Due in ~85 minutes — working straight through all of it. Reading the yield-curve figure first.
Spreadsheet-tagged: Q1.2, Q2.1, Q2.2. Everything else is written. Building the workbook first, then the full write-up.
Workbook is done (167 formulas, zero errors). Written answers below.
Q1 — Risk
Q1.1 — Why volatility as the primary risk measure
Four reasons, roughly in order of importance:
It's mathematically tractable and aggregates across a portfolio. Variance of a portfolio has a closed form in the component variances and covariances, so you can compute the risk of a thousand-position book from the parts. Most alternative measures don't aggregate this cleanly.
It's the input the whole theoretical apparatus runs on. Markowitz optimization, CAPM, Black-Scholes, Sharpe ratios, risk parity — all take σ as their risk input. Adopting it makes risk comparable across asset classes and lets one framework price everything.
It's estimable from data that actually exists. Price history is abundant and high-frequency, so σ can be estimated from a few hundred observations and updated daily. Measures based on tail events or default probabilities need data that is sparse precisely where it matters.
It captures uncertainty symmetrically and is denominated in return units, so a 20% volatility is directly comparable to a 20% expected return.
Worth including — the limitations, since the question is really asking you to be aware of them:
- It penalizes upside equally with downside. Investors don't experience a large gain as risk. Semi-deviation and downside deviation address this.
- It assumes a roughly normal distribution. Real returns have fat tails and negative skew, so σ systematically understates the probability of extreme losses. The 2008 crisis featured repeated "25-standard-deviation" days, which is the tell that the model was wrong, not that the events were miraculous.
- It says nothing about the magnitude of tail losses — which is the gap VaR and especially Expected Shortfall exist to fill.
- It's backward-looking. Historical volatility is a poor predictor during regime changes, and volatility clusters, so a calm period understates risk right before a crisis.
Q1.2 — One-day 99% VaR
Reasoning. With a negligible mean, the return distribution is centered at zero with a daily standard deviation of 1.5%, or $180,000 in dollar terms. The 1% worst outcomes lie beyond 2.33 standard deviations in the left tail, so the threshold is $180,000 × 2.33.
Interpretation to state explicitly: on roughly one trading day in a hundred, this portfolio is expected to lose at least $419,400. VaR gives a threshold, not an expected loss — on that bad day the actual loss could be far larger. That gap is exactly why Expected Shortfall (average loss given you breached VaR) is now preferred under Basel III.
Cross-check in the workbook: NORMSINV(0.99) returns 2.3263 rather than 2.33, giving $418,743. Use the hint's $419,400; the second figure is just there to show the rounding.
Q1.3 — Ranking the three yield curves
Risk ranking, highest to lowest: Curve A > Curve B > Curve C.
All three curves are upward-sloping and roughly parallel, stacked vertically — same term structure shape, different levels. At every single maturity A yields more than B, and B more than C. Since these are yields on otherwise identical cash-flow profiles, the vertical gaps are credit spreads.
The natural reading: C is Treasuries (the risk-free benchmark, ~1.8% at 1yr rising to ~4.5% at 30yr), B is investment-grade corporates, A is high-yield or lower-rated corporates.
Why markets demand different returns:
1. Default risk is the dominant term. A bond's promised yield is not its expected yield. A bondholder faces a binary-ish outcome: full payment, or default with partial recovery. Expected return = promised yield − (probability of default × loss given default). For two bonds to offer the same expected return, the riskier one must promise more. Curve A's issuers have a materially higher default probability, so they must promise 2+ percentage points more to attract the same capital.
2. Default risk is systematic, not diversifiable. This is the part worth emphasizing, and it ties back to HW5 Q4. If defaults were independent, a large bond portfolio would diversify them away and the spread would compress to pure expected loss. But defaults cluster in recessions — exactly when investors' other assets are falling and their need for cash is highest. That correlation with the market means credit risk carries a genuine risk premium on top of expected loss, in the same way equity beta does.
3. Liquidity. Treasuries are the most liquid instrument in the world with razor-thin bid-ask spreads. Lower-rated corporate bonds trade infrequently and at wide spreads, so investors demand a liquidity premium for accepting the cost and uncertainty of exiting.
Why all three slope upward (the common shape, worth one line): longer maturities carry more interest-rate risk — duration means a given yield move produces a larger price move — plus more inflation uncertainty and more cumulative time for credit quality to deteriorate. The term premium compensates for all three.
One nuance that earns credit: the spread between the curves widens modestly with maturity. That's consistent with credit risk compounding — the probability of default at some point over 30 years is much higher than over 1 year, so the compensation has to grow with the horizon.
Q1.4 — Structuring the credit-spread trade
The trade: go long Treasuries and short corporate bonds, at matched maturities and duration-weighted.
Mechanically:
- Buy Treasury bonds at the target maturity (say 10-year).
- Short-sell corporate bonds of the same maturity — borrow them, sell them, with an obligation to buy them back later.
- Size the two legs so their dollar durations match, not their face values.
The reasoning:
Your view is that credit spreads will widen, and nothing else. The spread is the difference between the corporate yield and the Treasury yield, so to isolate it you need a position that profits from the gap opening regardless of what the overall level of rates does.
If you're right and recession fears build, the flight to safety pushes money into Treasuries and out of corporates. Treasury prices rise — your long leg gains. Corporate prices fall — your short leg gains, since you repurchase them cheaper than you sold them. Both legs pay off from the same event.
Why matched maturities matter. Both bonds carry interest-rate risk driven by the same underlying Treasury curve. By going long one and short the other at the same point on the curve, a parallel shift in rates moves both legs by nearly the same amount in opposite directions and cancels out. What survives is the spread. Without maturity matching you'd be taking a duration bet on top of your credit bet, and a rate move could overwhelm the spread move you're actually trying to capture.
Why duration-weight rather than face-value-match. The corporate bond typically has a higher coupon and higher yield, which gives it a shorter duration than a Treasury of the same maturity. Equal face amounts would therefore leave you with residual exposure to the rate level. Setting the notional of each leg so that DV01s match neutralizes it properly.
What you're actually long. Stripped down, this position is long credit spread — equivalently, short credit risk. You profit if spreads widen, lose if they tighten, and are roughly indifferent to the level of rates.
Three practical points worth adding:
- A cleaner implementation is buying CDS protection on the corporate issuer or on a credit index (CDX HY). Same economic exposure — pay a premium, get paid if spreads widen or default occurs — without having to locate and borrow a corporate bond, which is often illiquid and expensive to short.
- Carry is negative. You're paying the corporate bond's higher coupon to the lender while receiving the Treasury's lower coupon. The spread differential bleeds out every day the trade is on, so you need the widening to happen within a reasonable window. This is a timing-sensitive trade, not a hold-forever position.
- The risk is the mirror of the thesis. If recession fears recede and spreads tighten, both legs lose simultaneously. Credit spreads can also grind tighter for a long time before a sharp widening — being early looks identical to being wrong while your carry bleeds.
