The Intake Valve: What the IRS Disclosure Overhaul Really Says About Crypto's Cost of Compliance

A tax case that takes more than a year to close is not really a penalty. It is a queue. And a queue, unlike a fine, is priced continuously by everyone standing in it.

That distinction is the reason I read the recent remarks from Jarod Koopman, who leads criminal investigation at the Internal Revenue Service, more carefully than the headline warranted. Speaking at the UCLA Extension Tax Controversy Conference — a venue built for practitioners who argue about procedure, not for press releases — Koopman described a rebuild of the agency's voluntary disclosure program. Cases resolved in roughly 120 days instead of a year or longer. Penalties described as "significantly lower." Payment terms that would include installment agreements and offers in compromise. A target of going live before December.

Three things were absent from that description, and all three carry more weight than what was present. No number was attached to "significantly lower." No dedicated lane was promised for digital assets — Koopman was explicit that crypto cases would travel through the same pipeline as everything else. And nothing was said about the systems that must sit underneath a 120-day clock: intake tooling, data matching, staffing ratios, and the administrative record that would survive a challenge.

For most of the past decade, American crypto holders have treated the tax authority as weather — something that happens to you, occasionally, with little warning and no appeal. What Koopman described is different in kind. It is the first clear signal that the agency intends to convert its enforcement posture from sporadic summons-and-audit campaigns into a standing, metered intake process. That is not leniency. It is throughput. And throughput, once installed, changes the economics of every participant who touches a US person's transaction history.

I have spent thirteen years watching this industry's regulatory plumbing get built one joint at a time, and I have learned to distrust announcements that arrive without a Federal Register citation. But this one deserves attention, because it lands in the same eighteen-month window as the largest coordinated expansion of cross-border asset reporting in the history of tax administration. The reform is not happening in a vacuum. It is happening because the data is coming, and the pipe it will arrive through was never designed for this volume.

The Long Tail of a Compliance Backlog

To understand why a 120-day target is the interesting part of this story, it helps to remember how the United States arrived at its current position on crypto taxation — not through a single statute, but through a decade of improvisation layered on top of a property-tax framework never intended for divisible, pseudonymous, globally transferable assets.

Notice 2014-21 set the foundation in 2014 by declaring that convertible virtual currency would be treated as property for federal tax purposes. That single classification decision determined almost everything that followed. Property means basis tracking. Basis tracking means lot identification, holding periods, and a requirement that every disposal — including every purchase of a coffee, every swap of one token for another, every gas fee paid in ETH — be reported as a gain or loss event. It also meant that the ordinary machinery of third-party information reporting, the 1099 series that underpins wage and brokerage compliance, did not fit cleanly onto crypto rails for another decade.

Revenue Ruling 2019-24 then resolved the fork and airdrop questions in the least forgiving direction available: a hard fork produces ordinary income measured at fair market value on the date of receipt, whether or not the holder ever claimed, moved, or sold the new asset. I have sat with fund administrators in Nairobi who spent weeks reconstructing the fair market value of assets their clients received involuntarily and never touched. That is not evasion. That is a reporting obligation that arrives before the ability to pay it.

The enforcement side moved in parallel, and more aggressively. The 2016 John Doe summons against Coinbase, granted in 2018 with a threshold that swept in accounts above roughly twenty thousand dollars, established the template. Then came the 2019 letter campaign — the 6173, 6174, and 6174-A notices sent to more than ten thousand identified taxpayers — which functioned less as an audit and more as a psychological operation. Subsequent summonses against Kraken, Circle, sFOX, and banking intermediaries with crypto exposure followed the same logic: identify the exchange, compel the records, work backward to the individual.

What that sequence produced, by the early 2020s, was a vast inventory of unresolved cases. Some were genuinely fraudulent. Many were not. A large middle band consisted of people who had reported something, or nothing, in good faith under rules that were themselves unsettled, and who now faced a choice between silence and a disclosure process that took a year or more to conclude. For a taxpayer with a modest liability, a one-year resolution cycle is often worse than the penalty itself, because uncertainty does not sit quietly. It contaminates every other financial decision — whether to sell, whether to migrate, whether to keep records in a form that could be defended.

