Future Capital Markets

Capital markets are the machinery standing between household savings and the firms that use them: exchanges, funds, banks, brokers, raters, index providers and the disclosure regimes that price what they all hold. The framing under test is that this machinery is being reshaped by new instruments and intermediaries — green bonds, ESG mandates, tokenised assets, digital settlement.

The evidence supports a narrower and stranger claim, and it inverts the framing's own emphasis. The instruments the framing points at are the ones with the smallest measured effects: a sovereign greenium of about 2 basis points, an ESG cost-of-capital channel calibrated at 0.44 basis points, and $16.2 billion of tokenised real-world assets. The things with measured scale are not new: private credit at about $2.1 trillion, a buyout industry sitting on 32,000 unexited companies worth $3.8 trillion, and index funds — an instrument invented in 1975 — concentrating 41% of the S&P 500's weight in ten companies and about 25% of its votes in three firms.

What has actually been reshaped is not the instrument but the venue. Ordinary leveraged lending and ordinary leveraged buyouts now happen outside the quarterly-reporting perimeter, and more than half of US equity volume trades off-exchange. The one unambiguous reallocation in the record is from disclosure to non-disclosure, and that is a change in the information available to price capital rather than a change in the technology of financing it.

This brief covers the intermediaries and instruments in general — who stands between savings and firms, what they charge, what they disclose, and whether measured returns justify the arrangement. It does not cover horizons longer than any instrument's tenor: sovereign wealth fund mandates, pension liability duration, century bonds and intergenerational discounting belong to Civilization Scale Investment. The seam is duration.

Established The best-measured change in this subject is a disappearance, and it is smaller than it is usually reported to be. Doidge, Karolyi and Stulz established that the number of publicly listed US firms peaked in 1996 at 8,025 and fell to roughly 3,500 by 2012, with listings per million inhabitants falling from 30 to 13. Their decomposition is the part that gets dropped: about 54% of the gap is missing new listings and about 46% is excess delisting — not only firms declining to float, but firms being taken off the board. Frontier The trough was not the end. An investment consultant's count puts the series at over 6,500 in 1997, 3,800 in 2012 and 4,700 as of June 2024 — a partial recovery. The direction is established; the level depends on whether one counts SEC registrants, exchange listings or domestic operating companies, and the sources use different bases. Established Mark the interest: that consultant advises institutions into private markets and concludes that more than 85% of companies above $100m of revenue are private — a finding that supports the product. The count is citable; the conclusion is advocacy.

Established The mechanism by which a public company stops being public is that somebody buys it. The World Federation of Exchanges analysed 69 exchanges across 64 jurisdictions from 2010 to 2024 and finds M&A accounts for more than 60% of delistings, voluntary go-private transactions generally below 10% a year, and involuntary delistings rising to roughly 30% in 2023 as credit tightened. That is a capital-markets fact, not a compliance-cost fact, and it is the single strongest piece of evidence against the standard regulatory-burden story.

Frontier Private equity's return premium is the most contested empirical question in the field, and it has a clean structure: two competent literatures on different vintages with different fee treatments. Harris, Jenkinson and Kaplan, on Burgiss data sourced from over 200 institutional investors — 598 buyout funds and 775 VC funds, vintages 1984–2008 — find US buyout at a public market equivalent of about 1.27 weighted-average, roughly 20–27% total outperformance over fund life or more than 3% a year against the S&P 500, with venture capital substantially outperforming in the 1990s and underperforming in the 2000s. Frontier Phalippou takes 2,132 North American funds, 2006–2015 vintages, $1.7 trillion raised, and finds net multiples of money of 1.55–1.63x, implying about 11% a year, which matches the relevant public indices; large pension funds did slightly worse at 1.51–1.54x. Against that parity he sets an estimated $230 billion of carried interest on those vintages and a rise in private-equity multibillionaires from 3 in 2005 to 22 in 2020.

Established The two results are not in contradiction; they are about different decades. The Harris–Jenkinson–Kaplan premium is concentrated in pre-2006 funds and the Phalippou parity is a post-2006 finding. The defensible synthesis is that buyout outperformed public markets materially before roughly 2006 and has performed at or near parity since, while fees stayed at pre-2006 levels. Established And an interested party has said so. Bain & Company, whose private-equity practice is a major business line, reported that over the ten years to June 2019 US private equity fund IRR was 15.3% against an S&P 500 public market equivalent of 15.5%, and wrote that parity with public markets is not what private-equity investors are paying for. Interest running against the finding raises its weight, and this is a consultancy conceding its own clients' underperformance. Handwave Bain's 2026 report still claims outperformance, but only for top-quartile funds — a claim restricted to the top quartile is a concession about the mean, and top-quartile status is not investable in advance.

Established Phalippou's methodological point is load-bearing and is not in dispute: IRR is not a rate of return, because of the reinvestment assumption embedded in the calculation. IRR is nonetheless what the industry reports. Established The liquidity problem is now the industry's own headline. Bain's Global Private Equity Report 2026, published 23 February 2026, records 32,000 unsold portfolio companies worth $3.8 trillion, $1.3 trillion of buyout dry powder, distributions to limited partners at 14% of net asset value in 2025, an average holding period at exit of about seven years against five to six in 2010–2021, and a fourth consecutive year of falling buyout fundraising at $395 billion, down 16%. Deal value rose 44% to $904 billion and exit value 47% to $717 billion on fewer transactions, the recovery concentrated in 13 megadeals of $10bn or more worth $274 billion. Frontier The structural reading: an intermediary that marks its own assets, controls the timing of exit and reports IRR has three degrees of freedom a public market does not give a manager, and the unexited backlog is where all three meet.