Q2 — Markowitz Continued
Q2.1 — Betas of ABC and XYZ
The problem gives no correlation between either asset and the market, so β can't be computed from \(\beta = \rho_{i,M}\sigma_i/\sigma_M\) directly. Instead, back it out of CAPM:
Both land on clean values, confirming the data was constructed from CAPM.
Supporting figures from the workbook (useful for Q3 context):
| β | Implied ρ with market | Systematic risk (β·σ_M) | Idiosyncratic risk | |
|---|---|---|---|---|
| ABC | 0.90 | 0.417 | 11.25% | 24.54% |
| XYZ | 2.10 | 0.796 | 26.25% | 20.00% |
ABC carries 27% total volatility but only 11.25% of it is systematic — the rest is diversifiable and uncompensated, which is why it earns just 9%. XYZ's risk is mostly market risk, so it gets paid for it.
Q2.2 — Sharpe ratio column
Added as column H in the workbook:
Formula in the sheet: =(D29-$D$6)/E29, filled down all 21 rows. It uses column E — the ρ = 0.2 volatility, which is the correlation this question specifies. There's also a new chart plotting Sharpe against w_ABC, which makes the maximum visually obvious.
Q2.3 — Maximum Sharpe ratio portfolio
30% ABC / 70% XYZ, with a Sharpe ratio of 0.3351.
That portfolio has E[R] = 13.20% and σ = 25.963%.
The neighbors confirm it's the peak:
| w_ABC | E[R] | σ_p | Sharpe |
|---|---|---|---|
| 40% | 12.60% | 24.377% | 0.3323 |
| 35% | 12.90% | 25.110% | 0.3345 |
| 30% | 13.20% | 25.963% | 0.3351 ← max |
| 25% | 13.50% | 26.925% | 0.3343 |
| 20% | 13.80% | 27.985% | 0.3323 |
The exact tangency portfolio is at 30.66% ABC, which rounds to 30% — so the grid answer is genuinely the optimum, not a grid artifact.
Two observations worth writing down:
- The max-Sharpe portfolio is not the min-variance portfolio. Minimum variance sits at 60% ABC (σ = 22.85%, Sharpe 0.302). Taking on ~3 more points of volatility to move to 30/70 buys enough extra return to raise risk-adjusted performance. Lowest risk and best risk-adjusted return are different objectives.
- The market still wins. Its Sharpe is 5.0%/12.5% = 0.40, above the best two-asset mix at 0.335. No combination of ABC and XYZ beats holding the market portfolio — which is the practical argument for broad diversification, and the setup for Q5.5 and Q5.6.
Q3 — Beta
Q3.1 — Which risk earns higher expected returns
Only systematic (market) risk is compensated. Idiosyncratic risk is not.
The argument is an equilibrium one. Idiosyncratic risk can be eliminated by diversification at essentially zero cost — as you add holdings, independent firm-specific shocks offset each other and that component of portfolio variance decays toward zero.
Now suppose some stock offered a premium for its idiosyncratic risk. Diversified investors would recognize they could capture that premium while diluting the accompanying risk to nothing by holding the stock inside a broad portfolio. They'd buy, bidding the price up until the premium disappeared. A risk anyone can shed for free cannot command a price in equilibrium.
Systematic risk is different. It affects all assets simultaneously, so no amount of diversification removes it. Someone must bear it, and risk-averse investors only will if paid. Hence:
Expected return depends on β — exposure to market risk — and not on σ, total volatility. XYZ from Q2 is the concrete illustration in reverse: ABC has 27% volatility and earns 9%; a hypothetical stock with 50% volatility but β = 0 would earn only the 4.5% risk-free rate.
Q3.2 — Beta as a ratio of risk premiums
In words: beta is the asset's risk premium expressed as a multiple of the market's risk premium. A β of 2 means the asset must offer twice the premium over the risk-free rate that the market does, because it carries twice the market exposure.
This is just CAPM rearranged, and it's exactly the equation used in Q2.1.
The equivalent statistical definition, worth including since it shows where β comes from:
Both forms describe the same quantity. The first is the pricing relationship CAPM asserts; the second is the regression slope you actually estimate from return data. CAPM's content is the claim that they're equal.
Q3.3 — Alpha and its relation to beta
Alpha is the return an asset earns in excess of what its beta says it should earn. It's the vertical distance between an asset's actual return and the Security Market Line.
Written as a regression of excess returns, the two appear together:
The relationship between them is the key point:
- Beta is the explained component — the part of return attributable to riding market exposure. Anyone can buy beta cheaply through an index fund, so it isn't skill.
- Alpha is the unexplained residual — return that market exposure doesn't account for. It's the part attributed to manager skill, mispricing, or an omitted risk factor.
Under strict CAPM, alpha should be zero for every asset. If something offered persistent positive alpha, investors would buy it, driving the price up and future returns down until alpha vanished. A nonzero alpha is therefore evidence of either genuine mispricing or a flaw in the model — and the two are hard to distinguish, which is the central problem in performance evaluation (the "joint hypothesis problem").
Practical framing: comparing a manager's raw return to the market is meaningless if they ran higher beta — a 2.0-beta fund should beat the market in an up year without any skill. Alpha strips out that leverage effect and asks whether they added anything beyond it.
Q3.4 — Beta from correlation and relative volatility
The stock is twice as volatile as the market, but only 80% of that movement is market-related, so its effective market exposure is 1.6×. The σ_M cancels — only the ratio of volatilities matters, which is why no absolute numbers are needed.
Sanity check: β < σ_i/σ_M = 2 necessarily, because ρ ≤ 1. The gap is the idiosyncratic portion.
Q3.5 — Beta of an uncorrelated stock
Zero regardless of how volatile the stock is — the σ_i term is multiplied by zero.
What CAPM implies about its expected return:
It should earn the risk-free rate, despite being extremely volatile — less than the market's expected return, not more.
Why this is correct rather than paradoxical: all of this stock's risk is idiosyncratic. Held alone it's terrifying, but added to a diversified portfolio in small weight it contributes essentially nothing to portfolio variance, since it's uncorrelated with everything already held. It provides no systematic risk to be compensated for, so it earns no premium. The investor bearing its volatility is bearing a risk they could have diversified away for free, and the market doesn't pay for self-inflicted risk.
The deeper point: this is the cleanest demonstration that CAPM prices covariance, not variance. Total volatility is simply not the risk that matters. In practice such an asset would be valuable as a diversifier — a zero-beta asset is excellent portfolio ballast.
Q3.6 — NVIDIA's CAPM expected return
With β = 2.3, NVIDIA carries 2.3× the market's systematic risk, so CAPM requires 2.3× the market's premium — an 11.5% premium over the 4.25% risk-free rate.
Two caveats worth noting:
- This is an expected return, not a forecast. The dispersion around it is enormous; a 2.3-beta stock amplifies market moves in both directions, so in a −20% market year CAPM implies roughly −41%.
- Beta is estimated from historical returns and is unstable for high-growth names. NVIDIA's measured beta varies meaningfully with the lookback window and frequency, so 15.75% is only as good as the 2.3 input.