That is the backlog the reform is aimed at. Not the sophisticated evader, who was never going to volunteer. The long tail of ordinary people with an unresolvable file.

The Global Map: 2026 Is a Reporting Year

Here is where the domestic story becomes a macro one, and where I think most crypto-native commentary has missed the frame entirely.

Three reporting regimes are converging on roughly the same window. In the United States, the digital asset broker reporting rules finalized in 2024 created Form 1099-DA, with gross proceeds reporting phased in from 2026 and basis reporting following in 2027, alongside a real-time transaction reporting option that some brokers are expected to exercise early. In the European Union, DAC8 — Directive (EU) 2023/2226 — applies from 1 January 2026, with the first exchanges of information scheduled for 2027. And the OECD's Crypto-Asset Reporting Framework, published in June 2023, has been transposed or committed to by a large bloc of jurisdictions that together account for the overwhelming majority of global financial centre activity, with first exchanges targeted for 2027.

Read those three lines together and the picture is unambiguous. Within a single reporting season, a substantial share of the world's crypto transaction data will begin moving between tax authorities automatically, in standardized format, at account level. The United States has signaled alignment with the CARF architecture rather than building a purely domestic parallel. That means an American taxpayer's exchange activity in 2026 becomes, in 2027, a data packet that arrives in another jurisdiction's system without a summons, without a treaty request, and without the eighteen-month lag that used to make cross-border enforcement practically theoretical.

I want to be precise about why this matters for the reform, because the connection is mechanical rather than rhetorical. A voluntary disclosure program is, structurally, an intake valve. It exists so that noncompliant actors have a route into the system that costs less than being found. The efficiency of that valve determines how much of the arriving data volume converts into resolved cases, and how much of it simply piles up as unopened files. When the underlying information flow was sporadic and summons-driven, a year-long resolution cycle was tolerable — the agency could work its inventory at whatever pace its staffing allowed. When the information flow becomes automatic and annual, a year-long cycle becomes a structural failure. The valve has to be rebuilt before the pressure arrives, not after.

The reform is therefore best understood not as a softening of American crypto tax enforcement, but as the construction of the intake infrastructure required by the 2026–2027 reporting convergence. The penalty discount is the price the agency is willing to pay to keep the valve from clogging. That framing changes what an investor should take away from the announcement, and it is the single most useful thing I can offer anyone reading this.

There is a second-order consequence that receives almost no attention. Once account-level data flows automatically, the binding constraint on enforcement stops being information and becomes capacity. An agency that can see everything but can only process a fraction will inevitably triage — it will pursue the largest and most visible liabilities first, and it will increasingly rely on automated matching to generate notices at scale. That is precisely the environment in which a functioning voluntary disclosure channel becomes valuable to the taxpayer, because it is the only route that lets you choose your own timing rather than receiving a letter that chooses it for you.

What a 120-Day Clock Actually Prices

Now to the substance, and to the part of the announcement I think is being misread in both directions.

The consensus reading, where it exists at all, is that the headline is the penalty reduction. I think that is backwards. The headline is the duration.

Consider what a taxpayer is actually buying when they enter a voluntary disclosure program. They are not buying forgiveness — the tax, interest, and penalties are still owed. They are buying three things: the avoidance of criminal referral, a defined penalty framework instead of an open-ended one, and a resolution date. The first has always been available to those who qualify. The second has always been negotiable at the margins. The third has never been available at all, and it is the one that changes behaviour.

A one-year-plus resolution cycle converts a tax liability into an unpriced option written against the taxpayer's entire financial life. You cannot close a property sale cleanly. You cannot answer a lender's compliance questionnaire honestly without inviting a cascade. You cannot plan a business exit. In fund terms, you are carrying an unhedged, uncollateralized exposure of unknown size with no maturity date — the worst possible instrument to hold, because it cannot be modelled and therefore cannot be managed. Reducing that to 120 days does not reduce the liability. It reduces the variance around the liability. And in my experience, from the Terra aftermath onward, sophisticated holders will pay a surprising amount to reduce variance.

That is why I believe the 120-day clock, if it holds, will move more people into compliance than the penalty discount will. Duration is the binding constraint. Magnitude is secondary.

