercial office. People started working from home and that hasn’t fully returned, and in certain markets especially there’s too much space. A lot of people that have been in real estate have not done that well — a lot of people got in at a wrong vintage, and a lot of properties have become structurally obsolete. So that sounds like a mess — why would you touch it? But it also means that competition is hardly looking. So we think there are opportunities right now, for example in assisted living. The population is aging. You can make a very strong case for fundamentals. Rents haven’t moved up in years, and there’s probably pent-up growth in rents to come. COVID was obviously a giant problem, because any facility tended to empty out as people pulled their relatives out to save their lives during COVID, understandably. A lot of newly built facilities from that era, from 2021, 2022, never got filled, and a lot of them have run into bankruptcy or financial distress. So it’s been an opportunity to build a position in an area with strong fundamentals.

In his 2023 investor letter, Klarman blamed the Federal Reserve's easy-money response to 2008 for more than a decade of distorted asset prices. He argued that the central bank's suppression of interest rates had driven investors up the risk curve, into leverage, and into lower-quality assets in a search for yield that the policy itself had made impossible to find safely. His critique was structural rather than cyclical. The Fed had not merely lowered rates but had committed, implicitly, to preventing large losses in financial assets. That commitment changed the behavior of every other actor in the system: corporates leaned on cheap debt to buy back stock, private equity bid up asset prices using leverage that depended on low rates forever, and retail investors learned to buy every dip. Klarman's conclusion was that the unwinding of this regime would not be orderly. As rates normalized, the entire scaffolding of leverage built on the assumption of permanently free money would have to be repriced. He framed 2022 as the first installment of that repricing, not as a one-off shock, and warned that the second-order effects - bankruptcies, distressed sales, redemption pressure at leveraged funds - would compound over several years rather than resolve in a single quarter.

Klarman's 2008 letter catalogues twenty lessons from the financial crisis that he argued investors had failed to learn. His central observation was that an entire generation of market participants had come to believe that central-bank action and innovation in financial engineering had eliminated the possibility of system-wide loss. He saw that belief, not the housing market itself, as the precondition for the eventual collapse. He argued that the mania had been built on layer upon layer of low-quality credit, each tranche of which had been rated by agencies paid by issuers, insured by counterparties whose own balance sheets were impaired, and bought by investors who had not read the offering documents. The complexity was not accidental - it concealed the absence of true underwriting. Each link in the chain assumed someone else had done the diligence. The lesson Klarman drew was that the absence of recent losses breeds the conditions for the next loss. He warned that even after 2008, the structural response - bailouts, quantitative easing, fiscal stimulus - would teach market participants that downside had been socialized. That expectation would, in turn, set up the next episode of moral hazard. He treated the post-crisis regime as the soil in which the next bubble would grow, not as a guarantee that bubbles could no longer occur.

Klarman's 2008 lessons also include a sustained critique of modern risk management. He argued that value-at-risk and similar quantitative frameworks had confused measurable recent volatility with the risk of permanent loss. By defining risk as the standard deviation of recent returns, the industry had effectively asserted that whatever had not recently happened could not happen. He pointed out that this is precisely backwards. The events that ruin portfolios are, almost by definition, those that did not appear in the recent sample. A risk model calibrated on the prior decade's data is most useful when least needed - in calm markets - and most dangerous when most needed, because the regime change it cannot anticipate is precisely the regime in which it is asked to perform. The proper approach, in his view, is scenario analysis grounded in fundamental downside cases rather than statistical extrapolation of normal times. Baupost's risk process therefore asks what a position is worth if the worst plausible fundamental scenario occur, and treats that as the binding constraint. Volatility is treated as a feature, not a measure of risk; the only risk that matters is the one that prevents the investor from holding the position through the recovery - which is, structurally, leverage or liquidity mismatch rather than price fluctuation.

In the same interview, Klarman reflected on the recurring pattern by which markets convince each generation that this time is different. He noted that the phrase appears, almost without fail, in the late stages of every bubble: the technology bubble of the 1990s, the housing bubble of the 2000s, the crypto and special-purpose-acquisition-company episodes of the early 2020s. The substance changes; the rhetorical move does not. He argued that the pattern is rooted in the institutional memory of the market. Each generation enters finance without having lived through the prior cycle's deflation. By the time the prior lesson would have been useful, the people who learned it have retired, and the new entrants have only seen the rising part of the curve. The phrase 'this time is different' is, in this view, less an analytical claim than a confession that the speaker has not studied the comparable prior episode. Klarman's prescription was deliberately old-fashioned: read the histories, study the prior episodes, and notice that the architectural similarity across cycles is greater than the surface similarity of the underlying assets. An investor who has read the 1929, 1969-1974, and 1990 episodes will recognize the shape of the 2008 episode while it is unfolding, rather than treating each new development as unprecedented. The willingness to read backward is, in his framing, an underappreciated source of edge.

In his reflections on his own errors, Klarman distinguishes mistakes of analysis from mistakes of process. An analytical mistake is being wrong about facts; a process mistake is reaching for risk because the environment punished patience. He treats the latter as far more dangerous because it tends to compound, eroding the discipline that produced the firm's edge in the first place. He is candid that the most common mistake at Baupost has been under-investing when prices were genuinely attractive - the asymmetric risk aversion that protects the firm in normal times costs it in recoveries. He frames this as a tolerable error: the asymmetry between the permanent loss from overreach and the temporary opportunity cost from caution is so large that the firm would rather err on the side of caution. What he refuses to tolerate is the mistake of changing one's standards to fit the market. Lowering the discount to value required for entry, reaching for yield in late cycles, or buying lower-quality assets because high-quality ones are scarce are all process errors that look rational in the moment and reveal themselves only when the cycle turns. The firm's risk system is therefore designed less to predict drawdowns than to detect, in real time, when its own underwriting standards are drifting.