Emerging Topics
The specific readings rotate every cycle. The analytical frame does not — and the frame is what the essay rubric rewards. Learn the frame first, then hang each year's readings on it.
The five-part essay frame
For any emerging asset class, instrument or technique, a complete answer moves through five stages in this order. Practise until it is automatic; under time pressure it is the difference between a rambling half-answer and full marks.
Two habits raise scores immediately. First, always compare: "relative to public high yield, direct lending offers…". A feature list scores badly; a comparison with a consequence scores well. Second, always name the investor: the same instrument is appropriate for a sovereign savings fund and inappropriate for a P&C insurer, and saying why demonstrates the whole curriculum in one sentence.
Digital assets & tokenization
Valuation without cash flows. A native digital asset produces no contractual cash flow, so discounted cash flow analysis has nothing to discount. The curriculum's substitutes: network value arguments (Metcalfe-style, value rising with the square of participants), cost of production floors (mining economics for proof-of-work assets), store-of-value substitution (a share of the monetary premium currently held in gold or reserve currencies), and protocol cash flows where a token has an actual claim on fees. Each is an analogy, not a model — and you should say so.
Tokenization is a wrapper change, not a risk change. Putting a building, a fund interest or a private loan on a distributed ledger alters fractionalisation, transferability, settlement speed and record-keeping. It does not alter the underlying asset's cash flow risk, its appraisal lag, or the fact that a thin secondary market can still gap. The exam's favourite trap: tokenization creates tradability, which is not the same as liquidity — liquidity requires willing counterparties at scale, and a tokenised illiquid asset is still illiquid.
Where the real risk sits. Operational, not market: private-key custody and loss, exchange and stablecoin counterparty failure, smart-contract code risk, chain forks and settlement finality, plus an unsettled and jurisdictionally fragmented regulatory perimeter. For an institution, the ODD question is who holds the keys and under what controls — qualified custody, multi-signature or MPC arrangements, proof of reserves, and insurance.
Climate risk & ESG integration
Two risk channels, kept separate. Physical risk is damage from the climate itself, split into acute (hurricanes, floods, wildfire) and chronic (sea-level rise, heat stress, water scarcity). Transition risk is the cost of moving to a low-carbon economy: policy and carbon pricing, technology displacement, shifting consumer preference, litigation, and reputational damage. They trade off — an aggressive transition raises transition risk and lowers long-run physical risk. Confusing the two is a reliable MCQ.
Three distinct practices, often conflated. Integration treats ESG factors as financially material inputs to valuation and risk — no values judgement, no return sacrifice claimed. Screening (negative, positive, norms-based) applies constraints to the universe and therefore mathematically cannot improve the unconstrained efficient frontier. Impact requires intentionality, measurability, and — the hard one — additionality: the outcome would not have occurred without the investment. Buying a listed green stock on the secondary market usually has no additionality; providing primary capital to a project that could not otherwise be financed does.
Why ESG scores disagree. Vendors differ in scope (what is included), measurement (what proxies are used), and weighting (how pillars are aggregated). Ratings are backward-looking, disclosure-dependent, and systematically favour large firms with reporting resources. Correlations between major providers' ESG scores are far lower than between credit rating agencies'. Consequence for the exam: an ESG-tilted portfolio's exposures depend as much on the data provider as on the investment thesis.
Alternatives-specific angles. Private markets have control and time horizon — a GP can actually change a portfolio company's emissions trajectory, which is a genuine impact case that public equity struggles to make. Against that: disclosure is voluntary and sparse, the denominator (financed emissions) is hard to compute, and greenwashing risk is elevated where no auditor checks the claim. Real assets carry the sharpest physical exposure, embedded in insurance cost and eventually in cap rates.