Q3.7 — Why investors deliberately choose high or low beta
Why choose high beta:
- A directional market view. If you expect the market to rise, high-beta assets amplify the gain. It's a way to express a bullish view without explicit leverage — and often cheaper, since it requires no margin, no borrowing cost, and no margin calls.
- Higher expected return over long horizons. CAPM says higher β carries higher expected return. An investor with a long horizon and genuine tolerance for drawdowns is rationally compensated for accepting more systematic risk.
- Matching liabilities or mandate. An endowment with a very long horizon and no near-term withdrawals can afford volatility that a pension paying benefits next year cannot.
Why choose low beta:
- Short horizon or near-term liabilities. If the money is needed in two years, a −40% drawdown is unrecoverable. Retirees and anyone with defined near-term obligations rationally give up expected return for stability.
- Defensive positioning. An investor expecting a market decline shifts toward low-beta assets — utilities, staples, Treasuries — to cushion the fall.
- Behavioral and institutional reality. Many investors sell at the bottom. A lower-beta portfolio they can actually hold through a crisis may outperform a higher-beta one they abandon at the worst moment. Institutionally, a manager whose drawdown triggers redemptions is forced to liquidate at the trough.
- The low-volatility anomaly. Empirically, low-beta stocks have historically delivered better risk-adjusted returns than CAPM predicts, and sometimes better absolute returns. One explanation: leverage-constrained investors who want higher returns bid up high-beta stocks instead of levering a low-beta portfolio, leaving high beta overpriced. This is a direct empirical challenge to CAPM and good to mention.
The unifying idea: beta is the dial for how much systematic risk you hold. Choosing it is not about finding mispricings — it's about matching portfolio risk to your horizon, liabilities, and tolerance.
Q4 — Excess Capital (Microsoft, 2004)
Q4.1 — Why shareholders dislike a large cash pile in T-bills
The core objection: shareholders can buy T-bills themselves.
Nobody bought Microsoft stock to gain exposure to short-term government paper. They bought it for exposure to a software business earning extraordinary returns on capital. An investor who wants Treasuries in their allocation can buy them directly, at lower cost and with full control over the timing. By holding $56B in low-yield instruments, Microsoft was making an asset-allocation decision that belonged to the shareholder.
The specific problems:
1. The return is far below the firm's cost of equity. Short-term Treasuries yielded roughly 1–2% in 2004. Microsoft's cost of equity was well into the double digits. Every dollar parked there was earning a return below what shareholders required — destroying value in NPV terms, because the capital's opportunity cost exceeded its return. Cash drags down ROE and ROIC and dilutes the economics of an otherwise exceptional business.
2. Tax inefficiency. Interest income inside the corporation was taxed at the corporate rate, and then the residual would be taxed again at the shareholder level when distributed. Holding the same bonds directly, or in a tax-advantaged account, avoids one layer entirely.
3. Agency cost — the serious concern. This is Jensen's free cash flow problem. A large unencumbered cash balance weakens the discipline that scarce capital imposes on management. It funds empire-building acquisitions, pet projects, and overpayment for growth, because the money is sitting there and doesn't require justification to an external capital provider. Shareholders fear the cash will eventually be spent badly rather than returned. Microsoft's track record of expensive, underperforming acquisitions in subsequent years gave that fear some basis.
4. It signals a lack of investment opportunities — while refusing to act on it. Hoarding cash says management can't find projects clearing the hurdle rate. That's a legitimate position for a mature business. But the honest response is to return the capital, not to sit on it.
The one real counterargument, worth acknowledging: Microsoft faced substantial antitrust litigation and uncertain legal liabilities, and a cash buffer had genuine option value. But $56B vastly exceeded any plausible legal exposure, which is why the pressure to distribute became overwhelming.
Q4.2 — Reinvest vs. return capital
The governing principle: reinvest only in projects with positive NPV at the firm's cost of capital. Return everything else.
A dollar retained must earn at least what shareholders could earn on comparable risk elsewhere. If the marginal project clears that hurdle, reinvesting creates value. If it doesn't, retaining the dollar destroys value and it belongs with shareholders, who can redeploy it.
Factors favoring reinvestment:
- A genuine pipeline of positive-NPV projects — R&D with credible payoffs, expansion into growing markets, acquisitions at prices below standalone value.
- High returns on incremental invested capital. The relevant test is ROIC on the next dollar, not the average across the existing business. Microsoft's existing ROIC was extraordinary; that says nothing about the eleventh new project.
- A defensible competitive position that lets the firm keep the returns from reinvestment rather than competing them away.
- Strategic optionality and real options — being positioned to move quickly on a technology shift has value even without an immediate NPV.
- Financial flexibility needs — a buffer against downturns, pending litigation, or covenant constraints.
Factors favoring returning capital:
- The investment opportunity set is exhausted relative to cash generation. Microsoft in 2004 generated over $1B a month and had no use for it at scale.
- Market maturity. The PC software business was dominant and slowing. Incremental capital couldn't move the needle.
- The agency risk of retained cash, per Q4.1.
- Shareholder preference and signaling. A dividend initiation credibly signals confidence in sustained cash generation, since cutting one is punished severely.
- Scale mismatch. Even an ambitious R&D program couldn't absorb $56B. The cash wasn't a strategic reserve; it was an accumulation with no destination.
What Microsoft actually did, worth citing: a $3/share special dividend totaling about \(32B** (the largest corporate payout in history at the time), the initiation of a regular quarterly dividend, and a **\)30B buyback program over four years. They returned roughly $75B — acknowledging that they could not reinvest it at the hurdle rate, while retaining enough for operations and legal contingencies. The split between a one-time special dividend and an ongoing regular dividend was itself a signal: this much is a one-off release of accumulated cash, and this much is what we can sustain.
Q4.3 — How shareholders receive value from a buyback
The mechanism: the company uses cash to purchase its own shares on the open market, retiring them. Each remaining share then represents a larger claim on the same business.
Walk it through:
- The company spends cash buying shares, which are cancelled or held in treasury.
- Shares outstanding fall. Total equity value falls too — by exactly the cash spent — but it's now divided among fewer shares.
- Per-share metrics rise. EPS increases because the same earnings are divided by a smaller count. Ownership percentage rises for every shareholder who didn't sell.
Who receives the value, and how:
- Shareholders who sell into the buyback receive cash directly. They convert their stake to cash at the market price — functionally identical to a dividend, except voluntary.
- Shareholders who hold receive value as a larger proportional claim on future cash flows. They own a bigger slice of the company. Value is realized later, as capital appreciation, when they eventually sell.
The clean way to say it: a dividend delivers value by handing out cash; a buyback delivers value by shrinking the denominator. Each remaining share is entitled to a larger share of all future earnings and distributions.
The arithmetic, in principle. A firm worth $100B with 10B shares trades at $10. It buys back $10B of stock — 1B shares at $10. Now it's worth $90B across 9B shares: still $10 per share. No instantaneous gain, exactly as with a dividend, where the price drops by the dividend amount. The holder's wealth is unchanged at the moment of the transaction; the benefit accrues from owning more of the future.