Established Private credit is the genuinely new intermediary, and the IMF has stated precisely what the risk is. Its April 2024 Global Financial Stability Report chapter puts the sector at roughly $2.1 trillion globally in assets plus undeployed commitments in 2023, $1.6 trillion in the US as of June 2023, growing about 20% a year over the prior five years and reaching 7% of non-financial corporate credit in North America — comparable to broadly syndicated loans and high-yield bonds. Median borrower size is $0.5 billion against $4.6 billion for leveraged loans. Over one-third of firms with private-credit-like characteristics have unsustainable interest coverage ratios, and the payment-in-kind share of business-development-company interest income has doubled since 2019. Insurers influenced by private equity hold a median 20% Level 3 assets against 6% for typical large insurers. Established The valuation finding is the one that matters for this brief: adjustment of the valuation of private credit loans is insufficient during market shocks — prices move much less than in high-yield or leveraged loans despite higher borrower risk. Speculative The Fund's verdict runs both ways and should be quoted both ways: the financial stability risks at present appear contained, but if the sector remains opaque and grows exponentially under limited prudential oversight the vulnerabilities could become systemic. The containment assessment is the Fund's finding; the systemic claim is a conditional prediction.

Established Ownership concentration is the largest measured change in the whole subject and it comes from a fifty-year-old instrument. Bebchuk and Hirst document the Big Three — BlackRock, Vanguard, State Street — holding roughly 20.5% of outstanding S&P 500 shares and casting about 25% of the votes at S&P 500 companies as of 2017, on more than $3.4 trillion of index-fund inflows over 2009–2018 with the Big Three taking about 82% of all fund inflows. Frontier Their projection of as much as 40% of S&P 500 votes within two decades is an extrapolation of a flow trend, not a measurement. Established Passive overtook active on assets at the end of 2023, on 2023 flows of +$244 billion to passive US equity funds against −$257 billion from active; the crossover dates by asset class are older than the headline — US equity flows turned passive-favouring in 2005, international equity in 2008, bonds in 2013. Established And index concentration itself: the top ten S&P 500 companies reached nearly 41% of index weight at end-2025, against roughly 19% in 1990, about 27% at the 2000 peak and about 19% at end-2015 — more than doubling in a decade, on companies producing about 32% of index earnings. That figure comes from a wealth manager, which is an interested party on concentration narratives; the weights are citable, the feedback-loop mechanism asserted around them is not.

Frontier Whether that ownership changes prices is where the field actually splits, and the split has not been resolved. Azar, Schmalz and Tecu reported anticompetitive price effects of common ownership in US airlines using the MHHI-delta measure. Dennis, Gerardi and Schenone re-ran it in the same journal and concluded that the documented positive correlation between common ownership and airline ticket prices stems from the market-share component of the common ownership measure, not the ownership and control components, and that results are sensitive to control measures and to bankruptcy-period ownership assumptions. Frontier Azar and Vives then offered a third position rather than a resolution: intra-industry common ownership raises prices while inter-industry common ownership lowers them, with the average effect positive only because some shareholders are concentrated in airlines, and antitrust needing to account for the procompetitive inter-industry effects of diversified investors. José Azar is an author of both the original finding and the revision, which raises the revision's weight. Established The honest terminal position: the ownership fact is established, the governance-influence claim is plausible and under-measured, and the price-effect claim is disputed by a direct replication in the same journal. Anyone stating the price effect as settled, in either direction, is overstating.

Frontier ESG is where the volume of claims most exceeds the strength of the evidence, and the cleanest test comes back near zero. Berk and van Binsbergen model whether divestment by socially conscious investors moves the cost of capital enough to change firm behaviour, and test it on FTSE4Good index inclusions over 2002–2021. Inclusion produced a price increase of about 24 basis points, not significantly different from zero. Their calibration of the cost-of-capital effect at current participation is 0.44 basis points in the base case, 0.79 bp on an alternative definition, and 10.6 bp even under an optimistic assumption that a third of assets under management divest. Established The measurement layer is worse. Berg, Kölbel and Rigobon compare six major ESG raters and decompose the divergence: measurement contributes 56%, scope 38% and weights 6%, with a documented rater effect in which an agency's overall view of a firm contaminates its category-level scores. If the ratings disagree mostly on measurement, empirical work regressing outcomes on “ESG” is regressing on the rater. Frontier The flagship instrument prices at almost nothing: 332 matched pairs of sovereign green and conventional bonds issued 2014–2023 give a greenium of about 2 basis points in advanced economies and 13 bp in emerging markets, with the authors concluding the fiscal impact is negligible once certification, reporting and monitoring costs are counted. Established And the flows have turned: $8.6 billion of global sustainable-fund outflows in Q1 2025, assets down to $3.16 trillion, launches down to 54 from 105 in the prior quarter, and 335 funds rebranded, of which 116 dropped ESG terms entirely and 3 added them — reported by trade press sympathetic to the sector, which raises its weight. Frontier Read the four channels together: the cost-of-capital mechanism is measured under a basis point, the ratings disagree on the component that should be objective, the flagship instrument prices at 2bp, and the flows have reversed. What ESG demonstrably reshaped is the fund-labelling industry.

