Global Macro: A Deep Econometric Analysis and Theoretical Explanation of Key Square Capital Management's Performance Record and Fund Industry Evaluation Standards

The Bulwark. (2025, Sep 9). Scott Bessent’s Hedge Fund Lost BILLIONS (w/ Marshall Brandt)[Video]. YouTube. https://www.youtube.com/watch?v=ZO0uvjmRRA0

Author’s Note:

This article presents a detailed analysis and interpretation of the views expressed in a discussion published by The Bulwark on YouTube (hosted by Tim Miller with guest Marshall Brandt). The analysis herein is based solely on the publicly available content of that video and is intended for academic, research, and commentary purposes. All interpretations of the speakers’ statements—and any related quantitative or theoretical extensions—are the author’s own and do not necessarily reflect the views of The Bulwark, its hosts, guests, or YouTube. Readers are encouraged to watch the original video to form their own conclusions.

This paper aims to conduct a rigorous academic dissection of the public discussion between Marshall Brandt and Tim Miller regarding Scott Bessent and the performance of the hedge fund he led, Key Square Capital Management (hereinafter "Key Square"). The discussion (cited from a YouTube video and related transcripts) raises a series of sharp criticisms regarding Bessent's investment record, touching not only on his specific returns but also on the complexity of evaluating Global Macro strategies, the dynamic mechanisms of capital flows, and the core issue of investment philosophy consistency. This paper will analyze the arguments in the discussion point by point, utilizing advanced financial theory, econometric models, behavioral finance perspectives, and verifiable public data on Key Square (strictly based on 13F filing data and limited media-disclosed performance), to conduct an in-depth critical assessment and expansion of these views.

Part One: The Epistemology of Performance Evaluation: Deconstructing the Definition and Measurement of "Failure"

The hedge fund industry, particularly the Global Macro specialization, has long been viewed as the apex of financial elitism. Legendary figures in the field, such as George Soros and Stanley Druckenmiller, created astonishing wealth through high-leverage, directional bets on global economic trends and policy inflection points, thereby shaping the image of the "Macro Master" who transcends the market. However, academic research in financial economics generally views with skepticism the ability of Active Management to consistently generate excess risk-adjusted returns (Alpha). The Efficient Market Hypothesis (EMH) poses a theoretical challenge, suggesting that consistently outperforming market benchmarks is systematically impossible in an informationally efficient market.

Even if we relax the assumptions of the EMH and acknowledge the existence of market frictions and irrationalities, the rational expectations equilibrium model by Berk and Green (2004) profoundly points out that manager Skill is a scarce and finite resource. As Assets Under Management (AUM) increase, Diseconomies of Scale—due to Market Impact Costs and constraints on the investment opportunity set—will inevitably erode Alpha. Investor capital flows toward high-skill managers until the expected Alpha, net of fees, approaches zero.

Within this theoretical context, the performance record of Scott Bessent and his fund, Key Square Capital, provides an excellent case study for examining the gap between the ideals and realities of the hedge fund industry. In the video, Marshall Brandt bluntly labels Bessent a "failed hedge fund manager," citing core arguments including a 50% frequency of down years and an AUM shrinkage of nearly 90%. While these metrics are intuitive and impactful, they appear overly coarse and simplistic within the framework of advanced financial analysis. Evaluating a hedge fund manager requires a more complex and sophisticated toolkit, encompassing Higher Moments of Risk, Non-linear Exposures, Liquidity Constraints, Incentive Distortions, and structural changes in the macroeconomic environment across multiple dimensions.

This paper will deconstruct each viewpoint in the video concerning Bessent's performance and industry perspectives, employing rigorous financial theory, quantitative models, and empirical evidence to deeply explore the underlying logic, potential biases, and broader industry implications. Our aim is not simply to confirm or refute Brandt's conclusions, but to demonstrate how, at a doctoral level of financial research, we systematically deconstruct, quantify, and critically evaluate a fund manager's performance and its impact on the financial system.

The discussion begins with Marshall Brandt's intuitive and sharp assessment of Bessent's performance record (00:00 - 00:06): "So 50 percent of his years were down. And I mean, I don’t see any way to view him as anything but a failed hedge fund manager." This assertion is repeatedly emphasized in the video (01:35-01:49) and forms the core argument for evaluating Bessent's career. On the surface, this metric has strong intuitive appeal, but using it as a sufficient condition for evaluating a Global Macro strategy manager is highly controversial in financial academia.

1.1 "50% Down Years": The Theoretical Flaws of the Batting Average Metric and the Importance of Higher Moments

The Batting Average (or Hit Rate), defined as the proportion of observations realizing positive returns within a specific time series ($P(R_t > 0)$), completely ignores the Magnitude and the shape of the return distribution. The core of financial markets lies in the trade-off between risk and return, not merely the pursuit of high-frequency wins. A fund manager might have an extremely high batting average, but if the average gain of profitable trades is far less than the average loss of losing trades (i.e., a negatively skewed strategy), the long-term expected return may be negative.

We can illustrate this with a simple expected return model. Assume an investment strategy has a win probability $P_{win}$, an average win magnitude $R_{win}$, a loss probability $P_{loss} = 1 - P_{win}$, and an average loss magnitude $R_{loss}$ (typically a negative value). The expected return $E[R]$ of the strategy is:

$E[R] = (P_{win} \times R_{win}) + (P_{loss} \times R_{loss})$

Brandt's argument implies that if $P_{win} \leq 0.5$ (based on annual data), the manager is a failure. However, this only necessarily holds true if the magnitudes of gains and losses are roughly symmetrical, or if $R_{win} < |R_{loss}|$. If a strategy can achieve a significant $R_{win} \gg |R_{loss}|$ (i.e., High Payoff), then even if $P_{win} < 0.5$, $E[R]$ can be positive and potentially very attractive on a risk-adjusted basis.

In the Global Macro domain, this low-hit-rate, high-payoff return structure is particularly prevalent. Macro traders often construct positions with Convexity, aiming to profit from significant macroeconomic imbalances, policy shifts, or Tail Events. These strategies resemble purchasing Out-of-the-Money Options, whether through explicit derivatives or synthetic positions constructed via dynamic hedging.

The characteristic of these strategies is that under normal market conditions, they may continuously pay "time value" or "premiums" (manifesting as frequent small losses or returns below the benchmark), but they can generate enormous non-linear returns when the market experiences severe volatility or major inflection points. Such strategies seek a positively skewed return distribution. Skewness, as the third moment, measures the asymmetry of the distribution relative to the mean.

$S = E\left[\left(\frac{R - \mu}{\sigma}\right)^3\right]$

Where $R$ is the return, $\mu$ is the mean, and $\sigma$ is the standard deviation. Positive skewness ($S>0$) means the distribution has a longer Right Tail, i.e., a higher probability of extreme positive returns.

Many successful systematic macro strategies, such as Trend Following (also known as CTA strategies), have been empirically proven to have significant positive skewness or "Crisis Alpha" (Hamill, Rattray, & Van Hemert, 2016). For such strategies, a batting average below 50% is an inherent feature of their design, not a sign of failure. They achieve long-term profitability through strict stop-losses (controlling $R_{loss}$) and letting profits run (maximizing $R_{win}$).

For discretionary Global Macro managers (like Bessent), they also frequently engage in similar asymmetric bets. Therefore, concluding Bessent's failure solely based on the 50% frequency of down years might be a fundamental misunderstanding of the return characteristics of Global Macro strategies. To conduct an effective evaluation, one must analyze the higher moments of the return distribution. We also need to consider Coskewness, the covariance between asset returns and the square of market returns, which measures the asset's performance during extreme market volatility. If Bessent's fund provided a hedge during market downturns (i.e., negative coskewness), it would have high portfolio value even if its average returns were not high.

Conversely, if Bessent's strategy exhibited significant Negative Skewness ($S<0$)—frequent small gains and occasional massive losses (similar to "picking up nickels in front of a steamroller" or volatility selling strategies)—then even if the batting average were above 50%, it might conceal massive tail risk. In this case, Brandt's criticism (though based on the wrong metric) might have arrived at the correct conclusion through flawed reasoning. While negatively skewed strategies may provide a stable income stream over long periods, they often face the risk of catastrophic collapse (e.g., LTCM or certain fixed-income arbitrage funds).

If Bessent's 50% down years were accompanied by massive Drawdowns, this strongly suggests his strategy might suffer from negative skewness or uncontrolled risk management. Global Macro strategies, due to their high leverage, indeed carry significant drawdown risk. If Bessent, in pursuit of high convexity, failed to effectively manage the Negative Carry or leverage of his positions, then frequent losses combined with deep drawdowns constitute evidence of performance failure.