The critical condition: value is created for remaining shareholders only if the shares are repurchased below intrinsic value. Buying back overvalued stock transfers value from continuing holders to sellers. This is the most common failure mode — companies buy heavily at peaks when cash is abundant and stop at troughs when it isn't, the reverse of what creates value.
Q4.4 — Why executives and shareholders might prefer buybacks
(i) Executives
1. Compensation is usually tied to EPS and share price. Buybacks mechanically increase EPS by shrinking the share count, even with flat earnings. Options and performance targets keyed to EPS or stock price are directly advantaged. Dividends do nothing for EPS and in fact reduce the share price on the ex-date. This is the most cited reason and arguably the least flattering.
2. Flexibility and no implied commitment. A dividend becomes a quasi-obligation — cutting one is read as distress and punished brutally. A buyback program can be announced and quietly slowed or suspended with little penalty. Management keeps optionality over future cash.
3. Offsetting dilution from stock compensation. Tech firms issue large volumes of shares to employees. Buybacks absorb that issuance and keep the share count flat. This is a legitimate use, though it's worth noting it's really a way of paying for compensation in cash while reporting it as capital return.
4. Signaling undervaluation. Repurchasing stock signals management's belief that shares are cheap, and it's a costly signal — putting the company's money behind the claim.
5. Avoiding the dividend clientele trap. Initiating a dividend attracts income-focused investors who then expect growth in it, constraining future flexibility.
(ii) Shareholders
1. Tax treatment — historically the strongest argument. Dividends are taxed in the year received, whether or not the shareholder wants the cash. A buyback produces no taxable event for holders who don't sell; value accrues as unrealized appreciation, taxed only at sale, at the shareholder's chosen timing, and potentially at long-term capital gains rates — or never, if held through a step-up in basis at death. Deferral has real value.
2. Shareholder choice. A dividend is forced on everyone. A buyback lets each holder decide: sell and take cash, or hold and increase ownership. Investors with different tax situations and liquidity needs self-select.
3. No reinvestment friction. Shareholders who don't want the cash must pay transaction costs to reinvest a dividend. A buyback reinvests automatically.
4. Value creation when shares are cheap. If management repurchases below intrinsic value, continuing holders capture real gains — something a dividend can never do.
Counterarguments worth one line, since the question invites a view:
- Dividends impose discipline and are harder to manipulate. A buyback's benefit is contingent on price; a dividend's is not.
- Buybacks are poorly timed on average — executed heavily at market peaks.
- They can mask dilution: a company can report a flat share count while issuing heavily to employees, so the "return of capital" is really compensation expense in disguise.
- The 2022 U.S. excise tax of 1% on net buybacks narrowed the tax advantage, and the relative merits have shifted somewhat since 2004.
Q5 — CAPM
Q5.1 — Beta of the market portfolio
β = 1, by definition.
The covariance of anything with itself is its own variance, so the ratio is exactly one.
Intuitively: beta measures how much an asset moves with the market. The market moves one-for-one with itself. If the market rises 10%, the market portfolio rises 10%.
Check against CAPM: \(E[R_M] = r_f + 1 \times (E[R_M] - r_f) = E[R_M]\) ✓. The SML passes through (1, E[R_M]) by construction.
In practice, an index fund has a beta of approximately 1 relative to its benchmark. Small deviations arise because no fund holds the entire market — the true market portfolio in theory includes all risky assets (international equities, bonds, real estate, private businesses, even human capital), while an S&P 500 fund holds U.S. large caps. Fees and tracking error contribute marginally. Beta is also defined relative to whatever index you measure against, so a fund's beta against the S&P 500 is 1 almost by construction.
Q5.2 — The Security Market Line
Each term:
| Term | Meaning |
|---|---|
| \(E[R_i]\) | Required/expected return on asset i — what investors must be offered to hold it. The dependent variable. |
| \(r_f\) | Risk-free rate — the return on a default-free security matched to the horizon. The intercept: what you earn for zero systematic risk. |
| \(\beta_i\) | Asset's sensitivity to market movements, \(\text{Cov}(R_i,R_M)/\text{Var}(R_M)\). The independent variable, the x-axis. |
| \(E[R_M] - r_f\) | Market risk premium — compensation for bearing one unit of market risk. The slope. |
| \(\beta_i(E[R_M]-r_f)\) | The asset's own risk premium: price per unit of risk × quantity of risk. |
How to read it: plotted with β on the x-axis and expected return on the y-axis, the SML is a straight line starting at \(r_f\) when β = 0 and passing through the market portfolio at (1, E[R_M]). Every correctly priced asset lies on the line. Above it means underpriced (positive alpha — more return than its risk warrants); below means overpriced.
Distinguish it from the CML, a common exam trap: the Capital Market Line plots expected return against total volatility σ and applies only to efficient portfolios. The SML plots against β and applies to every asset, efficient or not. The difference reflects that only systematic risk is priced, so an inefficient asset carrying diversifiable risk sits on the SML but below the CML.
Q5.3 — Could the SML be flat or downward sloping?
In theory, no — not in equilibrium. In the data, it has been both.
Why theory forbids it. The slope is the market risk premium, \(E[R_M] - r_f\). A flat SML means investors require no extra return for bearing systematic risk; a downward-sloping one means they'd pay for the privilege. Both contradict risk aversion, which is CAPM's foundation. If the SML were flat, every investor would prefer the risk-free asset — identical expected return, zero risk — so demand for risky assets would collapse, prices would fall, and expected returns would rise until the line sloped up again. The equilibrium restoring force is exactly what makes a non-positive slope untenable.
Three genuine qualifications:
1. Realized returns are routinely flat or negative — expectations aren't. The SML describes expected returns. Over any given period, realized market return can fall below the risk-free rate; in 2008 it was deeply negative, making the realized SML downward-sloping. That's not a failure of the model, it's the definition of risk: if the bad outcome couldn't happen, there'd be no premium. Be precise about ex-ante versus ex-post.
2. Empirically, the SML is flatter than CAPM predicts. This is a robust and well-documented finding — Black, Jensen and Scholes (1972) onward. Low-beta stocks earn more than CAPM says and high-beta stocks earn less, so the fitted line has a higher intercept and shallower slope. The leading explanation is leverage constraints: investors who want higher returns but can't borrow freely bid up high-beta stocks instead of levering a low-beta portfolio, leaving high beta overpriced. This is the "betting against beta" anomaly, and it's one of the strongest empirical challenges to CAPM.
3. A negative-beta asset has expected return below \(r_f\). That's not a downward-sloping SML — the line still slopes up — but it's the same insight the Q3.5 logic produces. Gold and some hedges behave this way: investors accept low expected returns for assets that pay off in bad states, because that insurance is valuable.
Q5.4 — Key CAPM assumptions
On investors:
- Investors are rational and mean-variance optimizers. They care only about expected return and variance, preferring more return and less variance. This implicitly requires either normally distributed returns or quadratic utility.
- Homogeneous expectations. All investors share identical estimates of expected returns, variances, and covariances. Without this they'd compute different efficient frontiers and hold different tangency portfolios, and Q5.5 collapses.