Established Market structure moved, and it moved in where trades happen rather than in who owns what. Off-exchange trading in US equities topped 50% in 2024, with November 2024 the first month in which off-exchange volume exceeded on-exchange volume and December 2024 and January 2025 also above 50%; off-exchange share exceeds 45% across all market caps, all price ranges and ETFs. Frontier The source is an exchange — a party with a direct commercial interest in framing off-exchange trading as a problem — and its interpretive claim that bilateral non-ATS liquidity is inaccessible to asset managers and raises their trading costs should be read as advocacy. Its concession that retail does not account for all non-ATS prints runs against its own framing and is worth noting. Frontier On payment for order flow the only fetchable quantification here is an SEC staff working paper, explicitly not a Commission position, on Robinhood crypto token introductions. Its equities-relevant datum is comparative: wholesalers pay about 0.8 basis points per dollar in equities, 8 bp in options and 35 bp in crypto — roughly 45 times the equity rate. Its estimate that spread widening costs crypto traders $4.8–6.77 million a day is for a different asset class and should not be generalised to equities.

Frontier Tokenisation is a pilot, and the arithmetic says so. A crypto-industry research report puts stablecoin market capitalisation at $317 billion in March 2026 — Tether $184bn, USDC $79.0bn — and tokenised real-world assets at $16.2 billion, of which tokenised Treasuries are $9.00 billion, 55.3%. Established Mark the interest heavily: the publisher is part of a listed digital-asset platform and discloses that its employees including journalists may receive that platform's equity-based compensation. These are the figures to cite because they are the ones published, and the honest reading is arithmetic: $16.2 billion of tokenised real-world assets is a rounding error against $3.8 trillion of unexited private-equity portfolio companies or $2.1 trillion of private credit.

Established The most consequential single item in this brief is barely a year old and it arrived by executive fiat rather than market demand. Executive Order 14330, Democratizing Access to Alternative Assets for 401(k) Investors, signed 7 August 2025, directs the Secretary of Labor within 180 days to reexamine ERISA fiduciary guidance, consider rescinding the 21 December 2021 Supplemental Private Equity Statement, propose appropriately calibrated safe harbours and prioritise action to reduce ERISA litigation; and directs the SEC to consider facilitating access, potentially by revising accredited-investor and qualified-purchaser rules. The named asset classes are private equity and debt, real estate and real estate debt, actively managed digital asset vehicles, commodities, infrastructure and longevity risk pools. The order states that more than 90 million Americans participate in employer-sponsored defined-contribution plans. Frontier Set it against the rest of this section and the sequence is the finding: a fee load justified by a premium the post-2006 evidence does not clearly show, a $3.8tn illiquidity backlog, four consecutive years of falling institutional fundraising, an opaque credit market whose marks do not adjust in shocks — and then a policy opening the largest pool of unsophisticated capital in the world. Whether that is democratisation or distribution is exactly the question, and it is not yet answerable.

Established The tokenisation figure above measures the wrong layer, and the layer underneath it holds an order of magnitude more money. Tokenised real-world assets at $16.2 billion is an instrument count; stablecoins — tokenised claims on bank deposits and Treasury bills — are the settlement layer those instruments move on, and their outstanding stock has run in the low hundreds of billions of dollars through 2025, dominated by two issuers. Frontier The largest issuer’s reserve holdings of US Treasury bills are on the scale of a mid-sized sovereign’s, which makes a stablecoin panic a forced-sale event in the bill market rather than a crypto event. This is the first thing in the subject with the scale the framing claims, and it is not a new financing technology: it is a money-market fund with a payments interface and without the disclosure, liquidity-fee and gating machinery imposed on money-market funds after 2008 and 2016 — the same relocation out of a protection perimeter this brief finds everywhere else, arriving in the plumbing rather than in the instrument.

Established The run risk is measured rather than hypothetical, and it has been measured twice. The algorithmic stablecoin TerraUSD lost its peg in May 2022 and the arrangement went to approximately zero within days, taking tens of billions of notional with it. The fully reserved USDC depegged to about $0.87 in March 2023 on news that $3.3 billion of its reserves sat at a failed bank, and repegged when the deposit guarantee was extended. Frontier Read together the two episodes say the binding variable is the reserve asset and its access to a central bank, not the ledger technology: the design that failed permanently had no reserve, and the design that recovered did so because a public authority backstopped a bank. Distributed settlement was load-bearing in neither direction.

Established Rulemakers have now answered the reserve question and left the settlement question open. The EU markets-in-crypto-assets regime brought e-money and asset-referenced tokens inside a licensing perimeter from the end of 2024, and the US GENIUS Act, signed 18 July 2025, requires one-for-one reserves in cash and short-dated government paper, monthly reserve disclosure, and no interest paid to holders. Frontier The interest prohibition is the consequential clause and is rarely read that way: it makes issuer seigniorage the business model and caps the instrument’s appeal as a savings vehicle, which is the design choice determining whether tokenised money competes with bank deposits for funding or stays a transaction balance. Handwave Which it becomes is not yet observable, and any deposit-substitution figure quoted today is a projection.

Frontier Is the public-to-private shift a regulatory artefact? Held by much of the US securities bar, parts of the deregulatory wing of the SEC, and the private-markets industry. The timing correlates loosely with Sarbanes–Oxley and Dodd–Frank and compliance costs are real. Against it: M&A drives more than 60% of delistings, and the listing count has risen from its 2012 trough. What would settle it is a cross-jurisdictional comparison of listing counts against compliance-cost indices — the listing-gap method makes it possible and it has not been done at scale since 2015. The strongest version of this claim is not the one usually made.

Frontier Do private markets outperform because they are illiquid and patient? The industry's position, and most institutional consultants'. For: a pre-2006 public market equivalent of 1.27. Against: post-2006 parity, and a consultancy's own 15.3% against 15.5%. What would settle it: mandatory standardised public-market-equivalent reporting on a common benchmark with a common fee treatment, which no regulator has imposed and which is a rule rather than a research programme.