1.2 Empirical Testing of Key Square's Performance: The Search for Missing Convexity

However, theoretical possibilities require empirical data support. We must examine Key Square's actual performance data to determine if it achieved the expected positive skewness. Although hedge fund Net Asset Value (NAV) data is typically private, we can rely on scattered information from public reports (Reference Material "Table 3"). Investor materials cited by Reuters show:

• 2016: +13%

• 2017: -7%

• 2018-2021: "lost money or were roughly flat"

• 2023, 2024 (as of November): Double-digit positive returns

In this record, we observe a prolonged period of stagnation (2017-2021). Even under the most optimistic assumption that "roughly flat" means an annualized return of 0%, the cumulative return from 2016 to 2021 (six years) is approximately:

$R_{2016-2021} = (1+0.13) \times (1-0.07) \times (1)^4 - 1 \approx 5.11\%$

This implies an annualized return of only about 0.83%. This level is far below the hedge fund industry average and potentially below the risk-free rate during the same period (especially after deducting the standard hedge fund fees of 2% management fee and 20% performance fee).

This empirical result indicates that the magnitude of Key Square's gains (e.g., +13% in 2016) was insufficient to compensate for its losses and long-term stagnation. The fund does not appear to have exhibited the positive skewness or Convexity sought by Global Macro strategies; instead, it failed to generate significant absolute returns in the long-term market environment. Therefore, although the metric used by Brandt has theoretical flaws, when combined with Key Square's actual data, his conclusion regarding Bessent's "failure" is empirically valid. Key Square failed to realize the core value proposition of a Global Macro strategy.

1.3 Modern Asset Pricing Theory and the Stochastic Discount Factor (SDF) Framework

Beyond the traditional mean-variance framework, the core of modern asset pricing theory is the Stochastic Discount Factor (SDF). In a no-arbitrage market, there exists a stochastic discount factor $M_{t+1}$ such that the price of any asset $i$ satisfies: $P_{i,t} = E_t[M_{t+1} X_{i,t+1}]$. The SDF represents the marginal utility of investors in different economic states.

The value of a truly valuable investment strategy depends on the covariance of its returns with the SDF. If a strategy provides high returns when the investor's marginal utility is high (i.e., poor economic conditions, market downturns), then the strategy has extremely high value (hedging value). Global Macro funds often claim to provide "Crisis Alpha."

Merely observing "50% of the time losing money" is insufficient. We need to analyze the timing of these losses. If Key Square's losses were concentrated during market booms, and profits concentrated during crises, then even with a high frequency of losses, the fund might have provided valuable macro hedging. Cochrane (2005) emphasizes that the evaluation of hedge funds must be conducted under the SDF framework.

However, during Key Square's prolonged slump from 2018-2021, the market experienced significant volatility (e.g., the Q4 2018 market sell-off or the March 2020 pandemic shock). If Key Square failed to provide this protection during these critical periods, instead falling or remaining flat along with other risk assets, then its contribution to an investor's portfolio is negligible. Key Square's performance record seems to indicate that it failed to effectively provide this macro hedging value, further supporting the assertion of "failure."

1.4 Beyond the Sharpe Ratio: Tail Risk and Downside Risk Measures

Given the non-normal characteristics of Global Macro strategies, the traditional Sharpe Ratio ($SR = (E[R] - R_f) / \sigma$) can be misleading. We need to introduce measures that are more sensitive to downside risk.

Sortino Ratio: Considers only the volatility below a target return (downside deviation, $\sigma_d$): $Sortino = (E[R] - T) / \sigma_d$. If Bessent's fund had large and frequent losses, its Sortino Ratio might be very low.

Omega Ratio: Considers the entire return distribution, measuring the ratio of probability-weighted gains above a threshold $r$ to the probability-weighted losses below that threshold: $\Omega(r) = \int_r^\infty (1-F(x))dx / \int_{-\infty}^r F(x)dx$. If Key Square's Omega Ratio (around the zero threshold) was consistently below 1, it indicates that the magnitude of potential losses exceeded potential gains.

Maximum Drawdown (MDD): Measures the largest peak-to-trough decline. A fund that experiences a deep drawdown requires higher subsequent returns to recover to the High Water Mark. For example, a 50% loss requires a 100% subsequent gain to break even. Key Square's multi-year period of "flat or losses" suggests the fund likely suffered significant drawdowns and recovered slowly, severely impairing the compounding growth of capital.

1.5 Behavioral Finance Perspective: Loss Aversion and the "Never Lose Money" Rule

Brandt mentioned a crucial industry maxim in the discussion (01:36-01:44): "And the golden rule from Soros and Druckenmiller, his mentors, was you never lose money, right? ... I mean at least you can put it in 10-year Treasuries."

Brandt's citation of Soros and Druckenmiller's "never lose money" rule, and the suggestion to at least invest in 10-year Treasuries, emphasizes the importance of Opportunity Cost and benchmark selection.

As a philosophy, "never lose money" centers on the extreme importance of risk management and the devastating impact of large drawdowns on compounding effects. The impact of large drawdowns on the Geometric Mean Return is much greater than on the Arithmetic Mean Return. The approximate relationship is:

$R_{geo} \approx R_{arithmetic} - \frac{1}{2}\sigma^2$

This means high volatility and deep drawdowns rapidly erode long-term capital accumulation.

However, interpreting "never lose money" as necessarily surpassing a specific risk-free asset benchmark requires caution. The objective of hedge funds is typically to generate Absolute Return, i.e., striving for positive returns in any market environment, with low correlation to traditional asset classes (like stocks and bonds). Whether the 10-year Treasury is an appropriate benchmark is debatable. 10-year Treasuries carry significant Duration Risk and inflation risk. In a rising interest rate cycle (such as 2022), 10-year Treasuries themselves can suffer significant nominal losses.

A more appropriate benchmark would be the cash rate (e.g., 3-month T-bill rate), plus a reasonable risk premium. If Bessent's fund, after deducting high fees (typically "2 and 20"), failed to significantly outperform the cash rate over the long term, then regardless of its batting average, it can be considered to have failed to achieve its core objective.

Furthermore, in the strictest sense, we can use Stochastic Dominance to compare Bessent's fund with the benchmark. If the benchmark exhibits Second-Order Stochastic Dominance (SSD) relative to Bessent's fund, it means that for any risk-averse investor, the benchmark is superior to the fund. Given the high fees of hedge funds, achieving stochastic dominance over a low-cost benchmark is extremely challenging.

This perspective is deeply rooted in the theory of Loss Aversion in behavioral finance. Kahneman and Tversky's (1979) Prospect Theory posits that the psychological pain of losses is far greater than the pleasure derived from equivalent gains. The utility function is steeper in the loss domain ($U(x) = -λ(-x)^β$ for $x < 0$, where $λ > 1$ represents the loss aversion coefficient).

Therefore, frequent losses (even if small in magnitude) cause severe erosion of investors' psychological capital, leading to redemption pressure. Hedge fund investors pay high fees expecting Absolute Return and capital preservation. Key Square's prolonged slump clearly violated this psychological contract. The "golden rule" of Soros and Druckenmiller is not literally about never losing money, but emphasizes the extreme importance of risk management—cutting losses quickly when judgments are wrong, avoiding permanent impairment of capital. Bessent's frequent losses indicate he may have failed to effectively implement this principle. From this perspective, Brandt's criticism captures one of the key elements of success in the hedge fund industry.

Part Two: Analyzing the 90% AUM Collapse: The Dynamic Feedback of Performance, Capital Flows, and Diseconomies of Scale

The second core argument raised by Brandt is the catastrophic decline in Key Square's assets under management (00:07-00:13; 01:50-01:56): "So he started with 4.5 billion of assets under management, he shut down the fund with 577 million. So he lost 90% of his capital." This statement, coupled with the chart showing AUM decline (from Convergence Inc.), has a strong visual impact and is central to evaluating the Key Square case.

2.1 The AUM Dynamics Equation and the Ambiguity of Brandt's Assertion

The change in assets under management over time can be described by the following equation:

$AUM_t = AUM_{t-1} \times (1 + R_t) + Flows_t$

Where $R_t$ is the investment return rate for the period, and $Flows_t$ is the net capital flow for the period. A decline in AUM can be caused by investment losses ($R_t < 0$) and net capital outflows ($Flows_t < 0$).

Brandt's statement "he lost 90% of his capital" is seriously ambiguous. In financial terminology, this usually refers to investment losses causing the NAV to drop by 90%. However, as analyzed in Part One, Key Square's cumulative investment return, though mediocre (annualized around 0.83% to 1.35%), was far from the -90% level. Therefore, Brandt logically conflated the change in AUM with investment performance. The decline in AUM from $4.5 billion to $577 million (a decrease of about 87.2%) must have been predominantly caused by massive investor redemptions.