- Single-period horizon. Everyone optimizes over the same one period.
- Investors are price takers. No one is large enough to move prices.
On markets:
- Unlimited borrowing and lending at the risk-free rate. This is what makes the Capital Market Line a straight line and lets investors reach any point on it. It's the assumption that fails most visibly in practice and generates the flat-SML anomaly in Q5.3.
- No taxes or transaction costs. Otherwise optimal portfolios differ by tax status and rebalancing is costly.
- All assets are tradable and infinitely divisible, including the ability to short without restriction.
- Information is costless and available to all, which supports homogeneous expectations.
Which ones actually matter. Homogeneous expectations and unlimited risk-free borrowing do the heavy lifting — together they force every investor to hold the same tangency portfolio and generate the whole beta-return relationship. The frictionless-market assumptions are more cosmetic; relaxing them modifies CAPM without breaking it. Richard Roll's critique adds a deeper problem: the true market portfolio includes all risky assets and is unobservable, so CAPM is arguably untestable — any test is really a joint test of CAPM and of whether your proxy index is the market portfolio.
Q5.5 — Why all investors hold the same tangency portfolio
Because with a risk-free asset available, risk preference determines only how much to allocate to risky assets, not which risky assets to hold. That's the Two-Fund Separation Theorem.
The argument:
- With homogeneous expectations, every investor computes the same efficient frontier — they share identical inputs, so they get identical output.
- Introducing a risk-free asset lets any investor combine it with any risky portfolio. The set of achievable combinations is a straight line from \(r_f\) through the chosen risky portfolio.
- The investor wants the steepest such line, since steeper means more expected return per unit of risk at every level. The steepest line from \(r_f\) is the one tangent to the efficient frontier — the Capital Market Line, touching at the tangency portfolio.
- The tangency portfolio maximizes the Sharpe ratio, and the Sharpe-maximizing portfolio doesn't depend on risk aversion at all. It's determined purely by the return and covariance inputs, which everyone shares.
So the decision separates into two independent steps:
- Which risky portfolio? The tangency portfolio. Same for everyone.
- How much of it? This is where risk tolerance enters. Conservative investors hold mostly the risk-free asset with a slice of the tangency portfolio. Aggressive investors borrow at \(r_f\) and hold more than 100% in it. Both sit on the same CML, just at different points.
The key insight: a conservative investor does not respond by picking a safer mix of risky assets. That would put them on an inferior line with a worse Sharpe ratio. They hold the same optimal risky portfolio and dial down exposure with cash. Adjusting the blend of risky assets to reduce risk is strictly worse than adjusting the cash weight.
Your Q2.3 result illustrates it concretely: the max-Sharpe mix is 30% ABC / 70% XYZ with Sharpe 0.335. Every investor should hold that mix, then scale it with borrowing or lending. Choosing 60/40 to reduce volatility gives Sharpe 0.302 — a worse risk-return tradeoff than simply holding less of the 30/70 portfolio alongside cash.
Q5.6 — Why the tangency portfolio must be the market portfolio
The mechanism is market clearing: every asset must be held by someone, and prices adjust until that's true.
Step by step:
- From Q5.5, every investor holds the identical tangency portfolio of risky assets, differing only in how much.
- Therefore the aggregate of all investors' risky holdings has the same composition as the tangency portfolio — summing scaled copies of one portfolio preserves its weights.
- But the aggregate of all investors' holdings is, by definition, the total supply of all risky assets outstanding — which is the market portfolio, value-weighted.
- Hence tangency portfolio = market portfolio.
The price adjustment that enforces it — the actual mechanism the question asks for:
Suppose the tangency portfolio didn't match market weights. Say it called for 5% in some stock while that stock is 2% of total market value. Then collective demand exceeds supply, and excess demand pushes the price up. A higher price means lower expected future return for the same cash flows, which reduces the stock's attractiveness and lowers its weight in the recomputed tangency portfolio. Simultaneously, the higher price raises that stock's market capitalization and therefore its weight in the market portfolio.
The two weights converge from opposite directions. The process continues until demand equals supply for every asset — at which point the tangency portfolio's weights are exactly market-cap weights.
The implications that follow:
- Market-cap weighting is the theoretically optimal portfolio, not merely a convenient default. This is the intellectual foundation of index investing.
- Every asset's expected return is pinned to its covariance with the market portfolio, which is what produces the SML.
- The market portfolio must include all risky assets, not just stocks — Roll's critique again, and the reason CAPM is cleaner in theory than in testing.
Q5.7 — The three forms of the Efficient Market Hypothesis
Fama's taxonomy distinguishes the levels by what information set is already reflected in prices.
1. Weak form. Prices fully reflect all past price and volume data. Future price movements cannot be predicted from historical price patterns — prices follow a random walk.
Implication: technical analysis cannot generate excess returns. Charts, trends, momentum indicators and support/resistance levels have no predictive power, because any exploitable pattern in past prices would be arbitraged away immediately.
Evidence: broadly supported, though the documented momentum anomaly — winners continuing to outperform over 3–12 month horizons — is a persistent challenge.
2. Semi-strong form. Prices reflect all publicly available information — past prices plus financial statements, earnings announcements, news, analyst reports, economic data. Prices adjust to new public information essentially instantaneously.
Implication: neither technical nor fundamental analysis can generate excess returns. By the time you read a 10-K, the information is in the price. Only genuinely new information moves prices, and news is by definition unpredictable.
Evidence: the most-debated form, and the practically important one. Event studies consistently show rapid price adjustment to announcements, supporting it. Against it: documented anomalies like value and size effects, post-earnings-announcement drift, and the existence of a few investors with long records that are hard to attribute to luck.
3. Strong form. Prices reflect all information, public and private, including insider information.
Implication: no one can earn excess returns, not even corporate insiders.
Evidence: almost certainly false, and essentially nobody defends it. Insiders demonstrably earn abnormal returns on their own trades — which is precisely why insider trading is illegal. The existence of enforcement is itself evidence against the strong form.
The nesting structure, worth stating: each form contains the previous one. Strong implies semi-strong implies weak. Evidence against weak form would doom all three; evidence against strong form says nothing about the other two.
The practical takeaway: the consensus position is that markets are approximately semi-strong efficient — not perfectly, but efficiently enough that after fees and taxes, the large majority of active managers underperform index funds. Grossman and Stiglitz's paradox is worth a closing line: markets cannot be perfectly efficient, because if prices already reflected everything, nobody would be paid to gather information, and then prices wouldn't reflect it. Some inefficiency must persist to compensate the analysts whose work creates efficiency.
Q6 — Incorporation
Q6.1 — The wingsuit flying school
(i) Is a sole proprietorship a good choice?
No. It's close to the worst possible structure for this specific business.
The decisive issue is unlimited personal liability. A sole proprietorship is not a separate legal entity — the owner and the business are legally the same person. Every business debt and every legal judgment reaches the owner's personal assets: house, savings, car, future wages.