AI & alternative data
The statistical problem. Machine learning is powerful where data is abundant and the underlying process is stable. Financial return data is neither: samples are short, the signal-to-noise ratio is brutally low, and the data-generating process shifts with regime and with the crowding of the strategy itself. This makes overfitting and data-snooping bias the central risks — with enough model specifications, an impressive backtest is guaranteed. Defences: walk-forward and out-of-sample testing, cross-validation that respects time ordering, limiting the number of trials, and a required economic rationale for any signal.
Prediction is not causation. A model can predict well in-sample and still fail out-of-sample because it learned a spurious correlation. For an allocator, the question "why should this relationship persist?" precedes "how good is the backtest?".
Alternative data. Satellite imagery, card transactions, web traffic, geolocation, natural-language processing of filings and calls. Practical issues the exam raises: short histories that don't span a full cycle, survivorship in the data vendor's panel, changes in panel composition that masquerade as signal, privacy and legality (personal data, material non-public information), and rapid capacity decay — once a dataset is widely licensed, its alpha compresses.
Fiduciary and governance angle. A trustee must be able to explain the process. Black-box models create explainability and accountability problems, complicate ODD (how do you diligence a model you cannot inspect?), and raise model-risk governance requirements: independent validation, version control, and defined kill criteria.
Private-market democratization
The structural push to bring private assets to wealth and retail channels through evergreen funds, interval funds, tender-offer funds, non-traded REITs and BDCs, and registered feeder structures. The economics for managers are obvious — a large, sticky, fee-tolerant pool of capital. The analytical issue is one thing: liquidity mismatch.
Periodic redemption (often quarterly, commonly capped at ~5% of NAV), a published NAV, low minimums, no capital calls, simplified tax reporting.
Appraised, lagged valuations; multi-year holding periods; exit only via sale or secondary at a discount. The mismatch is structural, not a failure of management.
Consequences to be able to state. Redemption caps, queues and gates are the designed release valve, and they bind precisely when investors most want out — so the vehicle is least liquid when liquidity is most valued. The cash and liquid-credit buffer held to meet redemptions creates permanent cash drag on returns. Because NAV is struck off appraisals that lag public markets, redeeming investors can transfer value from remaining holders (or vice versa) — a genuine fairness problem, and the reason valuation policy and NAV-strike timing dominate diligence on these vehicles. Finally, fee load is often higher than institutional share classes once distribution and servicing fees are included.
The balanced answer notes the real benefits too: access to a return stream previously reserved for institutions, diversification for portfolios otherwise entirely in public markets, and — for a genuinely long-horizon individual — an illiquidity premium they are well placed to earn. The judgement call is whether the wrapper's cost and the mismatch leave enough of that premium for the end investor.
Private credit & secondaries
Private credit. Directly originated, mostly senior secured, mostly floating-rate loans to middle-market companies, typically sponsor-backed. Decompose the yield: base rate + credit spread + illiquidity premium + complexity/origination premium + OID and fees. Because the coupon floats, the exposure is credit, not duration — rising rates help income but stress borrower coverage ratios, which is the point of the cycle where the asset class is genuinely tested. Watch: covenant erosion, PIK income substituting for cash interest (a warning sign, since it defers the cash test), rising loan-to-value at entry, valuation discretion on unquoted loans, and correlation with the sponsor's own equity outcome.
Secondaries. Two families, and the exam wants the distinction.
Secondaries pricing is quoted as a percentage of reference NAV, but the reference NAV is typically a quarter or two stale, so a "discount" during a falling market may be no discount at all against true current value. This lag effect is a favourite computational trap.
Insurance-linked & niche real assets
The cleanest example of a genuine diversifier, and the best place to rehearse the five-part frame end to end.
Definition. Insurance-linked securities transfer insurance risk to capital markets. Catastrophe bonds pay a spread over collateral yield and forgive principal if a defined trigger is breached — indemnity (the sponsor's actual losses), industry loss (a market-wide index), or parametric (a measured physical event such as wind speed at a location). Collateralised reinsurance, industry loss warranties and quota-share sidecars are private variants of the same trade.