Frontier Or is the illiquidity premium a volatility-laundering premium? This is the most under-covered serious hypothesis in the field. The claim is that private marks are smoothed, that the reported low correlation with public markets is an artefact of stale pricing, and that investors are paying for the appearance of stability rather than for return. The IMF's finding that private credit valuations adjust insufficiently in shocks is direct supporting evidence in the adjacent asset class. What would settle it: forced mark-to-market of a private portfolio through a genuine liquidity event, or a matched comparison of transacted secondary prices against reported net asset values across a full cycle.

Frontier Is common ownership anticompetitive? Held by Azar, Schmalz, Elhauge and parts of the FTC and DG COMP; contested by a direct replication in the same journal; refined by Azar and Vives into a sign-ambiguous position where the inter-industry effect is procompetitive. What would settle it: an event study on a large forced divestiture, or on an index-composition change that moves common ownership without moving market share. Frontier A separate and weaker claim is better supported: that index funds are a new bureaucracy owning the economy without governing it — not colluding, but under-investing in stewardship because the benefits are shared with competitors. The evidence is the arithmetic of stewardship staffing against portfolio size. What would settle it is outcome measurement: do firms with higher Big Three ownership differ on governance outcomes, controlling for index inclusion?

Speculative Does passive investing degrade price discovery, and does it break past some threshold? Two claims usually merged. The strong version — that indexing produces a bubble that must unwind — is a recurring prediction with a poor forecasting record, held by a persistent minority including self-interested active managers and some serious microstructure researchers. Established The weak version is arithmetic and is not in dispute: the marginal informed trader is a shrinking share of volume, and at 41% top-ten weight, index performance is a bet on ten firms. What would settle the strong version: an identified shock to passive share with a measured effect on cross-sectional return-to-fundamental relationships.

Frontier Does ESG change the cost of capital and therefore corporate behaviour? Held by the sustainable-finance industry, many regulators and the bulk of the practitioner literature. The evidence for it is largely claimed assets under management. The evidence against is four independent measurements: 0.44 basis points, a 2bp greenium, 56%-measurement-driven rating divergence, and reversing flows. The volume-to-evidence ratio here is the highest in the brief. Speculative The fallback is a different hypothesis and should be named separately: that ESG works through engagement, mandate and reputational risk rather than through price. It is much harder to test, it has not been tested at scale, and it is routinely offered as an answer to the near-null cost-of-capital result as though it were established. It is not.

Speculative Will tokenisation re-plumb settlement and thereby reallocate capital? Held by the digital-asset industry and by central banks running wholesale settlement pilots. The evidence is real, growing and three orders of magnitude below private credit. What would settle it: a tokenised instrument achieving primary-issuance scale in a regulated market with measurable cost savings against conventional settlement. Speculative And the newest open question has no evidence at all yet: is retail access to private markets democratisation or distribution of the illiquidity backlog? The sequence — a $3.8tn unexited book, four years of falling institutional fundraising, then an order opening 90 million defined-contribution participants — is a correlation of timing, not a demonstrated intent. What would settle it: fee and net-return data on defined-contribution-channel private funds once they exist, benchmarked against the plan's default target-date fund. There is no evidence either way, because the policy is a year old. The absence is itself the finding.

Frontier Does atomic settlement reduce cost, or move it? The case for tokenised settlement is that delivery and payment execute in one indivisible step, removing the principal risk that existing systems manage with collateral and time. The unstated cost is liquidity: settling every trade gross in real time needs far more intraday funding than netting does, and netting is why the wholesale plumbing has the shape it has. The BIS proposal answers this by putting central bank reserves on the same programmable ledger as the assets, so that elasticity — the central bank’s ability to create settlement balances on demand — survives the redesign. Frontier No pilot has published the intraday liquidity cost of running its own volumes atomically. The experiments exist — a sterling system settling in central bank money through an omnibus account, tokenised correspondent-banking trials across several currencies, a cross-border project that reached a minimum viable product before its convening institution withdrew — and each publishes architecture rather than a cost line. Until one publishes the funding requirement against the netting baseline it displaces, the efficiency claim is a diagram.

Speculative The deeper question is whether the singleness of money survives tokenisation. The BIS sets three tests — singleness, meaning every form of money trades at par; elasticity, meaning settlement balances expand on demand; integrity, meaning the system resists illicit use — and argues stablecoins fail all three. Frontier That is an argument from institutional design rather than a measurement, and this brief marks it as one. The falsifiable part is narrow: if tokens issued by different institutions persistently trade away from par against each other in ordinary conditions rather than only in stress, singleness has failed in the observable sense; if they do not, the test passes at the level the data can reach. Handwave Nobody publishes a continuous cross-issuer par-deviation series, so the cheapest decisive measurement in this subject is one nobody has assembled.

Established The binding constraint in this subject is that the reported numbers are produced by the parties being measured. A private fund marks its own assets, chooses when to exit, and reports an internal rate of return that is not a rate of return. Those three degrees of freedom are not fraud; they are the reporting convention. No regulator requires a standardised public market equivalent on a common benchmark with a common fee treatment, which is why two competent literatures can reach opposite conclusions and neither can be adjudicated on the data.

Established The ESG measurement layer fails at the component that should be objective. Across six major raters, measurement contributes 56% of divergence, scope 38% and weights 6% — and there is a documented rater effect in which an agency's overall view of a firm contaminates its category scores. Frontier That is not merely inconvenient: it means the field's replication problems are downstream of a construct-validity failure, and that a regression of anything on an ESG score is a regression on the rater's house view.