Why that's disqualifying here specifically: a wingsuit school is about as high-liability as a business gets. Students will be jumping off cliffs and out of aircraft. Serious injury and death are foreseeable outcomes, not remote tail risks. A single wrongful-death suit could produce a judgment of millions of dollars, and in a sole proprietorship that judgment attaches directly to your friend personally. He could be bankrupted for life by one accident.
Secondary problems:
- No ability to raise capital. A sole proprietorship can't issue equity. Growth is limited to personal funds and personal loans.
- No continuity. The business legally terminates on the owner's death or incapacity — a pointed concern in this industry.
- Credibility. Instructors, insurers, and landowners leasing jump sites take a formal entity more seriously.
- It changes the moment you join. A sole proprietorship has exactly one owner by definition. If you come in as a partner, it automatically becomes a general partnership — which is worse, because general partners carry unlimited joint and several liability. You would be personally liable for your friend's negligence as an instructor, and a plaintiff could collect the entire judgment from whichever partner has more assets. That's you.
The only arguments for it: it's free, requires no filing, and has the simplest possible taxes (Schedule C, pass-through). Those advantages are real but trivially outweighed here.
(ii) What structure instead?
An LLC — specifically a multi-member LLC if you join, single-member if you don't.
Why it fits:
- Limited liability. The LLC is a separate legal entity. Members' personal assets are shielded from business debts and lawsuits. A judgment against the school reaches the school's assets, not your friend's house. This is the entire reason to incorporate here.
- Pass-through taxation. Profits and losses flow to members' personal returns, avoiding the double taxation of a C-corp. A small school will likely run losses in early years, and those losses can offset other personal income.
- Operational flexibility. No board, no formal meetings, no stock classes, minimal filings. An operating agreement governs how you want it run. For a two-person business this is exactly right.
- Flexible profit allocation. Unlike an S-corp, an LLC can allocate profits disproportionately to ownership — useful if your friend does all the instructing and you mainly contributed capital.
- Tax election optionality. An LLC can elect to be taxed as an S-corp later if self-employment tax savings become meaningful.
Critical caveat worth writing down: the LLC shield is not absolute, and in this business the gaps matter.
- An instructor remains personally liable for their own negligence. The LLC protects your friend from liability for other people's acts and general business obligations, but if he personally makes an instructional error that injures a student, he can be sued individually. The entity doesn't license negligence.
- Courts can pierce the veil if formalities are ignored — commingling personal and business funds, undercapitalizing the entity, treating it as an alter ego.
- Personal guarantees on leases or loans bypass the shield entirely.
So the entity is necessary but not sufficient. The complete answer: form an LLC and carry substantial commercial general liability and professional liability insurance, and use well-drafted waivers and assumption-of-risk agreements signed by every student. Insurance is arguably the more important protection — the LLC protects personal assets, but insurance protects the business from being wiped out by the first claim. In practice, most liability-heavy small businesses rely on this layered combination, and the LLC is the cheapest layer.
Q6.2 — Skydrop AI
(i) Do S-corps have favorable tax treatment?
Yes, but with meaningful qualifications — and "favorable" depends entirely on circumstances.
The real advantages:
1. No double taxation. An S-corp is a pass-through: income is taxed once, at the shareholder level. A C-corp pays corporate tax (21% federally) and shareholders pay again on dividends or capital gains — an effective combined rate that can approach 40%+. For a profitable company distributing earnings, this is a genuine and substantial saving over a C-corp.
2. Self-employment tax savings — the most distinctive benefit. This is what separates an S-corp from an LLC taxed as a partnership. An owner-employee pays themselves a "reasonable salary" subject to payroll taxes (15.3% on the Social Security wage base, 2.9%+ Medicare above it), and takes remaining profits as distributions not subject to self-employment tax. In an LLC taxed as a partnership, essentially all active income is subject to SE tax. On meaningful profits this saves real money.
3. Pass-through losses offset personal income, subject to basis and at-risk limitations — valuable in early years.
4. The 20% QBI deduction (Section 199A) may apply, reducing the effective rate further.
The qualifications that make "favorable" conditional:
- The salary must be "reasonable." The IRS actively audits S-corps that pay token salaries to maximize distributions. The saving is bounded.
- It's not automatically better than a C-corp. For a company reinvesting all earnings rather than distributing them, the 21% corporate rate can beat individual rates up to 37%. Double taxation only bites when money comes out.
- It loses to a C-corp on QSBS. Section 1202 Qualified Small Business Stock allows exclusion of up to $10M (or 10× basis) of capital gains on a qualifying C-corp stock sale held five years. For a startup with a plausible large exit, this single provision can dwarf every S-corp tax advantage combined. S-corp stock doesn't qualify.
- Pass-through income is taxed whether or not distributed. Shareholders can owe tax on income they never received — a real cash-flow problem for a company reinvesting everything, which is exactly what a growth startup does.
Bottom line: S-corps are genuinely advantageous for profitable small businesses with few owners that distribute earnings — consulting firms, local service businesses, professional practices. That is not Skydrop AI.
(ii) Is an S-corp right given rapid growth and VC funding?
No. An S-corp is fundamentally incompatible with venture financing, and the obstacles are structural, not matters of preference.
1. Only one class of stock is permitted. This is the fatal problem. VCs invest almost universally through convertible preferred stock, carrying liquidation preferences, anti-dilution protection, dividend rights, and protective provisions. An S-corp cannot issue preferred stock at all without immediately terminating its S election. There is no workaround. No institutional VC will restructure their entire investment framework around your entity choice.
2. A 100-shareholder cap. Fine at founding, but stock option pools, employee exercises, angel rounds and SAFEs converting all consume shareholder slots. A growth company blows through this.
3. Shareholders must be U.S. individuals, certain trusts, or estates. Partnerships and corporations cannot hold S-corp stock — and essentially every VC fund is structured as a limited partnership or LLC. A VC fund legally cannot invest in an S-corp. Non-resident aliens are also barred, excluding foreign investors and complicating hiring international founders or employees who take equity.
4. Pass-through income creates phantom tax liability. Shareholders owe tax on their share of profits regardless of distributions. A rapidly growing company reinvests everything, so investors would face tax bills on income they never see. VCs will not accept this.
5. No QSBS benefit. Section 1202 can exclude up to $10M or 10× basis in gains for C-corp stock held five years. For founders and early employees of a company with a serious exit, this is often the single largest financial consideration in entity choice — and it's unavailable to S-corps.
6. An S-corp election terminates automatically the moment any requirement is violated, potentially with retroactive tax consequences and the need for IRS relief. Having it break mid-financing is a genuine deal risk.
(iii) What structure instead?
A Delaware C-corporation. This is the near-universal standard for venture-backed startups, and the reasons are specific:
Why a C-corp:
- Multiple stock classes. Common stock for founders and employees, preferred for investors — with whatever rights each round negotiates. This is the structural requirement VCs cannot waive.
- Unlimited shareholders of any type. Funds, corporations, foreign investors, option holders — no restrictions.
- QSBS eligibility under Section 1202. Up to $10M or 10× basis in gains excluded after a five-year hold. For founders this is frequently the most valuable single feature.