However, this does not mean Brandt's criticism is baseless. There is strong Endogeneity between investment losses and capital outflows. As Brandt pointed out (04:22-04:27): "And as a hedge fund manager, typically, only people will draw money when you're not doing well."

2.2 Berk and Green's Rational Market Model: Skill, Scale, and Capital Flows

The key theoretical framework for understanding AUM dynamics is Berk and Green's (2004) rational fund flow model. This model provides profound insight into the equilibrium of the active management market. The model posits that investment skill is scarce, but fund managers face Diseconomies of Scale when deploying their skills. As the fund size increases, the marginal cost of finding and exploiting mispricing opportunities increases (e.g., market impact costs, liquidity constraints), leading to the gradual erosion of Alpha.

In a rational market, investors will continuously allocate capital to funds expected to generate positive Alpha until the fund size reaches an equilibrium point where the expected Alpha is zero (net of fees). If the fund performs poorly, investors update their beliefs about the manager's skill (Bayesian Updating) and begin to redeem capital.

In Key Square's case, based on Bessent's reputation at Soros Fund Management (SFM), investors initially expected him to possess high skill, thus investing a massive $4.5 billion. However, the persistent poor performance (2017-2021) sent a strong negative signal to investors, indicating that Bessent's skill might be lower than expected, or his strategy Capacity was far less than $4.5 billion.

The process of AUM declining by 90% is essentially the market rationally repricing Bessent's investment skill. Capital withdraws from inefficient allocations, seeking a new equilibrium scale. If the AUM eventually stabilized at $577 million, this might reflect the market's belief that at this scale, the expected Alpha Bessent can generate is close to zero.

2.3 The Curse of Initial Scale and Capacity Constraints

Key Square's launch size of $4.5 billion was enormous. Brandt expressed sympathy (01:28 - 01:35): "I don’t know how exactly, you know, you put 4.5 billion dollars to work on a Thursday." This touches upon the capacity constraint issue in Global Macro strategies.

If Bessent's strategy capacity was far below $4.5 billion, the excessive scale doomed its mediocrity from the start. Large-scale capital deployment generates significant market impact costs. Market impact costs can be approximated as: $Cost \approx \sigma \times \sqrt{V/ADV}$, where $\sigma$ is volatility, $V$ is the transaction volume, and $ADV$ is the average daily volume. For a $4.5 billion fund, even allocating a small fraction of capital to a position results in a transaction size large enough to cause significant market impact, especially in liquidity-constrained markets (like the Thai baht market mentioned by Brandt).

Furthermore, the massive scale might force the manager to invest in suboptimal opportunities or over-diversify, thereby diluting Alpha. Key Square's AUM collapse can be viewed as the market correcting its initial capital allocation error (allocating too much capital to a manager with limited capacity).

2.4 The Convex Flow-Performance Relationship and Liquidity Spirals

Brandt's view that investors withdraw capital when performance is poor is supported by empirical research, but this relationship exhibits specific non-linearity and asymmetry in the hedge fund domain.

Financial academic research shows a non-linear, Convex Flow-Performance Relationship (Chevalier and Ellison, 1997; Goetzmann, Ingersoll, and Ross, 2003). Investors tend to chase past winners, but the punishment for losers is swifter and harsher.

However, the flow-performance sensitivity of hedge funds is higher and exhibits stronger asymmetry. Fung and Hsieh (2000) found that hedge fund investors react more quickly and intensely to negative performance. This stems from the structure of hedge fund investors (typically institutional investors and high-net-worth individuals who are more risk-sensitive and have more options) and the contractual terms of hedge funds (such as Lock-up periods and Notice periods). When the lock-up period ends, accumulated negative performance triggers concentrated redemption pressure.

Key Square experienced a typical negative feedback loop:

1. Poor performance (negative Alpha or bad luck) leads to a decline in NAV.

2. Investors lose confidence and begin massive redemptions.

3. To meet redemption requests, the fund is forced to liquidate positions under adverse market conditions (Forced Selling).

4. Forced selling generates market impact costs, further deteriorating performance. This phenomenon is detailed in the Liquidity Spiral model by Brunnermeier and Pedersen (2009).

5. The decline in AUM leads to reduced management fee income, affecting the fund's operational capacity and talent retention, further weakening Alpha generation capability.

2.4.1 High-Water Marks (HWM) and the Dynamic Game of Redemption Incentives

Goetzmann, Ingersoll, and Ross (2003) proposed a key model explaining why hedge fund investors accelerate redemptions when facing losses. Due to the existence of High-Water Mark (HWM) provisions (managers can only extract performance fees when the NAV exceeds the historical peak), investors face a complex decision when the fund's NAV declines.

When $NAV_t < HWM$, the manager's incentive fee is zero. From the investor's perspective, they receive "free" management services for a future period without paying performance fees, which is equivalent to the value of a call option. However, the value of this option depends on the probability of the fund recovering above the HWM. If investors believe the manager's skill has permanently deteriorated (Skill Deterioration), or the probability of recovery is extremely low due to diseconomies of scale (see Part Five), the value of this "free option" decreases.

In this situation, investors make an intertemporal comparison: stay in the current fund, expecting recovery (and enjoying free management), or redeem capital and switch to another fund with higher expected Alpha (even if fees must be paid immediately). When there are numerous alternatives in the market, investors are more inclined toward the latter.

In Bessent's case, Key Square Capital launched with a massive $4.5 billion scale, and investor expectations were extremely high. When initial performance showed losses, the fund quickly fell below the high-water mark. As losses deepened, the difficulty of recovering to the HWM increased exponentially. For example, a 50% loss requires a 100% future gain to recover. This greatly weakened the investors' willingness to hold, triggering massive redemptions. The collapse of investor confidence in Bessent's skill was the core driver of the redemption wave.

2.4.2 Liquidity Mismatch, Forced Selling, and Spiral Effects

When a fund faces massive redemption pressure, liquidity management becomes critical for survival. If the fund's asset liquidity is lower than its liability liquidity (i.e., investor redemption terms), a Liquidity Mismatch occurs. To meet redemption requests, the manager may be forced to conduct Fire Sales of assets under adverse market conditions.

Shleifer and Vishny's (1997) theory suggests that in the presence of market frictions and limited arbitrage, fire sales further depress asset prices, leading to a decline in the fund's NAV, thereby triggering more redemptions and forming a vicious cycle. Brunnermeier and Pedersen's (2009) model further explores the interaction between Market Liquidity and Funding Liquidity, demonstrating how liquidity spirals can lead to systemic risk.

$\Delta P_t = -\lambda \cdot Flows_t$

Where $\Delta P_t$ is the price change, $Flows_t$ is the capital flow (negative value indicates outflow), and $\lambda$ is the market liquidity parameter (measuring price sensitivity to flows). When market liquidity dries up ($\lambda$ increases), even small redemptions can lead to significant price drops.

For a fund with an initial scale of $4.5 billion, even when investing in relatively liquid Global Macro markets, large-scale concentrated liquidations can generate significant market impact costs. We can measure this risk using Liquidity-Adjusted VaR (L-VaR), which considers the potential market impact costs of liquidating positions.

$L\text{-}VaR = VaR + Exogenous\ Liquidity\ Cost$

The sharp decline in Key Square's AUM was likely accompanied by significant liquidity pressure and forced selling effects, which further amplified its actual investment losses. The fund might have been forced to prioritize liquidating the most liquid assets, leading to the deterioration of the remaining portfolio's liquidity condition (Liquidity Profile Deterioration), further exacerbating the risk.

2.4.3 Incentive Distortions and Risk-Seeking Behavior

When the fund's NAV is far below the high-water mark, the manager's compensation structure resembles a Deep Out-of-the-Money Call Option. In this situation, the manager has a strong incentive to take excessive risks, i.e., adopting a "Gambling for Resurrection" strategy.

Carpenter's (2000) model demonstrates how this convex compensation structure induces risk-seeking behavior. When the sensitivity (Delta) of the manager's expected future compensation $V_{manager}$ to the fund's NAV $NAV_t$ is low, but the sensitivity to volatility (Vega) is high, increasing the portfolio's volatility can increase the manager's expected compensation, even if it increases the investors' risk.

$\frac{\partial V_{manager}}{\partial \sigma} > 0 \quad \text{when} \quad NAV_t \ll HWM$

In Bessent's case, as AUM and NAV continued to decline, he likely faced enormous pressure and distorted incentives. To salvage the fund and his career, he might have been compelled to undertake aggressive trades contrary to his professed risk management principles (such as Soros's "never lose money"). While this behavior might theoretically bring a turnaround, empirically it is more likely to accelerate the fund's collapse. The 90% decline in AUM suggests that this potential Moral Hazard ultimately led to catastrophic consequences. Such behavior is a breach of Fiduciary Duty to investors.