There is a real qualification, and it is not a small one. A 120-day target is a statement about internal process design, not a statutory guarantee. Nothing in the remarks indicated whether the clock runs from preclearance submission, from acceptance into the program, or from the completion of the taxpayer's narrative submission — and those three readings differ by months in practice. Under the existing framework, a voluntary disclosure begins with a preclearance request, proceeds to a comprehensive narrative covering a six-year disclosure period, and requires cooperation plus full payment of tax, interest, and penalties or a bona fide arrangement to pay. Any of those stages can stall on documentation. A clock that starts after the taxpayer has already spent four months assembling records is a much weaker commitment than it appears.

I would also flag the administrative-law dimension, because it is the kind of detail that decides whether a reform survives contact with litigation. Internal procedural changes of this sort are typically implemented through the Internal Revenue Manual and internal memoranda rather than through notice-and-comment rulemaking. That is generally permissible when the change affects only internal operations. But a program that alters the penalty exposure available to the public, even by discretion, sits closer to the line. If a taxpayer is denied the reduced-penalty treatment on the strength of an unpublished internal standard, the challenge writes itself. I assign this a low probability and a moderate impact, but it belongs in the risk register rather than the footnotes.

The Penalty Is a Made-Up Number, and That Is the Point

Here I want to step back and say something that will annoy practitioners on both sides, because it is true and because it explains why the missing number matters so much.

The American penalty architecture for tax noncompliance is not a market price. It is an administrative convention, assembled over decades by legislators responding to constituencies, and it bears only an incidental relationship to the actual social cost of the conduct it punishes. The civil fraud penalty under section 6663 is 75% of the underpayment. The fraudulent failure-to-file penalty under section 6651(f) escalates to 75%. The accuracy-related penalty under section 6662 is 20%. Failure-to-file under section 6651(a)(1) runs 5% per month to a 25% cap. These numbers do not emerge from any observable supply-and-demand curve for compliance. They are negotiated political settlements, and they function as reference points in a bargaining process rather than as clearing prices.

I have made this argument before in a different context — that the interest rate curves governing decentralized lending markets are largely arbitrary constructs dressed in the language of market discovery, set by governance parameters that respond to tokenholder sentiment rather than to the actual marginal cost of capital across the credit surface. The same critique applies with equal force here. A 75% fraud penalty and a 4% supply rate on a lending pool share a common defect: both are numbers someone chose, defended by a narrative of necessity, and neither is continuously validated against the thing it claims to measure.

Which is why "significantly lower" is not a detail. It is the entire substance. A discount from 75% to 60% is a rounding error that changes no behaviour. A discount from 75% to 25%, paired with a defined 120-day resolution, converts an unmanageable exposure into a calculable one and will move a genuinely large population. The difference between those two worlds is a single number the agency has not published, and it is the number that determines whether this reform matters.

My working assumption, based on how these programs have historically been structured, is that the reduction will be differentiated by conduct category rather than uniform — a larger concession for the inadvertent reporter with a clean narrative and full cooperation, a much smaller one for anyone whose facts suggest wilfulness. That is the sensible design. It is also the design that produces the most litigation over characterization, because the line between "I did not understand the fork rules" and "I chose not to report the fork" is exactly where the money sits.

The practical implication for anyone with an open file is uncomfortable but clear. Do not model your exposure on the announced direction. Model it on the published range. Until a number exists in a citable document, any plan built on "significantly lower" is a plan built on a rumour with a good source.

Why There Is No Crypto Lane — and Why That Is Correct

Koopman's statement that digital assets will not receive a separate disclosure track has been read in some quarters as indifference, or as evidence that the agency does not understand the asset class. I read it as the opposite, and I think it is the most technically sound decision in the whole package.

The instinct to build specialized infrastructure for every new category is a familiar one in this industry. We have spent five years watching teams stand up dedicated data availability layers for rollup ecosystems that, in the overwhelming majority of cases, do not generate enough throughput to justify the marginal cost of a separate availability market. The architecture is elegant. The demand is not there. Most rollups would be perfectly served by posting to the base layer and accepting the cost, and the ones that genuinely need independent DA are a small minority whose economics are atypical. Building bespoke infrastructure ahead of demonstrated volume is a reliable way to produce something expensive, underused, and difficult to retire.