Frontier The common-ownership measure carries its own bottleneck. MHHI-delta mixes an ownership component with a market-share component, and the replication attributes the entire airline price correlation to the latter. A measure that cannot separate the thing being tested from the thing being controlled for is not a measure of the thing being tested. The endogeneity objection is the same one Azar and Vives raise against their own earlier work.

Established Private-market prices do not move when public prices do, and the IMF has measured it. Private credit loan valuations adjust insufficiently during market shocks — less than high-yield or leveraged loans despite higher borrower risk. Frontier That is the technical core of the volatility-laundering hypothesis, and it means that any correlation or volatility statistic computed on private marks is computed on a smoothed series. Portfolio construction that treats a smoothed series as a low-volatility asset is buying a measurement artefact.

Established And the disclosure perimeter itself is the bottleneck this brief keeps returning to. Public firms report quarterly to a regulator. Private-credit borrowers, private-equity portfolio companies and off-exchange bilateral counterparties do not, or do so to a far smaller audience. Median private-credit borrower size is $0.5 billion against $4.6 billion for leveraged loans, so the migration is concentrated in exactly the firms least likely to have public financials in the first place. Frontier Level 3 assets at a median 20% of holdings for private-equity-influenced insurers, against 6% for typical large insurers, is the same fact appearing on a balance sheet the public does see.

Established The bottleneck in tokenised settlement is not throughput; it is the settlement asset. The international standards for financial market infrastructures already require money settlements in central bank money where practical and, where not, in an asset carrying little or no credit and liquidity risk. Every tokenised platform therefore faces a question first posed in 2012, and the available answers are a deposit token issued by a commercial bank, a stablecoin issued by a non-bank, or tokenised central bank reserves that only a central bank can create. Frontier The first two carry issuer credit risk into the settlement leg; the third exists in pilots and in no wholesale production system at scale. That is a central-bank decision rather than an engineering problem, and it has been deferred in most jurisdictions.

Frontier Fragmentation is the second bottleneck, and it is being created faster than it is being solved. Each bank consortium, market infrastructure and public pilot builds its own ledger with its own identity, privacy and finality rules, which reproduces the correspondent-banking problem the technology was meant to remove — more venues, each internally efficient, joined by bridges that are the weakest link. Speculative Retail central bank digital currency is the case where the measured record is already unkind: launched retail schemes show adoption in the low single-digit percentages of their populations, and the largest pilot reports cumulative transaction volume rather than the balance or velocity figures that would show the money being held. Handwave Weak adoption of a first-generation design says little about a later one.

Established Nothing on this map produces a result this brief waits on, and the honest statement of the dependency is that it is a rule rather than a discovery. The three things that would move the field furthest are all things a regulator could require tomorrow: a standardised public market equivalent, a periodic reconciliation of reported net asset values against transacted secondary prices, and dependency disclosure for private-credit borrowers comparable to what a listed issuer files. None is a research programme; each is a legislative or rulemaking choice. Both are recorded as typed requirements below.

Frontier What it waits on from research is narrower and genuinely hard. A common-ownership measure that separates ownership from market share; an identification strategy for stewardship intensity that is not confounded by index inclusion; and a shock to passive share exogenous enough to test the price-discovery claim. Speculative On retail access there is no dependency to record because there is no data: the first cohort of defined-contribution private-asset allocations has not yet produced a fee or return series, and until it does the question cannot be asked empirically.

Established Adjacent to this brief and not inside it: whether the tax rate on carried interest changes the fee equilibrium is a taxation question and belongs to Future Taxation Models; this brief owns the fee itself, as a claim on returns. Whether a capital structure can outlive the institution that issued it belongs to Civilization Scale Investment.

Established The highest-value experiment in this subject is a disclosure rule, and it is cheap. Require every fund marketing to institutional or retail investors to report a public market equivalent against a named benchmark on a standardised fee treatment, alongside the IRR. The data already exist inside the funds. The Harris–Jenkinson–Kaplan and Phalippou dispute is not a dispute about the world; it is a dispute about sample and convention, and a common convention dissolves most of it.

Frontier Second: a mark-to-market reconciliation across a full cycle. Compare transacted secondary prices against reported net asset values, fund by fund, through a period containing a genuine liquidity event. That is the direct test of the volatility-laundering hypothesis, and the secondaries market is now large enough to supply the transacted leg. Speculative The obstacle is not method but access: secondary transaction prices are private, and the parties holding them are the parties the test would embarrass.

Frontier Third: an event study on a shock that moves common ownership without moving market share. A forced divestiture, or an index reconstitution that reshuffles which diversified holders own which competitors, gives the variation the MHHI-delta critique says the existing literature lacks. This is the single experiment that would resolve a top-journal dispute currently at a stalemate.

Frontier Fourth: repeat the FTSE4Good design on a larger and more recent inclusion event. The Berk–van Binsbergen test is the closest thing to a clean identification of the ESG cost-of-capital channel and it returned 24 basis points, insignificant. A replication on a bigger index change, or on a mandated divestment by a large public pension, would either confirm the near-null or find the channel the practitioner literature assumes.

Established Fifth, and the one with a fixed start date: track the defined-contribution private-asset cohort from its first dollar. Executive Order 14330 sets a 180-day clock on the Labor Department and directs the SEC to consider accredited-investor changes. Fee load, net return and realised liquidity for that cohort, benchmarked against each plan's default target-date fund, is a measurement nobody has to invent — it only has to be collected before the series is retrospectively reconstructed by interested parties. Speculative If it is not collected prospectively, the question of whether retail access improved retail outcomes will be argued for a decade on selected vintages.