2.4.4 Investor Heterogeneity and "Run" Dynamics

The investor base of a hedge fund is usually heterogeneous, with different information channels, risk preferences, and redemption terms. When the fund performs poorly, informed investors (such as seed investors, strategic partners, or insiders) may redeem first. This behavior not only directly reduces AUM but, more importantly, sends a strong negative signal to other investors, potentially triggering a Coordination Game similar to a Bank Run. Even if the fund's fundamentals have not deteriorated to the point of requiring liquidation, the collective behavior of investors (even if rational) can lead to the fund's collapse.

Key Square's investor base (considering its large initial size) likely included large institutional investors (such as pension funds, sovereign wealth funds) and Fund of Funds. These institutions typically have strict risk controls and stop-loss provisions (e.g., trigger conditions based on VaR or MDD). Once triggered, these lead to mandatory, inelastic redemptions, accelerating the decline in AUM. The Herding Behavior of institutional investors may also have exacerbated the redemption wave.

In summary, the 90% shrinkage of Key Square Capital's AUM is undeniable evidence of its operational failure. However, this is not merely the result of investment decision errors but is driven by a combination of complex structural factors (high-water marks, liquidity mismatch), behavioral factors (incentive distortions, investor runs), and dynamic feedback mechanisms (liquidity spirals). Bessent's case vividly demonstrates how these theoretical mechanisms can lead to the rapid decline of a high-profile hedge fund in reality.

2.5 Empirical Analysis of Key Square's 13F AUM Data

We can further verify the downward trend of AUM by analyzing Key Square's 13F filing data (Reference Material "Table 1"). It must be emphasized that 13F AUM only includes long U.S. equity positions and is much smaller than the total AUM of a Global Macro fund. The 13F AUM at the end of 2016 ($806 million) being far below the total AUM ($4.5 billion) confirms this.

Key Square's 13F AUM history shows dramatic fluctuations and an overall downward trend:

• Peak: Reached $1.0408 billion in December 2017.

• Sharp Decline: Dropped to $306.8 million by the end of 2018 (a 70.6% decline within a year). This rapid decline indicates that after performance began to deteriorate (-7% in 2017), investors quickly withdrew or the fund significantly reduced its U.S. equity exposure.

• Continued Shrinkage: By September 2022, 13F AUM dropped to $56.13 million.

• Latest Data: In June 2024, only $14.19 million.

From the peak to the latest data, the 13F AUM declined by 98.6%. This trend is consistent with the decline in total AUM, confirming the catastrophic contraction of Key Square's asset scale.

2.6 Case Comparison: Tiger Global Management's AUM Volatility and Resilience

To understand Key Square's failure more comprehensively, it is beneficial to conduct a comparative analysis with other high-profile hedge funds that have experienced similar capital dynamics. The case of Tiger Global Management (hereinafter "Tiger Global") (Reference Material provided) offers an excellent reference point.

Tiger Global's total AUM peaked at about $86 billion at the end of 2021. However, as high-valuation tech stocks were hit hard in 2022, its flagship hedge fund plunged 56%. This led to its total AUM dropping to $46 billion by the end of 2023 (a decline of about 46%).

Both cases illustrate the vulnerability of concentrated betting strategies during macroeconomic reversals. However, Key Square's AUM decline (nearly 90%) was far greater than Tiger Global's, and ultimately led to its transformation into a family office. This reflects several key differences:

• Long-Term Track Record and Reputation Capital: Tiger Global had an excellent 20-year annualized record of 21%, accumulating deep reputation capital, making investors more patient. Key Square lacked a long-term independent track record of success.

• Resilience and Recovery: Tiger Global achieved a strong rebound in 2023 (+28.5%) and 2024 (approx. +24%). Although Key Square also had positive returns during the same period, its AUM had already shrunk significantly, and its recovery seemed to rely more on high-risk "Gambling for Resurrection" (see Part Six).

The comparison with Tiger Global highlights Key Square's failure in building long-term trust and coping with adversity. The collapse of AUM is the market's final verdict on Bessent's investment skill.

Part Three: The Epistemology of Global Macro: Trading Difficulty, Macro Forecasting, and Reflexivity

In the discussion, Brandt described the nature of Global Macro trading, emphasizing its complexity and challenges (00:41-01:11). He pointed out that "trading global macro is very difficult" because it involves making "really big bets" on a wide range of asset classes such as currencies, commodities, interest rates, and inflation. Understanding the inherent challenges of Global Macro strategies is crucial for objectively evaluating Bessent's performance.

3.1 The Efficiency of Macro Markets and the Limits of Forecasting

The difficulty of Global Macro strategies first stems from the Macro Efficiency of the markets they trade. Compared to the micro-level individual stock market, major macro markets (such as G10 currencies, government bond markets) are generally considered more informationally efficient. These markets are deep, liquid, and closely monitored by a large number of professional investors and researched by macroeconomists globally.

3.1.1 The Exchange Rate Puzzle and Random Walks

In the foreign exchange market, the manifestation of the Efficient Market Hypothesis is Uncovered Interest Rate Parity (UIP).

$E_t[\Delta s_{t+1}] = i_{d,t} - i_{f,t}$

Where $\Delta s_{t+1}$ is the rate of change in the exchange rate, and $i_d$ and $i_f$ are the domestic and foreign risk-free interest rates, respectively. Although empirical research has found the "Forward Premium Puzzle," where high-interest-rate currencies tend to appreciate rather than depreciate (Fama, 1984), this does not mean predicting exchange rate movements is easy. The classic study by Meese and Rogoff (1983) showed that structural models based on macroeconomic fundamentals perform worse in predicting exchange rates than a simple random walk model. This highlights the extreme difficulty of forecasting foreign exchange markets, as exchange rates are affected by too many unpredictable shocks, and macroeconomic data is typically low-frequency and lagging.

3.1.2 The Dynamic Complexity of the Yield Curve and No-Arbitrage Models

In the interest rate market, predicting the future movements of the yield curve (level, slope, curvature) requires accurate forecasts of central bank policy, inflation dynamics, and economic growth prospects. Modern interest rate theory (such as Arbitrage-Free Dynamic Term Structure Models, DTSMs) attempts to characterize the dynamics of the yield curve by imposing no-arbitrage conditions.

$P(t, T) = E_t^{\mathbb{Q}}\left[ \exp\left(-\int_t^T r(s) ds\right) \right]$

Where $P(t, T)$ is the zero-coupon bond price, $r(s)$ is the instantaneous short-term interest rate, and $\mathbb{Q}$ is the risk-neutral measure. Although these models can capture most of the movements in the yield curve, predicting changes in the Market Price of Risk remains extremely challenging. Central bank forward guidance and unconventional monetary policies (such as QE) further distorted the term premium, making predictions based on historical patterns more complex.

Global Macro traders attempt to profit from temporary imbalances or inefficiencies in these markets. But this requires them to possess superior information processing capabilities or more advanced analytical models than the market consensus. In today's era of rapid information dissemination, such advantages are increasingly difficult to obtain and maintain.

3.2 The Challenges of Macroeconomic Forecasting and the Lucas Critique

The core of Global Macro strategies lies in forecasting macroeconomic variables. However, the accuracy of macroeconomic forecasting has always been questioned. The macroeconomic system is a complex, non-linear dynamic system affected by numerous interacting variables and random shocks.

The Efficient Market Hypothesis (EMH) suggests that if the market is efficient, it is impossible to consistently earn excess returns by forecasting macroeconomic variables. Even if we relax the assumptions of the EMH, macro forecasting remains extremely difficult.

The Lucas Critique (Lucas, 1976) points out that the historical relationships between macroeconomic variables may change due to shifts in policy regimes. When policymakers attempt to exploit these relationships, the expectations and behaviors of economic agents also adjust accordingly, causing the original models to fail. $Y_t = \beta(Z_t) X_t + \epsilon_t$, if the parameter $\beta$ depends on the policy variable $Z_t$, then models estimated based on historical data will fail when policies change.

This means Global Macro traders cannot rely solely on historical data and models; they must have a deep understanding of the current political-economic environment, the reaction functions of policymakers, and the formation of market participants' expectations. For example, the difficulty of forecasting exchange rates is well known. The classic study by Meese and Rogoff (1983) found that models of exchange rates based on macroeconomic fundamentals perform worse in predicting short-term exchange rate movements than a simple random walk model ($E[S_{t+1}] = S_t$). Research by Goyal and Welch (2008) also shows that most macroeconomic variables have very poor ability to predict asset returns out-of-sample.