Economic exposure. The risk driver is meteorological and seismic, not financial, so unconditional correlation with equities and credit is close to zero and remains low in financial crises. That is the whole investment case. Basis risk is the price of the cleaner triggers: parametric structures settle fast but can pay nothing while the sponsor suffers a large loss.
Valuation. Expected loss comes from a vendor catastrophe model, so the multiple of expected loss is the pricing convention. Model risk lives in the hazard and vulnerability modules, not in a discount rate — and the models are recalibrated after every major event, which repriced the market repeatedly through the 2017–2022 loss years.
Risks and frictions. Returns are strongly negatively skewed: many small coupon-like gains, rare severe principal losses. Loss creep and development on indemnity triggers delays final settlement; trapped collateral locks capital past maturity; climate change makes the historical hazard record non-stationary; and the market is seasonal, concentrated in US wind exposure.
Portfolio fit. A diversifying sleeve for an owner with a long horizon and the analytical capacity to interrogate a catastrophe model — sized small, diversified across peril and region, and never sold as a bond substitute. Adjacent niche real assets (timber, farmland, water rights, royalties) rehearse the same argument: a biological or contractual driver, an appraisal-based NAV, and thin comparables.
Confusion pairs
Cover the right column and reproduce the distinction aloud.
Practice
Six multiple-choice questions in exam style, with the reasoning — not just the letter.
1. An allocator argues that tokenizing a portfolio of middle-market loans will make the position liquid. The most accurate critique is that tokenization:
2. A manager buys shares of a listed renewable-energy utility on the secondary market and markets the fund as impact investing. The weakest element of that claim is:
3. A quantitative team reports a Sharpe ratio of 3.1 from a machine-learning signal tested across 4,000 model specifications on eight years of daily data. The primary concern is:
4. An evergreen private-credit fund offers quarterly redemptions capped at 5% of NAV and holds a 10% liquid buffer. Which statement best describes the structural consequence?
5. In a GP-led continuation vehicle, the central governance concern is that the general partner:
6. A secondaries buyer purchases an LP interest at "92% of NAV" using a reference NAV struck two quarters earlier, during which comparable public markets fell 15%. The buyer has most likely:
Constructed-response practice
Write these under time. Each outline is the shape the rubric rewards, not a model answer to memorise.
Prompt A (12 minutes). A public pension plan with a 78% funded ratio is considering a 5% allocation to private credit, funded from public high yield. Recommend for or against, and justify with reference to the plan's liability profile, the exposures the allocation adds, and two risks the investment committee should monitor.
- State the liability profile first — long, inflation-sensitive, bond-like; the plan is underfunded so return-seeking is required but surplus volatility is the constraint.
- Identify what actually changes: this is a substitution within credit, not an addition of a new risk factor, so the marginal exposure is illiquidity + complexity + seniority, less duration (floating rate).
- Quantify the trade honestly: yield pick-up vs loss of daily liquidity and mark-to-market transparency.
- Two monitored risks, chosen and justified: borrower interest-coverage deterioration in a higher-rate regime, and valuation discretion on unquoted loans (with the governance mitigant).
- Close with a recommendation and the condition attached to it (pacing, manager selection, liquidity budget headroom).
Prompt B (10 minutes). An investment committee asks why two ESG data providers assign the same company scores in opposite quartiles, and what that implies for a planned ESG-tilted mandate. Explain and advise.
- Three sources of divergence — scope, measurement, weighting — with one concrete example of each.
- Structural biases: disclosure-dependence favours large caps; ratings are backward-looking; some providers score risk-to-the-firm, others impact-of-the-firm.
- Implication: the mandate's realised exposures are a function of the data vendor, so vendor choice is an investment decision requiring committee approval.
- Advice: define the objective first (integration for materiality vs values-based screening vs impact), then select data to serve it; specify benchmark, tracking-error budget, and how the tilt will be attributed.