Established The plumbing that decides who bears what is mostly reporting convention, and it is worth stating mechanically. A public market equivalent discounts a fund's cash flows at the realised return of a chosen index; a value above 1.0 means the fund beat that index over its own life. An internal rate of return instead solves for a single rate that sets the net present value of the cash flows to zero, which embeds an assumption that distributions are reinvested at that same rate. Where a fund exits early and well, IRR flatters; a public market equivalent does not. The industry reports the first and the literature argues over the second.

Established Fee arithmetic is the part that is not contested. An estimated $230 billion of carried interest was earned on 2006–2015 North American vintages whose net multiples of money were 1.55–1.63x, or about 11% a year, against public indices at similar levels. The number of private-equity multibillionaires rose from 3 in 2005 to 22 in 2020. Frontier Whether that is a transfer or a payment for a service depends entirely on the premium question in section 2, which is why the fee arithmetic and the return arithmetic have to be read together and usually are not.

Established On the credit side the mechanics are what produce the opacity. A private-credit loan is bilaterally negotiated, held to maturity by design, and marked by the holder against a model rather than a screen. Payment-in-kind interest — the borrower issuing more debt instead of paying cash — has doubled as a share of business-development-company interest income since 2019, which converts a cash-flow problem into an accrual. Frontier Over one-third of firms with private-credit-like characteristics have interest coverage ratios the IMF calls unsustainable, and that population is not visible in any public filing.

Established Market structure has a similar mechanical core. An alternative trading system reports its volume; a bilateral internaliser prints to a trade-reporting facility without displaying a quote. Off-exchange share above 50% of US equity volume means the majority of trades no longer contribute to the displayed quote that everything else is priced against. Frontier Whether that raises institutional trading costs is contested and the loudest claimant is an exchange. Payment for order flow at 0.8 basis points in equities against 35 bp in crypto is the cleanest available demonstration that the practice's cost scales with the opacity of the venue, not with the sophistication of the trader.

Frontier Tokenisation's engineering claim is settlement, not allocation. The case is that a token with atomic delivery-versus-payment removes settlement lag and collateral drag. The evidence available is $16.2 billion of tokenised real-world assets, over half of it tokenised Treasuries — instruments that already settle in one day at negligible cost. Speculative The engineering requirement that would make the case is a regulated primary issuance at scale with a published cost comparison against conventional settlement. No such comparison is in this pack.

Within this map: Civilization Scale Investment, which owns the tenor problem this brief deliberately stops short of; Future Taxation Models, which owns the rate at which carried interest and capital income are taxed while this brief owns the fee; Economic Resilience, where the 2023 bank episode appears as a resilience test rather than as an intermediation question; Long-Term Institutions, which owns whether a sovereign fund's fiscal rule binds; Institutional Design, where the incentive problem of a self-reporting intermediary is stated generally; and Digital Citizenship, which carries the identity and settlement rails the tokenisation case assumes.

Outside it: corporate finance and the empirical asset-pricing literature; market microstructure, which supplies the off-exchange and order-flow measurements; the law and economics of fiduciary duty, which supplies the ERISA question; antitrust economics, which supplies the common-ownership dispute; and prudential regulation, which supplies the IMF's framework for a non-bank credit sector.

Established Almost every number in this brief comes from a party with an interest, and saying which way the interest runs is load-bearing. Running against their own interest, and therefore weighted up: a private-equity consultancy reporting that ten-year fund IRR of 15.3% trailed a public market equivalent of 15.5%; ESG trade press reporting record outflows and 116 funds dropping ESG terms; and an exchange conceding that retail does not account for all off-exchange prints. Running with their interest, and therefore cited for facts rather than conclusions: an investment consultant counting listed companies while selling private-markets advice, a wealth manager framing index concentration, and a crypto-industry publication reporting crypto-industry scale.

Established The institutional requirement that would change this subject most is a disclosure perimeter that follows the credit. The migration of corporate lending from syndicated loans to bilateral private credit, and of corporate ownership from listed equity to buyout funds, moved a large share of corporate finance outside quarterly reporting without any decision that it should be there. Frontier No regulator has asserted that private-credit borrowers should file what a listed issuer files, and no regulator has explained why they should not. The IMF's position is the closest thing to an official statement, and it is conditional: contained now, potentially systemic if opacity and growth continue under limited prudential oversight.

Established The second requirement is a fiduciary standard that survives contact with an illiquid asset in a daily-valued account. A defined-contribution plan must value a participant's balance every business day and permit exchanges; a buyout fund reports quarterly on marks it sets itself and exits on a seven-year average. Executive Order 14330 directs the Labor Department to consider rescinding the December 2021 private-equity statement and to propose safe harbours, and directs the SEC to consider revising accredited-investor and qualified-purchaser rules. Speculative Whether a safe harbour can be written that makes an illiquid, self-marked asset appropriate for a daily-valued account of 90 million participants is an open design question, and the order does not answer it.

Frontier And the third is a supervisory answer to a question nobody has been assigned. Index providers are not regulated as owners; they are regulated as benchmark administrators. Three firms casting about a quarter of the votes in the largest equity market are exercising a governance function under an information-services licence. That is not an allegation of misconduct — it is an observation that the regulatory category and the economic function have come apart.

Established The distributional question in this brief is who is on the other side of the illiquidity backlog. A $3.8 trillion book of unexited companies, four consecutive years of falling institutional fundraising, distributions at 14% of net asset value, and then a policy opening the retirement accounts of more than 90 million people. Speculative The mechanism is coherent and the timing is suggestive; the intent claim is unevidenced and this brief does not make it. What can be said without inference is that the institutions best placed to price these assets have been reducing their commitments while the channel to the least-informed capital was being widened.