3.3 Reflexivity Theory and Global Macro Trading

The Theory of Reflexivity, proposed by Bessent's mentor George Soros, provides a crucial framework for understanding the complexity of macro markets. Reflexivity posits a two-way feedback loop between the Cognitive Function of market participants and the reality they attempt to influence (Manipulative Function). Investors' expectations affect asset prices, and changes in prices, in turn, affect economic fundamentals (e.g., through credit channels, wealth effects, capital flows) and investors' expectations.

$Perceptions \xrightarrow{Cognitive} Prices \xrightarrow{Manipulative} Fundamentals \xrightarrow{Feedback} Perceptions$

This feedback loop can cause the market to deviate from equilibrium, generating self-reinforcing bubbles and crashes (i.e., non-equilibrium dynamics). Global Macro traders attempt to identify and exploit these reflexive processes. However, reflexivity also makes market dynamics highly non-linear, Path Dependent, and prone to Regime Shifts. Predicting the Inflection Points of reflexive processes is extremely difficult and can lead to massive losses. Whether Bessent mastered the ability to identify and navigate reflexivity, or merely became a victim of reflexive processes, is key to evaluating his failure.

3.4 The Risk Characteristics of "Big Bets": Leverage, Concentration, and Asymmetric Risk Management

Brandt describes Global Macro as making "really big bets." This reveals the key risk characteristics of the strategy: high leverage and high concentration, and the extremely high demands on risk management capabilities.

3.4.1 The Amplifying Effect and Vulnerability of Leverage

Since the volatility of macro variables is typically lower than that of individual stocks, Global Macro funds commonly use high leverage (through futures, options, or repurchase agreements) to achieve significant returns. Leverage amplifies gains, but it also proportionally amplifies risks and sensitivity to errors. High leverage makes the fund extremely sensitive to market shocks and funding liquidity tightening. When liquidity dries up, highly leveraged funds may be unable to meet Margin Calls, forced to liquidate, triggering systemic risk (such as the LTCM crisis). The massive losses of Bessent's fund are likely related to the use of high leverage.

3.4.2 Concentration and the Test of Conviction

Global Macro strategies are typically based on High Conviction, concentrated trades, rather than broad diversification. Concentration increases Idiosyncratic Risk. If the core macro judgment is wrong, the fund will suffer heavy losses. Successful macro traders need to maintain extremely high cognitive flexibility and strict risk controls (such as stop-losses, position sizing management) while making high-conviction bets. The Kelly Criterion provides a theoretical framework for optimal position sizing:

$f^* = \frac{p(b+1)-1}{b}$

Where $f^*$ is the optimal betting proportion, $p$ is the probability of winning, and $b$ is the payoff odds. However, under fundamental uncertainty, accurately estimating $p$ and $b$ is extremely difficult, and overconfident application of the Kelly Criterion can lead to over-betting.

Soros's success lies in his ability to quickly change his views and liquidate positions when he discovers an error. This requires extreme psychological resilience and discipline. If Bessent lacked this ability, exhibiting Stubbornness or the Endowment Effect, his concentrated "big bets" could turn into one-way suicide attacks.

3.4.3 Behavioral Biases and the Pitfalls of Discretionary Macro

Bessent appears to employ a Discretionary macro strategy, relying on the manager's analysis, intuition, and judgment. While this approach offers flexibility, it is more susceptible to a range of Behavioral Biases:

• Overconfidence: Overestimating the ability to predict macro trends, leading to excessive risk-taking and overtrading.

• Confirmation Bias: Selectively seeking information that supports one's views and ignoring contrary evidence, especially when the macro narrative is complex.

• Disposition Effect: Taking profits too early (fear of losing paper gains) and cutting losses too late (unwillingness to admit failure). This runs counter to the positively skewed return structure required by Global Macro strategies (cut losses, let profits run).

• Anchoring: Relying excessively on historical data or initial judgments, failing to adjust views promptly based on new information.

In complex macro environments, especially under pressure (such as facing redemption pressure), these cognitive biases are amplified. Bessent's failure may be partly attributed to his inability to overcome these inherent behavioral pitfalls when facing market pressure and massive capital.

3.5 The Evolution of Global Macro Strategies and Adaptability Challenges

In the Post-GFC Era, Global Macro strategies faced unprecedented challenges. Central banks' unconventional monetary policies (QE, ZIRP) distorted market signals, suppressed macro volatility, and made traditional macro analysis more difficult. Central bank interventions changed the traditional relationships between macroeconomic variables, such as the flattening of the Phillips curve.

In the QE era, the drivers of asset prices might rely more on liquidity than traditional fundamentals. If Bessent failed to adapt to this change, his poor performance is understandable. Many traditional discretionary macro funds performed poorly during this period.

In recent years, with soaring inflation and the onset of aggressive rate hike cycles (post-2022), the macro environment has changed dramatically again. This provided new opportunities for Global Macro strategies. However, Key Square's AUM continued to decline significantly in 2022 (13F AUM dropped to $56 million in 2022Q3), indicating it may have failed to effectively capture this macro paradigm shift. For example, if the fund underestimated the persistence of inflation and the determination of central banks to raise rates, it could have faced huge losses in the interest rate and foreign exchange markets.

Research by Pastor, Stambaugh, and Taylor (2015) suggests that manager skill may be time-varying and correlated with the macroeconomic environment. A strategy successful under a specific market paradigm may fail after a paradigm shift. Bessent's failure may reflect his inability to adapt to the new macro environment. While Brandt acknowledges the difficulty of the field, this cannot excuse long-term poor performance. If the environment becomes too difficult to generate Alpha, the strategy should not attract so much capital.

Part Four: Pedigree, Reputation Capital, and the Curse of the "Soros Protégé"

Bessent's distinguished background is a significant part of his career. Brandt noted (01:12-01:17) that he worked under Soros and Druckenmiller. This experience accumulated massive reputation capital for him, enabling him to raise $4.5 billion when launching Key Square. However, Brandt also observed a common pattern (01:18-01:27): excellent PMs often underperform after spinning out. This phenomenon can be termed the "Mentor-Protégé Paradox."

4.1 Reputation Capital and Signaling Theory

In the asymmetric information market of asset management, investors rely on various signals to infer the manager's ability. Reputation capital is one of the most important signals (Spence, 1973). As the former Chief Investment Officer (CIO) of SFM, Bessent's reputation capital was undoubtedly top-tier. This reputation enabled him to overcome the information asymmetry barriers when starting a new fund and attract a large amount of initial capital. Research by Nanda, Wang, and Wysocki (2014) shows that managers with good reputations are more likely to obtain seed capital and larger launch sizes. Key Square's $4.5 billion launch size is a testament to the power of reputation capital.

4.2 Performance Persistence of Spin-off Funds: The Issue of Skill Transferability

However, the question is whether success on the original platform can be translated into success in independent operation. Academic research is skeptical about the performance persistence of spin-off funds (Dimmock and Gerry, 2012; Agarwal, Lu, and Ray, 2017). This phenomenon may have several reasons:

• Inseparability of Skills and Platform Dependence: A manager's success may depend not only on their individual investment skills but also on the support systems of their platform, including research teams, trading execution capabilities, risk management frameworks, and information networks. At SFM, Bessent had access to the wisdom of Soros and Druckenmiller and global information channels. At Key Square, he needed to build these systems from scratch. The loss of platform advantage might be a significant reason for his performance decline.

• Difficulty in Transferring Tacit Knowledge: The effectiveness of the "apprenticeship" system in transferring skills is particularly limited in the Global Macro field. The success of Soros and Druckenmiller relies heavily on their personal intuition and experience, which are Tacit Knowledge and difficult to systematically teach and replicate.

• Changes in Risk Appetite and Oversight Mechanisms: Within the parent fund, portfolio managers may be subject to stricter risk controls and oversight. When operating independently, they may face different incentive structures and fewer external constraints, potentially leading to changes in risk-taking behavior.

4.3 Role Transition and Agency Problems

Independent operation also brings the challenge of role transition. Shifting from a CIO focused on investment to a CEO who needs to manage the entire enterprise requires a different skill set. Many excellent traders are not good at managing organizations, handling investor relations, and dealing with operational challenges.

Furthermore, shifting from managing proprietary capital (or family office capital, like the later SFM) to managing Outside Investor Capital (OPM), the manager faces different agency problems (Harris and Raviv, 1979). They need to cope with investor redemption pressure and short-term performance evaluation, which may affect their long-term investment decisions. For example, to avoid redemptions, the manager might engage in Herding behavior or be forced to prematurely liquidate promising long-term positions.

4.4 Violation of the "Golden Rule": Failure of Risk Management

Brandt emphasized the importance of risk management and cited the golden rule of Bessent's mentors (01:36-01:39): "you never lose money." This emphasizes the extreme importance of capital preservation and risk management, stressing strict stop-loss discipline and dynamic risk exposure management.