- No phantom income. Corporate earnings are taxed at the entity level. Shareholders owe nothing until they receive distributions or sell — appropriate for a company reinvesting everything.
- Clean equity compensation. ISOs can only be issued by corporations. Option pools, vesting schedules, 409A valuations and early exercise all have well-established mechanics.
- It's what acquirers and IPO underwriters expect. Converting at exit is costly and creates friction precisely when you don't want it.
Why Delaware specifically:
- The Delaware General Corporation Law is the most developed and frequently amended corporate statute in the U.S.
- The Court of Chancery is a specialized business court with no juries, judges expert in corporate law, and over a century of precedent. Legal outcomes are more predictable than anywhere else.
- Investor familiarity. VCs, their counsel, and standard documents (NVCA model forms) all assume Delaware. Choosing elsewhere creates diligence friction and legal cost for no benefit.
- Flexible, management-friendly provisions for authorizing shares, indemnifying directors, and structuring boards.
The honest tradeoffs, worth one line to show balance: a C-corp faces double taxation if it ever distributes profits (irrelevant for a startup, which won't), early losses are trapped at the entity level as NOL carryforwards rather than flowing to founders' personal returns, and Delaware incorporation means franchise taxes plus registered agent fees plus foreign qualification in your operating state — a few thousand dollars a year. For a company raising venture capital these costs are rounding errors against the structural benefits.
The clean summary line: entity choice should match the capital strategy. The wingsuit school is a small, closely held, liability-heavy business that will never raise institutional money — LLC. Skydrop AI is a venture-scale company that will raise multiple priced rounds from institutional funds — Delaware C-corp. The co-founder's instinct about S-corp tax treatment isn't wrong in general; it's wrong for this company.
Q7 — Capital Markets
Q7.3 — Primary vs. secondary market
Primary market: where securities are created and sold by the issuer for the first time. Capital flows to the company.
- Stocks: IPOs, follow-on offerings, direct listings, private placements. The company issues new shares and receives the proceeds.
- Bonds: new issues sold to investors, with proceeds going to the issuing company or government.
- Typically intermediated by investment banks acting as underwriters, who price the offering, assume placement risk, and distribute to investors.
- This is the only point at which the issuer raises capital.
Secondary market: where existing securities trade between investors. Capital flows between investors; the company receives nothing.
- Stocks: NYSE, Nasdaq, and other exchanges, plus OTC markets.
- Bonds: mostly over-the-counter dealer networks rather than exchanges.
- When you buy Apple shares in your brokerage account, you're buying from another investor. Apple gets none of it.
| Primary | Secondary | |
|---|---|---|
| Seller | The issuer | An existing holder |
| Who gets the money | The company | The selling investor |
| Frequency | Episodic | Continuous |
| Pricing | Negotiated/book-built | Set by supply and demand |
| Volume | Small | Vastly larger |
Why the secondary market matters even though the issuer gets nothing — this is the part worth developing:
- It makes the primary market possible. Nobody would buy an IPO if they could never sell. The ability to exit is what makes investors willing to enter, so secondary liquidity directly lowers the cost of capital in the primary market.
- It provides price discovery. The secondary price tells the company what the market thinks it's worth, informing future issuance, M&A, and compensation.
- It enables capital allocation. Investors can continuously reallocate toward better opportunities without the issuer's involvement.
- It establishes the benchmark for new issues. A follow-on offering is priced off the prevailing market price.
Q7.4 — Buy side and sell side
The buy side manages capital and purchases securities for investment. These are institutions investing their own or their clients' money, seeking returns.
Examples:
- Mutual funds and ETF managers (Fidelity, Vanguard, BlackRock)
- Hedge funds (Bridgewater, Citadel, Millennium)
- Pension funds (CalPERS, CalSTRS)
- Private equity and venture capital firms (Blackstone, KKR, Sequoia, Andreessen Horowitz)
- Endowments and sovereign wealth funds (Yale's endowment, Norway's Government Pension Fund)
- Insurance company investment arms
The sell side creates, markets, and sells securities and services to the buy side. They are intermediaries earning fees, commissions, and spreads rather than investment returns.
Examples:
- Investment banks — underwriting, M&A advisory, capital raising (Goldman Sachs, Morgan Stanley, J.P. Morgan)
- Broker-dealers — executing trades and providing market access
- Market makers — quoting two-sided prices and providing liquidity (Citadel Securities, Virtu, Jane Street)
- Sell-side research analysts — publishing company reports and ratings for institutional clients
- Sales and trading desks at banks
- Credit rating agencies, arguably (Moody's, S&P) — serving issuers rather than investors
The distinction that matters, worth stating:
| Buy side | Sell side | |
|---|---|---|
| Revenue | Management fees + performance fees; returns on capital | Commissions, spreads, underwriting and advisory fees |
| Research | Proprietary, kept internal for trading advantage | Published and distributed to win client business |
| Success metric | Investment performance vs. benchmark | Transaction volume and fee generation |
| Relationship to risk | Takes principal risk | Primarily intermediates (though market makers take inventory risk) |
One nuance worth adding: the line isn't perfectly clean. Banks run proprietary trading and principal investments that look buy-side, and market makers take genuine principal risk. The useful distinction is who bears investment risk for return versus who intermediates for fees.
On research in particular: sell-side analysts publish freely because research is a marketing tool — it generates trading commissions and investment banking relationships. This creates a structural conflict, since an analyst covering a company their bank wants as a client faces pressure toward optimism. Buy-side research stays private precisely because its value lies in others not having it.
Q7.5 — Penny stocks
Penny stocks are low-priced, small-capitalization shares — conventionally under $5 per share, though the term often implies well under $1. The SEC's working definition is a security under $5 issued by a very small company.
Are they traded on exchanges? Mostly no.
The great majority trade over-the-counter, through:
- OTC Markets Group tiers: OTCQX (best-disclosure), OTCQB (venture stage), and Pink Open Market (minimal or no disclosure requirements)
- The "grey market" for securities with no market maker quoting them
Why they're generally not on major exchanges:
- They fail listing requirements. NYSE and Nasdaq impose minimum standards: share price (typically a $1 or $4 minimum), market capitalization, shareholders' equity, public float, number of round-lot holders, and corporate governance requirements like independent directors and audit committees. Penny stock issuers typically fail several.
- Continued-listing standards force them off. A listed company whose price stays below $1 for 30 consecutive trading days receives a deficiency notice and, absent a reverse split or recovery, is delisted to the OTC market. This is the most common path onto the OTC markets — fallen companies rather than new ones.
- Reporting burden. Exchange listing requires full SEC reporting (10-K, 10-Q, 8-K) with audited financials. Many penny stock issuers are shells or near-shells that can't bear the cost, or don't want the scrutiny.
- Liquidity and market-maker economics. Exchanges need sufficient trading volume to function. Penny stocks often trade sporadically.
Why this matters — the risks, which is what the question is really probing:
- Minimal disclosure. Pink Market issuers may file nothing at all. You may have no reliable financial information.