Established The second ethical fact is that a labelling industry grew to $3.16 trillion on a mechanism measured at under a basis point. Investors who bought sustainable funds to change corporate behaviour bought an outcome the best-identified estimate says did not occur through the channel they were told about. Frontier That is not a fraud finding — the engagement and mandate channels are untested rather than disproven — but it is a mis-selling risk stated in measurements, and the 335 rebrandings in a single quarter are the industry pricing that risk itself.

Established Third: fee incidence falls on savers who did not choose the arrangement. Carried interest of an estimated $230 billion on vintages that returned about what the index did was paid, in large part, out of public pension assets belonging to teachers, nurses and municipal employees who have no role in the allocation decision. Frontier The defence — that top-quartile funds outperform — is a defence only if quartile membership is identifiable in advance, and it is not.

Frontier And the ownership concentration raises a legitimacy question distinct from the antitrust one. Three firms voting roughly a quarter of the shares in the S&P 500 are voting other people's money under a fee-minimising mandate that gives them no economic reason to invest in stewardship. The complaint that they under-govern and the complaint that they over-govern are both live, and they cannot both be right. The honest position is that the influence is real, the direction of its exercise is under-measured, and the accountability path from a beneficiary to a vote is longer than in any prior ownership structure.

Established The terminal position is that the intermediary changed more than the instrument. Private credit, buyout funds and index providers are not new instruments; they are new gatekeepers between savings and firms. The genuinely new instruments price at 2 basis points and total $16.2 billion. A subject that was framed as financial technology turns out to be a subject about who gets to see the books.

Established The long-run consequence is informational rather than allocative. Prices are a society's best mechanism for aggregating dispersed information about what should be built, and they work in proportion to what is disclosed. Moving a growing share of corporate finance behind a private reporting perimeter degrades that aggregation whether or not it degrades returns, and the IMF's finding that private marks do not adjust in shocks is the measurable trace of it. Frontier Nobody has measured the aggregate cost, and the standard defence — that sophisticated counterparties price it correctly — is precisely the claim the valuation finding puts in doubt.

Frontier The second long-run fact is concentration on both axes at once. Ten companies are 41% of the index and produce 32% of its earnings; three firms cast a quarter of its votes. An index investor's outcome is now a bet on ten firms, and those firms' governance is decided largely by three fiduciaries with no incentive to spend on governing. That is a structural feature of a savings system that solved the fee problem, and it was not designed by anyone.

Speculative The most consequential open scenario is retail illiquidity at scale. If defined-contribution allocations to private assets grow into a normal single-digit percentage of $12 trillion of US plan assets, the buyer of the unexited backlog becomes a population that cannot assess it, cannot exit it, and is defaulted into it. No prior episode gives a base rate, because no prior episode combined a self-marked asset class with a daily-valued default. Handwave Forecasting the outcome requires assuming either that the safe harbours will be tight or that they will not, and that is a political prediction wearing a financial one's clothes.

These horizons track rule-making clocks, fund lives and reporting cycles rather than technology:

• 10 yr: Established The dated items dominate. Executive Order 14330's 180-day instruction to the Labor Department and its direction to the SEC on accredited-investor rules land inside this window, and the first defined-contribution private-asset cohort produces its first fee and net-return series. Frontier The $3.8tn unexited book resolves one of three ways — exit, continuation vehicle, or write-down — and the mix is the clearest available test of whether reported marks were real. Frontier Expect private credit to be brought inside some prudential perimeter if a default cycle arrives, and not otherwise; the IMF has already written the conditional. Frontier On the settlement side the dated items are a US stablecoin regime issuing its first licences and European and UK moves to T+1 equity settlement in 2027, each of which generates a published cost series; the item to watch is whether any wholesale pilot publishes an intraday-liquidity line, because no efficiency claim can be adjudicated without one.
• 25 yr: Speculative If the Bebchuk–Hirst flow trend continues, Big Three voting power passes 40% of S&P 500 votes, and the regulatory category for index providers becomes untenable. Speculative Tokenisation either achieves a regulated primary issuance at scale with published settlement savings, in which case there is finally something to measure, or remains a Treasuries wrapper at low single-digit billions. Frontier The one measurable thing inside this horizon is whether standardised public-market-equivalent reporting was ever imposed; if not, the return dispute will still be unresolved in 2050 on the same two literatures.
• 50 yr: Speculative The plausible split is that the disclosure perimeter is either restored by a crisis or normalised by habit, and there is no third path visible from here. Speculative A savings system in which most corporate ownership sits in vehicles that mark themselves is a different system from the one securities law was written for, and it will be regulated as such or not at all. Handwave Which of those happens is a political outcome, and any confident statement of it is assertion.
• 100 / 250+ yr: Handwave Beyond useful forecasting. The only base rates at that horizon are that disclosure regimes have been built after crises rather than before them, and that ownership concentration has reversed only through war, taxation or antitrust. Handwave Two regularities from a handful of episodes are a story, not a forecast.