Bessent's continuous losses at Key Square and the collapse of AUM indicate serious deficiencies in his risk management, failing to adhere to this principle. In Global Macro trading, due to the use of leverage and the concentration of bets, risk management is crucial. Soros and Druckenmiller are known for their ability to quickly change their views when discovering errors.

Risk management is not only technical but also psychological. It requires the manager to overcome Ego and Confirmation Bias. Among managers with distinguished backgrounds, ego might be particularly strong, leading to holding positions too long in adverse situations and failing to stop losses timely. While Bessent absorbed the bold betting style of his mentors, he seems to lack the same level of risk management discipline and execution capability.

Key Square's case is a textbook example of the Mentor-Protégé Paradox, showing that reputation cannot replace independently generated Alpha, and "pedigree" is a double-edged sword in the asset management industry.

When discussing Bessent raising $4.5 billion, Marshall Brandt expressed a complex attitude (01:28 - 01:32): "I feel for the guy a little bit. I mean, I don't know how exactly, you know, you put 4.5 billion dollars to work on a Thursday."

This comment touches upon the most core constraints in active asset management: Capacity Constraints and Diseconomies of Scale. For any active strategy, especially one like Global Macro that relies on identifying and exploiting temporary market imbalances, the increase in capital scale will eventually systematically erode its potential to generate Alpha. Launching with a $4.5 billion scale, one of the largest hedge fund launches at the time, may have put Key Square Capital at a significant structural disadvantage from the start, becoming a victim of the "curse of scale."

4.4.1 Berk and Green Model: The Systematic Erosion of Alpha by Scale

Berk and Green's (2004) model provides a rigorous theoretical foundation for understanding diseconomies of scale. The core idea of the model is that the manager's skill is finite, and as the scale increases, the manager is forced to invest marginal capital in suboptimal opportunities (Opportunity Set Constraints), or their trading behavior itself adversely affects market prices (price pressure), thereby reducing the fund's overall return rate.

The fund's Alpha $\alpha(q)$ is a decreasing function of its managed scale $q$:

$\frac{\partial \alpha(q)}{\partial q} < 0$

In rational expectations equilibrium, investors will allocate capital to the fund until the expected Alpha net of fees is zero.

$E[\alpha(q^*)] - Fees(q^*) = 0$

In Bessent's case, the market (incorrectly) attributed extremely high expected Alpha to him based on his pedigree, leading to the $4.5 billion initial capital inflow. However, if his actual strategy capacity was far below this scale, the fund might have been in a state of $q > q^*$ from day one, meaning the expected Alpha was negative. This implies that even if Bessent possessed some skill, the massive scale might have completely offset the advantage brought by this skill.

4.4.2 Market Microstructure and Market Impact Costs

The primary micro mechanism of capacity constraints is market impact costs. When a fund conducts large transactions, its buying and selling behavior directly affects asset prices, causing the actual execution price to be worse than the market price before the transaction. This is a manifestation of limited liquidity.

4.4.3 Modeling Market Impact Costs

Market impact costs are usually modeled as a function of transaction size and market liquidity. The classic Square Root Law posits that market impact costs are proportional to the square root of the transaction size (Almgren & Chriss, 2001; Torre & Bouchaud, 1997):

$Impact\ Cost \approx \beta \cdot \sigma \cdot \sqrt{V/ADV}$

Where $V$ is the transaction size, $ADV$ is the average daily volume, $\sigma$ is the asset's volatility, and $\beta$ is a parameter related to the market microstructure.

For a $4.5 billion fund, even relatively small position adjustments (e.g., establishing a 5% position, i.e., $225 million) can generate significant market impact costs in many macro markets. This directly reduces the strategy's potential returns. Assuming a strategy has a theoretical Alpha of 5%, if the market impact cost reaches 1%, the actual realizable Alpha will decrease by 20%. As the scale increases, the growth rate of market impact costs is faster than the growth rate of the scale, leading to rapid Alpha decay.

4.4.4 The Heterogeneity of Liquidity in Global Macro Markets

Brandt mentioned: "it's not everyone who can establish a position in the Thai baht" (01:00 - 01:02). This highlights the heterogeneity of liquidity in Global Macro markets. While G10 currencies and major government bond markets are extremely liquid, the Alpha sources of Global Macro strategies often lie in broader markets, including emerging market currencies, bonds, Credit Default Swaps (CDS), and less liquid Over-the-Counter (OTC) derivatives.

In these markets, a $4.5 billion scale is enormous. Attempting to establish large-scale positions in a short period not only leads to high transaction costs but may also attract the attention and counter-operations (Front-running or Predatory Trading) of other market participants (including local central banks and hedge fund peers). The fund's trading behavior itself might change the market dynamics, making originally profitable trades unprofitable. This is the opposite of a "self-fulfilling prophecy"—a "self-defeating trade."

For example, if Bessent attempted to short a liquidity-constrained emerging market currency, his selling behavior itself would cause the currency to rapidly depreciate, increasing the cost of establishing the position (causing Slippage), and potentially triggering central bank intervention or capital controls, thus causing the trade to fail. Managing such massive positions requires extremely high trading skills and a deep understanding of market microstructure.

4.4.5 Capital Deployment Pressure and the Psychology of Suboptimal Decision-Making

Brandt's comment, "I don't know how exactly, you know, you put 4.5 billion dollars to work on a Thursday," reveals the immense capital deployment pressure facing a new fund. Holding a large amount of cash creates Cash Drag, affecting the fund's NAV performance. Investors pay high management fees (even without performance fees, assuming a 2% management fee, this amounts to $90 million annually) for their capital to be effectively utilized, not left idle.

This pressure can lead to psychological and behavioral biases in the fund manager, resulting in suboptimal investment decisions:

• Rushed Trading: To quickly reduce cash holdings, the manager might rapidly establish positions without waiting for the optimal entry point or conducting sufficient research, increasing the risk of mispricing and market impact. This behavior is akin to "Forced Trading."

• Lowering Standards: To accommodate more capital, the manager is forced to invest in suboptimal opportunities (Second-best ideas) that do not meet their core strategy or conviction standards. This leads to a decline in portfolio quality and dilution of Alpha.

• Style Drift: To improve liquidity, the manager might be forced to allocate capital to markets with larger capacity but lower Alpha (e.g., shifting from emerging markets to developed markets, or from directional trading to relative value trading), thereby changing the fund's risk-return profile and diluting its core advantage.

In Bessent's case, the initial scale of $4.5 billion likely forced him to conduct large-scale, possibly imprudent capital allocation at the launch phase. If these initial positions performed poorly (considering the adverse factors brought by rushed trading), the fund would immediately fall into trouble, triggering the aforementioned performance-liquidity spiral. Therefore, the massive initial scale might be the "original sin" of his failure.

4.4.6 Organizational Capacity and the Non-linear Growth of Management Complexity

Capacity constraints exist not only at the market level but also at the organizational level (Organizational Capacity). Managing a $4.5 billion Global Macro fund requires a complex organizational structure, robust operational processes, and advanced risk management systems.

As discussed in Part Four, although Bessent had experience working in large institutions, independently establishing and managing an organization of this scale was a completely new challenge. Management complexity grows non-linearly with the increase in the number of positions, asset classes, and markets. Monitoring global cross-asset risk exposures, correlation matrices, scenario analysis, and liquidity conditions is a daunting task. If Key Square Capital's organizational infrastructure (including middle and back-office support, technology systems) failed to match its capital scale, the possibility of Operational Risk and management errors would greatly increase, leading to "Unforced Errors." For example, errors in trade execution, delays or inaccuracies in risk reporting, could all lead to significant losses.

In summary, Brandt's comment on the difficulty of deploying $4.5 billion highlights the enormous constraints of scale on active management. Key Square Capital's failure can largely be attributed to its initial scale exceeding the boundaries of its strategy capacity and organizational capabilities. This provides an important lesson for investors: in hedge fund investing, especially in Alpha-scarce strategies, scale is often the enemy of returns. Many successful hedge funds choose to close to outside capital after reaching a certain scale precisely to protect their Alpha generation capabilities.

Part Five: Deep Dive into Key Square's 13F Holding Data and Correlation with the Macroeconomic Background

To further evaluate Bessent's investment strategy and execution capabilities, we conduct a more detailed analysis of Key Square's 13F holding data and correlate it with the macroeconomic background of the same period. Although 13F data has limitations (only covering U.S. equity longs and some derivatives), it provides objective information about its exposure in the U.S. stock market, offering valuable clues to understanding the evolution of its macro views.