- Extreme illiquidity and wide spreads. A bid-ask spread of 10–30% is common, so you lose a large fraction immediately on a round trip, and large positions can be impossible to exit.
- Vulnerability to manipulation. Thin float plus low price plus no disclosure is the ideal setup for pump-and-dump schemes. Promoters accumulate shares, generate hype, sell into the retail buying, and the price collapses. This is common enough that the SEC maintains dedicated penny stock rules (Rule 15g-2 through 15g-9) requiring brokers to provide risk disclosures and determine suitability before selling them to customers.
- High failure rate. Many issuers are pre-revenue, distressed, or shells.
The nuance worth noting: a low share price alone doesn't make a stock dangerous — price depends on share count, and a company could do a split to put itself in penny territory. What makes penny stocks risky is the bundle that typically accompanies the low price: tiny market cap, no disclosure, no liquidity, no analyst coverage. Also note that OTCQX companies can be legitimate — including large foreign companies trading as ADRs — so "OTC" and "penny stock" aren't synonyms.
Q7.6 — Clearinghouses
A clearinghouse is an intermediary that stands between the buyer and seller of a trade, guaranteeing that it settles even if one party defaults.
Mechanically it performs novation: the original contract between buyer and seller is replaced by two contracts, each with the clearinghouse. The clearinghouse becomes the buyer to every seller and the seller to every buyer (a central counterparty, or CCP).
Core functions:
- Counterparty risk elimination. Neither side needs to assess the other's creditworthiness — both face the clearinghouse. This is the fundamental purpose, and it's what allows anonymous trading on exchanges.
- Netting. Rather than settling every trade individually, the clearinghouse nets each member's obligations to a single net position per security per day. This cuts settlement volume dramatically — often by 95%+ — and reduces operational risk and capital requirements.
- Margin collection. Members post initial margin (collateral against potential future loss) and variation margin (daily mark-to-market settlement of gains and losses). Daily marking prevents losses from accumulating unnoticed.
- Settlement management. Coordinates the actual exchange of securities for cash — delivery versus payment — ensuring neither side delivers without receiving.
- A default waterfall. If a member fails, losses are absorbed in sequence: the defaulter's margin, then its guaranty fund contribution, then the clearinghouse's own capital ("skin in the game"), then the mutualized guaranty fund of surviving members.
Examples: the DTCC (and its subsidiary NSCC) clears U.S. equities; the OCC clears listed options; CME Clearing and ICE Clear handle futures; LCH clears a large share of interest-rate swaps.
Why they matter: without a clearinghouse, every trade would require evaluating the counterparty's credit, which would make anonymous exchange trading impossible and drastically reduce liquidity. The 2008 crisis is the counterexample that proved the point — the OTC derivatives market was largely _un_cleared and bilateral, so Lehman's failure propagated unpredictably through a web of bilateral exposures nobody could map. Dodd-Frank responded by mandating central clearing for standardized OTC derivatives, precisely to convert that opaque web into a hub-and-spoke structure with visible, margined exposures.
The concern worth mentioning: clearinghouses concentrate risk rather than eliminating it. A CCP is by construction systemically critical and "too big to fail," so its risk management and default waterfall are now themselves a focus of regulatory attention.
Q7.7 — The block trade
(i) Why would she accept the lower price?
Because the $80.46 screen price is only available for a small number of shares. Selling $100M — roughly 1.24 million shares — into the open market would push the price down far below $80.35.
The reasoning:
1. The quoted price has limited depth. The NYSE price of $80.46 is the last trade or best bid for a normal-sized order, typically a few hundred to a few thousand shares. It is not a price at which 1.24 million shares can be sold. The order book thins quickly beyond the top of book.
2. Market impact would exceed the 11-cent concession. Working a 1.24M-share order into the open market walks down the book: you exhaust bids at $80.46, then $80.44, $80.42, and so on. The average realized price could easily be well below $80.35, and the cumulative loss would exceed the $136,500 she's giving up by accepting Morgan Stanley's bid.
3. Information leakage and front-running. A large order visible in the market signals that a big seller is present. Other participants pull their bids or trade ahead of the remaining flow, pushing the price further against her. The longer the order is exposed, the worse this gets. A block trade executes immediately and privately, revealing nothing until after it's done.
4. Certainty and immediacy. Morgan Stanley's bid is firm for the entire $100M right now. Working the order over hours or days exposes her to market risk — if Exxon falls 2% during execution, she loses $2M, which dwarfs the concession. She is buying immediacy and certainty, and $0.11 per share is the price.
The clean framing: the 11-cent discount is not a worse price — it's the cost of liquidity for size. She's paying a liquidity provider to absorb a position the open market can't take at once. Measured against the realistic alternative, she likely got a good execution.
(ii) Is Morgan Stanley guaranteed to make money?
No. Morgan Stanley is taking genuine principal risk, and can lose substantially.
What they're hoping for: buy at $80.35, then distribute the 1.24M shares to their own institutional clients at or near \(80.46, capturing roughly 11 cents per share — about **\)136,500** — for providing immediacy. That's the business model: earn the spread for warehousing a position the market can't absorb instantly.
Why it isn't guaranteed:
-
They now own \(100M of Exxon and bear full market risk.** If adverse news hits — an oil price drop, an earnings warning, a sector selloff — before they've distributed, the loss can be far larger than the spread. A 1% move against them is **\)1M, roughly 7× the expected profit. The risk-reward is highly asymmetric, which is why block desks hedge aggressively and distribute fast.
-
They may not find buyers at $80.46. If institutional demand is thin, they have to cut the price to move inventory. Selling at $80.20 turns a $136,500 gain into a $186,000 loss.
-
Their own distribution moves the price. Selling 1.24M shares creates the same market impact problem the hedge fund was avoiding. Morgan Stanley just thinks they can manage it better — through client relationships, patience, and hedging.
-
Adverse selection — the most important risk. The PM may be selling because she knows something. If she's exiting ahead of bad news, or is the first of many sellers, Morgan Stanley has bought from a better-informed counterparty. This is the structural hazard every market maker faces, and it's why block bids are discounted more for names with pending news or unusual activity.
-
Hedging is imperfect and costly. They can short an energy ETF or use futures to neutralize market and sector exposure, but that leaves residual idiosyncratic risk in Exxon specifically and costs money to maintain.
The correct framing: the 11-cent spread is compensation for bearing risk, not a guaranteed fee. Morgan Stanley expects to profit on average across many such trades — the law of large numbers works in their favor — but any individual block can lose money, and large blocks occasionally lose a great deal. That's exactly the market-making economics from HW4 Q3.2: earning a spread while managing inventory and adverse selection.
On submission: the file is EE156_HW6.xlsx, one workbook with both tabs, which satisfies the "do not submit multiple separate .xlsx files" requirement. Zip it with your solutions PDF.
Two things to note for your write-up: Q7 is numbered 7.3–7.7 in the PDF with no 7.1 or 7.2 — match their numbering so the Gradescope page mapping lines up. And there's no self-grade page in this one, unlike HW5.