• Depends on Nothing on this map. This brief waits on no result another brief produces: the measurements it needs are reporting conventions and disclosure rules, not discoveries. No typed depends-on edge is claimed. The nearest thing to a dependency is a seam rather than an edge — Civilization Scale Investment takes over exactly where an instrument's tenor runs out, and Future Taxation Models owns the rate at which the fees measured here are taxed.
• Requires (not on this map) Three constraints, none of them a research result and all of them things a rulemaker could supply. First, a disclosure perimeter that follows the credit: corporate lending has moved to a $2.1 trillion bilateral market whose median borrower is $0.5 billion against $4.6 billion in syndicated loans, whose payment-in-kind share of business-development-company interest income has doubled since 2019, where over one-third of comparable firms carry interest coverage the IMF calls unsustainable, and whose loan valuations adjust insufficiently during market shocks — none of which appears in a public filing. Second, a standardised public market equivalent on a common benchmark and a common fee treatment: the entire private-equity return dispute is two competent literatures on different vintages with different conventions, one finding a 1.27 public market equivalent pre-2006 and the other net multiples of 1.55–1.63x at parity with public indices post-2006, and a common convention dissolves most of the disagreement without a single new observation. Third, a fiduciary standard for self-marked assets in daily-valued accounts: Executive Order 14330 directs the Labor Department to consider rescinding the December 2021 private-equity statement and to propose safe harbours, and the SEC to consider revising accredited-investor rules, for a channel reaching more than 90 million defined-contribution participants — while the assets on offer report quarterly on marks the manager sets and exit on a seven-year average. All three are choices, and all three have so far been declined. A fourth constraint arrived with tokenised settlement, and it is the same kind of thing. No wholesale pilot has published the intraday liquidity cost of settling its own volumes atomically against the netting baseline it would replace, so the efficiency case for the redesign rests on architecture rather than on a funding line. The operators hold that number already; publishing it is a disclosure choice, and it has been declined in the same way as the other three.
• Enables Every allocation question downstream of this brief inherits its opacity: whether a long-horizon programme can be financed, whether a resilience investment can be priced, whether a household's retirement balance means what the statement says. No typed enabling edge is claimed, because the enabling relationship runs through a disclosure regime rather than through a result.
• Adjacent Corporate finance and empirical asset pricing; market microstructure; the law and economics of fiduciary duty; antitrust economics, which supplies the common-ownership dispute; prudential regulation of non-bank credit; and within this map Economic Resilience, Long-Term Institutions and Institutional Design.

Handwave “A few index funds own everything, and that is a de facto planned economy.” This circulates widely, sometimes with a conspiratorial framing, and it persists because the premise is true: the Big Three do hold about 20.5% of S&P 500 shares and cast about 25% of the votes. Established What fails is the inference. These are fiduciary holdings of other people's money; the three firms compete fiercely on fee; they vote inconsistently with one another; and the claim of coordinated control has no evidentiary support. The true premise is exactly what gives the false conclusion its durability — the same structure as every persistent fringe claim in finance.

Frontier “Public markets are dying because of regulation.” The count fell and then partially recovered — roughly 6,500 in 1997, 3,800 in 2012, 4,700 in June 2024 — and more than 60% of delistings are M&A, with involuntary delistings spiking to about 30% in 2023 on credit conditions. Established Neither the compliance-cost story nor the headcount itself establishes what it is used to establish: that firms have less access to capital, or that households have lost access to returns. Neither has been demonstrated, and a headcount cannot demonstrate either.

Frontier “Private equity outperforms public markets.” For the median investor in the post-2006 period the evidence is at parity, and one of the industry's own consultancies published 15.3% against 15.5% and wrote that parity is not what investors are paying for. Established Statements of outperformance that rest on IRR are not statements about a rate of return, and statements that rest on top-quartile funds are not statements about what an allocator can expect. The pre-2006 premium is real and is not evidence about a fund raised today.

Frontier “Common ownership demonstrably raises consumer prices.” A direct replication in the Journal of Finance attributes the original airline correlation to the market-share component of the measure rather than to ownership or control, and the original author's own revision makes the sign ambiguous by separating intra-industry from inter-industry effects. Established The question is open in both directions, and confident statements in either are overstatements of a live dispute.

Frontier “ESG investing has changed the cost of capital.” The best-identified estimate is 0.44 basis points, the cleanest inclusion event returned 24 basis points, insignificant, the flagship sovereign instrument prices at 2 basis points in advanced economies, and the ratings that the whole apparatus rests on diverge 56% on measurement. Speculative The standard rebuttal — that ESG works through engagement rather than price — is a different hypothesis, is much harder to test, and has not been tested at scale. Offering it as an answer to the price result is a change of subject, not a refutation.

Frontier “Green bonds materially lower sovereign borrowing costs.” Two basis points in advanced economies across 332 matched pairs, and the authors' own conclusion is that certification, reporting and monitoring costs plausibly exceed the saving. Frontier “Tokenisation has reallocated capital.” $16.2 billion of tokenised real-world assets, over half of it Treasuries, reported by a crypto-industry publication. That is a pilot, not a reallocation, and the comparison that settles it is $2.1 trillion of private credit.

Speculative “Passive investing is a bubble that must unwind.” A recurring prediction with a poor forecasting record, and it should be separated from the arithmetic version, which is not disputed: at 41% top-ten weight, index performance is a concentrated bet, and the informed marginal trader is a shrinking share of volume. Handwave The strong version requires a threshold nobody has identified and a mechanism nobody has measured; the weak version requires only addition.

Established “Private credit is a systemic risk.” The IMF says the opposite, explicitly, and then names the conditions under which it would not. Reporting the conditional as the finding inverts the source. Frontier And the mirror-image error is more common in industry commentary: quoting the containment sentence without the opacity condition attached to it. Both halves are the Fund's position and neither is quotable alone.

Speculative “Retail access to private markets will improve retail outcomes.” There is no evidence either way, because the policy is a year old and the first cohort has no return series. The absence is the finding, and it will stop being an absence only if somebody collects the data prospectively. Handwave Confident predictions in either direction — democratisation or distribution — are currently arguments about motive rather than about measurements.