5.1 Limitations and Interpretation Methods of 13F Data

Caution must be exercised when analyzing the 13F data of a Global Macro fund. This data cannot reflect its allocations in foreign exchange, interest rates, commodities, non-U.S. equities, and short positions. The large gap between Key Square's total AUM ($4.5 billion) and its peak 13F AUM ($1.04 billion) confirms this. However, 13F holdings are often used to express specific macro themes or for hedging purposes. We infer the evolution of its strategy by analyzing the industry concentration, style preferences, and correlation with macroeconomic events of its holdings.

5.2 Phase One: Trump Trade and Financial Reflation (2016Q4 - 2017Q4)

Key Square's 13F portfolio in the initial phase (Reference Material "Table 2") clearly reflects the prevailing macro narrative at the time.

• 2016Q4: 13F AUM was $806 million. Major holdings were concentrated in banks (BAC, JPM, GS) and energy (HAL). This is highly consistent with the "Trump Trade" or "Reflation Trade" following the 2016 U.S. election. The market expected tax cuts, deregulation, and increased infrastructure spending to push up interest rates and economic growth.

• 2017: Continued the financial theme, and added gold miners (GDX) as a hedge in Q2. By 2017Q4, 13F AUM reached its peak of $1.041 billion, with increased weight in energy (EOG).

In this phase, Key Square's strategy appears to have been successful. According to reports, it achieved a +13% return in 2016. This indicates that Bessent successfully captured the main macro trends in the initial period.

5.3 Phase Two: Rising Volatility and Defensive Shift (2018Q1 - 2019Q4)

From 2018 onwards, the market environment became more challenging. The Federal Reserve continued to raise interest rates, trade frictions escalated, and global economic growth slowed. Key Square's 13F AUM began to decline significantly.

• 2018: 13F AUM dropped from $873 million in Q1 to $307 million in Q4. The holding structure changed significantly. The weight of financial stocks decreased, shifting towards communications (DISH) and defensive assets.

• 2018Q4: During the sharp market decline, exposure to gold and miners (GDX, NEM) increased. This is a typical risk-off strategy.

In this phase, Key Square's performance began to deteriorate (-7% in 2017, followed by "flat or losses"). The significant decline in AUM indicates that investor confidence began to waver. While the defensive positioning was reasonable, the failure to translate it into positive returns suggests flaws in its macro judgment or hedging strategy.

5.4 Phase Three: Pandemic Shock and Recovery Trade (2020Q1 - 2021Q2)

During the 2020 pandemic, the market experienced extreme volatility. Key Square's holdings show its attempts to adapt to this unprecedented environment.

• 2020Q1: 13F AUM briefly rose to $709 million. The holding structure shifted towards large-cap tech stocks (GOOG, MSFT, META), capturing the trend of tech stocks benefiting in the early pandemic.

• 2020Q2: AUM dropped to $422 million. Hedging positions such as put options on AAPL appeared, indicating skepticism about the strong market rebound or risk management efforts.

• 2020H2-2021H1: Shifted towards the "Recovery Trade," increasing allocation to cyclical industries and commodities (CLF, TECK-B.TO, GM).

The AUM volatility in this phase was high, with frequent changes in holdings. Although certain themes were captured, the overall performance was reportedly "flat or losses," suggesting potential losses in other markets (such as interest rates, foreign exchange), excessively high hedging costs, or timing issues.

5.5 Phase Four: Inflation Trade and Energy/Materials Exposure (2021Q3 - 2022Q4)

From the second half of 2021, inflation pressure emerged, and commodity prices rose. Key Square's 13F portfolio shifted towards energy, materials, and extractive industries.

• 2021Q3-Q4: Heavy positions in CLF, OXY, DVN, BTU, RIG. This indicates Bessent attempted to capture the opportunities of rising inflation and the commodity bull market.

• 2022: Global inflation soared, and the Federal Reserve aggressively raised interest rates. Key Square continued to maintain exposure to energy and commodities, and added gold (NEM) in Q2.

Although these sectors performed well in the inflationary environment, its 13F AUM continued to decline during this period, reaching only $56 million by 2022Q3. This suggests the fund's overall performance might not have been ideal (possibly suffering losses in the interest rate or foreign exchange markets, e.g., failed shorting of government bonds), or investors continued massive redemptions.

5.6 Phase Five: High Concentration and Option Trading Dominance (Endgame) (2023Q1 - 2024Q2)

After experiencing a significant shrinkage in AUM, Key Square's 13F portfolio became highly concentrated, and option trading dominated. This is consistent with the characteristics of "Gambling for Resurrection" when a fund faces an existential crisis.

• 2023: AUM was at a low level, with holdings highly concentrated in a few stocks and options like RIG.

• 2023Q4: AUM suddenly surged to $470 million. Mainly composed of call options on China (FXI calls, KWEB calls) and Brazil (EWZ calls) ETFs. This is a high-leverage, high-risk macro bet, betting on global recovery and Chinese policy stimulus.

• 2024Q1: AUM continued to rise to $529 million. Holdings shifted towards QQQ put options and gold (GLD). This indicates its macro view shifted towards concerns about U.S. tech stocks and hedging demand.

• 2024Q2: 13F AUM sharply dropped to only $14.19 million, with holdings remaining only in two bank ETFs (KRE, KBE). This marks the end of the fund's active operation.

This dramatic change might mean its macro bets failed, or the fund conducted large-scale capital returns, transforming into a family office.

5.7 Evaluating Bessent's Trading Style and Capabilities

Through the analysis of 13F data, we can preliminarily evaluate Bessent's trading style and capabilities. He demonstrated significant strategy adaptability and flexibility, able to quickly adjust the portfolio's risk exposure in different market environments. His investment style is distinctly thematic and tends to take highly concentrated bets.

However, despite these seemingly positive characteristics, the ultimate result of Key Square was failure. This points to several possible explanations:

1. Deficiencies in Timing: Although Bessent could identify the main market themes, he might have systematic deficiencies in the timing of entry and exit.

2. Drag from Non-13F Positions: His performance in the core areas of Global Macro (interest rates, foreign exchange, commodities) might have been far worse than his equity portfolio performance.

3. Insufficient Risk Management and Position Sizing Control: Especially the high-risk option trading adopted in the fund's late stage shows a loss of control in risk management.

Overall, the 13F data analysis supports Brandt's viewpoint. Although Bessent demonstrated certain macro insights at times, he failed to translate these insights into sustained superior performance and failed to effectively manage risks, leading to a significant shrinkage of AUM.

Part Six: The Fallacy of Contrarian Trading and Inconsistency of Investment Philosophy

The latter part of the discussion turned to Bessent's controversial remarks about tariffs and Goldman Sachs. This reveals issues regarding the consistency of his investment philosophy and intellectual honesty, which are crucial for evaluating a macro fund manager.

6.1 The Logical Fallacy of "Trading Against Goldman Sachs" and Contrarian Investment Theory

Bessent claimed (02:30-02:37): "I made a good career trading against Goldman Sachs." He attempted to portray himself as a successful contrarian trader, using this statement to support his unconventional views on the impact of tariffs and refute the analysis of mainstream institutions like Goldman Sachs.

The theoretical basis of successful Contrarian Investing stems from behavioral finance, which posits that the market exhibits overreaction and herd mentality (De Bondt and Thaler, 1985). However, successful contrarian investing is not merely blindly going against the consensus. It requires "Variant Perception," i.e., a deep understanding of the fundamentals and the ability to identify errors in the market consensus.

Bessent's statement contains several serious logical fallacies:

• Hasty Generalization: Even if he successfully traded against Goldman Sachs' views in some past transactions, it does not prove that trading against Goldman Sachs is a universally effective strategy. Brandt pointed out (03:13-03:27) that Goldman Sachs is considered a "thought leader in the investment space," representing high-quality market consensus. Systematically trading against Goldman Sachs and succeeding is extremely difficult.

• Survivorship Bias and Post Hoc Attribution: Bessent might only remember the trades where he went against Goldman Sachs and succeeded, forgetting the failed trades. This is a typical Confirmation Bias.

• Ad Hominem and Irrelevant Evidence: In the debate about the impact of tariffs, Bessent did not provide economic theory or empirical data to support his views but instead attacked Goldman Sachs' credibility and claimed he made money on them in the past (03:00-03:12). This is the fallacy of irrelevant evidence.

6.2 The Economics of Tariffs: Theory and Evidence

The discussion focused on the issue of tariffs (02:38 - 03:12). Tim Miller pointed out that the mainstream economic view is that the cost of tariffs is mainly borne by U.S. companies and consumers, while Bessent opposed this.

In international trade theory, the Tariff Incidence depends on the relative price elasticities of supply and demand. Although the U.S., as a large country, might improve its Terms of Trade through tariffs, passing some costs onto foreign producers (optimal tariff theory), empirical studies generally show that the costs of tariffs imposed by the U.S. in recent years have been almost entirely borne domestically. For example, studies by Amiti, Redding, and Weinstein (2019) and Fajgelbaum et al. (2020) confirm this.

Therefore, Bessent's claim that the cost of tariffs is not borne by U.S. consumers or companies contradicts mainstream economic theory and empirical evidence.

6.3 Inconsistency of Investment Philosophy and Cognitive Dissonance

More seriously, Brandt pointed out Bessent's self-contradiction on the issue of tariffs (03:28-03:49): "We've got an investor letter of his where he talks about how tariffs are going to be incredibly inflationary." Tim Miller added (03:43-03:49) that Bessent had actually bet that tariffs would be inflationary when he was running a hedge fund.

If these allegations are true, then Bessent's current public stance is completely contrary to his past investment views and bets. This reveals a serious inconsistency in his investment philosophy and macro analysis framework. In asset management, Consistency of Investment Philosophy is crucial. It is the foundation of investor trust and an important basis for evaluating the manager's skill.

Bessent's shift in stance on tariffs seems not based on new economic evidence but driven by political motives (considering his background as a nominee for Treasury Secretary, 00:14-00:21). This behavior of changing basic economic views to cater to specific political narratives raises serious questions about his Intellectual Honesty.

This can be explained by Cognitive Dissonance (Festinger, 1957) or Motivated Reasoning (Kunda, 1990). Individuals tend to process information and form beliefs in ways that align with their own goals or interests.

6.4 Intellectual Honesty and Professional Credibility

For Global Macro traders, the ability to objectively analyze the real world is crucial. If their analytical framework is distorted by political bias, they are likely to suffer losses in the market. Bessent's record of failure at Key Square, coupled with the logical fallacies and inconsistencies demonstrated in public policy debates, may collectively reflect deeper flaws in his ability to analyze complex systems.

Bessent's remarks about "trading against Goldman Sachs" and his self-contradiction on tariffs reveal deeper issues than his poor investment performance: lack of logical rigor, inconsistency of investment philosophy, and potential intellectual dishonesty. This behavior damages his credibility as a professional. In the field of Global Macro, which relies on clear analytical frameworks and objective judgment, such flawed thinking can be fatal. The historical trajectory of Key Square shows that in Global Macro trading, lacking intellectual honesty, discipline, and adaptability cannot guarantee success, even with the most distinguished background and the largest initial capital.

Data Appendix: Key Square Capital Management LLC Public Data Compilation

Table 1｜Key Square 13F Reported AUM History (Unit: $ Thousands)

Source: Holdings Channel. Note: 13F AUM covers only U.S. equities and reportable options market value.

Filing Date	13F AUM ($ Thousands)
2016-12-16	805,804
2017-03-17	450,927
2017-06-17	671,394
2017-09-17	616,547
2017-12-17	1,040,796
2018-03-18	872,919
2018-06-18	476,920
2018-09-18	442,797
2018-12-18	306,784
2019-03-19	292,587
2019-06-19	274,837
2019-09-19	363,244
2019-12-19	302,226
2020-03-20	709,382
2020-06-20	422,122
2020-09-20	195,248
2020-12-20	417,028
2021-03-21	587,008
2021-06-21	528,605
2021-09-21	271,983
2021-12-21	245,472
2022-03-22	248,952
2022-06-22	192,436
2022-09-22	56,139
2022-12-22	103,516
2023-03-23	120,167
2023-06-23	49,940
2023-09-23	21,670
2023-12-23	469,992
2024-03-24	528,655
2024-06-24	14,198

Table 2｜Quarterly 13F Portfolio "Value + Holdings Snapshot + Line-by-Line Annotation" (2016Q4–2024Q2)

Source: 13f.info / SEC original filings.

Quarter	13F Value ($K)	Positions	Top Holdings (Example)	Annotation (Based on Holdings and Macro)
2016Q4	805,804	35	BAC, JPM, GS, HAL	High weight in banks/energy; tailwind from post-US election bank sector rally and yield rise.
2017Q1	450,927	24	BAC, JPM, MON, BABA	Still favoring financials; 2017 US equity rally continued, tax reform expectations supported bank valuations.
2017Q2	671,394	34	BAC, JPM, MON, GDX	Gold miners (GDX) appeared; mid-bull market but starting to add defensive/gold exposure.
2017Q3	616,547	34	BAC, JPM, MON, C	Financial theme continued.
2017Q4	1,040,796	41	BAC, JPM, TWTR, EOG	Strong US equities around tax reform passage, banks benefited; energy EOG weight increased.
2018Q1	872,919	33	BAC, HAL, JPM, CSCO	Early 2018 volatility climbed, rate hike expectations rose; portfolio still favoring financials/energy.
2018Q2	476,920	28	EOG, DISH, SLB, TEO	Energy/communications weight; market volatility during the year, rate hikes proceeded.
2018Q3	442,797	25	DISH, FHB, TEO, PM	Proportion of defensive/communications individual stocks increased.
2018Q4	306,784	18	DISH, GDX, NEM, GOOS.TO	Q4 US equities declined near bear market edge; gold and mining exposure (GDX/NEM) increased.
2019Q1	292,587	17	DISH, WFC, LNG, TEO	Context of "Fed Pivot" (2019 rate cuts) and 2019 US equity strong year.
2019Q2	274,837	17	DISH, LNG, TV, INTEQ	Stable scale, focused on individual stocks/themes.
2019Q3	363,244	22	GOOG, DISH, INTEQ, GE	Tech weight recovered; US equities remained strong during the year.
2019Q4	302,226	21	DISH, GOOG, INTEQ, XP	S&P closed the year with best performance since 2013.
2020Q1	709,382	25	GOOG, DISH, MSFT, META	Pandemic shock period; Fed cut rates to near zero in March, coexistence of crash/rebound.
2020Q2	422,122	11	NKLA, DISH, AAPL puts, MTCH	Q2 one of the strongest rebounds in history; appearance of hedges like AAPL puts.
2020Q3	195,248	13	DISH, TECK-B.TO, NSC, GM	Recovery trade continued, cyclical allocations visible.
2020Q4	417,028	17	DISH, TECK-B.TO, CLF, MSFT	Vaccine news and fiscal stimulus boosted risk assets; portfolio expanded again.
2021Q1	587,008	29	DISH, CLF, GOOG, MSFT	Cyclicals + Tech in parallel amid 2021 commodity rebound/economic reopening background.
2021Q2	528,605	28	DISH, OXY, GOOG calls, IBB	Coexistence of energy (OXY) and bullish GOOG options.
2021Q3	271,983	19	CLF, DISH, OXY, DVN	High weight in materials/energy, oil prices rebounded throughout the year.
2021Q4	245,472	26	DISH, VNT, BTU, RIG	Energy/extractive (BTU, RIG) significant.
2022Q1	248,952	37	RIG, DISH, JNPR, SNE	2022 inflation rising and rate hikes started, commodity-related exposure prominent.
2022Q2	192,436	21	XHB calls, NEM, JNPR, RIG	Inflation peaked at 9.1% in the period; gold/mining (NEM) and housing ETF options seen together.
2022Q3	56,139	6	RIG, GDX, FLR, BTU	Scale contracted, still retained energy/gold mine exposure.
2022Q4	103,517	12	RIG, WFRD, BTU, META	Expectations of inflation falling and rate hikes slowing emerged.
2023Q1	120,167	13	FXI, RIG, RIG calls, BTU	China reopening/policy expectations drove volatility-related trades in China stocks and commodities.
2023Q2	49,940	9	RIG calls, RIG, URI, WBD	Coexistence of energy and selected US individual stocks.
2023Q3	21,669	2	RIG calls, RIG	Portfolio highly concentrated.
2023Q4	469,992	8	FXI calls, EWZ calls, KWEB calls, RIG calls	SEC filings show option positions on China/Brazil ETFs and offshore drillers; market expectations for China stimulus and Brazil highs at the time.
2024Q1	528,653	5	QQQ puts, GLD, UNP, LYB	Nasdaq strong in the quarter, gold prices hit new highs; visible hedging (QQQ puts) and gold positions.
2024Q2	14,197	2	KRE, KBE	Only two bank ETF exposures; reflecting bank sector volatility and interest rate expectation changes in the quarter.

Table 3｜(Supplementary) Excerpts of Public Reports on "Flagship Hedge Fund Performance"

Source: Reuters reports citing investor materials and sources familiar with the matter.

Year/Period	Reported Performance	Notes
2016	+13%	Benefited from GBP/Brexit and long trades after Trump's election.
2017	-7%
2018–2021	Precise percentage not disclosed ("lost money or were roughly flat")
2023	Double-digit positive returns
2024 (as of Nov)	Double-digit positive returns, November best