A dedicated crypto disclosure channel would be the same mistake in regulatory form. What would it actually contain? Strip away the exotic branding and the median crypto tax file is unremarkable: a set of exchange statements, a handful of self-custody wallets, some DeFi interactions whose cost basis is reconstructible, a staking income schedule, and possibly a fork or airdrop event. The analytical difficulty is real, but it is a difficulty of record reconstruction, not of legal theory. It does not require a separate adjudicative track. It requires a functioning general track that can absorb messy documentation — which is precisely what a 120-day intake process with a cooperative narrative requirement is designed to do.

There is a second reason the unified approach is defensible, and it is about signalling. A dedicated crypto channel would communicate, accurately or not, that the agency views digital asset noncompliance as a categorically different and somewhat lesser species of error. That is a gift to nobody. It would be cited for years by people arguing that crypto occupies a privileged compliance position, and it would invite exactly the sort of aggressive positioning that ends in a fraud referral. Uniform treatment is not hostility. It is the refusal to grant a special status that would later have to be revoked.

Where the Congress-side proposal differs is instructive. The Digital Asset Taxation Certainty Act, which advanced through the House Ways and Means Committee with a provision numbered 502 establishing a separate disclosure route for taxpayers whose crypto errors were inadvertent, is attempting something narrower and more precise: it is legislating a subjective state — inadvertence — into a procedural entitlement. That is a legitimate thing for a legislature to do, and it is the kind of fine-grained responsibility allocation that administrative discretion handles poorly. But it is also slow, and it is contingent on floor votes in both chambers plus signature, and it is not a substitute for the intake capacity the agency needs this year.

So the two tracks will likely coexist, imperfectly. The administrative reform provides the floor. The legislative proposal, if it survives, provides a safe harbour for a defined class. Anyone planning around this should assume the floor arrives first and the harbour may never arrive at all.

Where the Identification Layer Actually Lives

There is a structural feature of this entire build-out that deserves more scrutiny than it is receiving, and it connects directly to a design choice this industry has been making for years without fully pricing the consequence.

Every reporting regime described above — 1099-DA, DAC8, CARF — depends on the same underlying capability: the ability to attribute an on-chain address to a legal person. That attribution does not happen on-chain. It happens at the boundary, in the custodial venues and fiat gateways where identity is captured, and it is then extended backward through chain analysis heuristics that are probabilistic rather than definitive.

The consequence is that the entire compliance edifice rests on a small number of chokepoints, and those chokepoints are permissioned by design. This is the structural risk I have written about in the stablecoin context for some time: a compliance-first issuance model that can freeze an address on short notice is efficient precisely because it concentrates control, and concentration of control is a single point of policy failure. The same architecture that makes a stablecoin issuer a good partner for a tax authority makes it a fragile foundation for a reporting system, because the authority's visibility is only as good as the issuer's continued willingness and legal ability to identify its holders.

I do not raise this as an argument against compliance. I raise it because it is the thing that determines whether the 2026 data inflow actually materializes at the volume implied by the policy documents. If a meaningful share of activity migrates to self-custody, to non-custodial swap venues, or to jurisdictions that implement CARF on a delayed or partial basis, the data inflow will be materially smaller than the architecture assumes — and the 120-day clock will be calibrated against a volume that never arrives, while the cases that do arrive will be the hardest ones, because those are the ones that surfaced through attribution rather than through voluntary reporting.

That is an underappreciated asymmetry. Automated reporting does not capture the population uniformly; it captures the custodial population thoroughly and the non-custodial population barely at all. The result, over time, is a compliance regime that taxes the visible and misses the invisible, which is both inequitable and unstable, because the visible participants eventually notice.

And this is where the older, harder truth reasserts itself. Chain analysis does not forget. Every transaction a holder has ever signed is permanently recorded, and the analytical tools that interpret those records improve every year. A wallet that looks opaque in 2026 may be trivially attributable in 2029. The ledger remembers what the algorithm forgets — and the algorithm, in this case, is the human capacity to forget what one did during a bull market with a hardware wallet and no records.

The Transmission Lag Nobody Models

I want to bring in something concrete from my own work, because the abstract argument about data flows tends to obscure a practical question that actually determines outcomes: how long does a regulatory or institutional signal take to reach the price of the asset in a market that is not the one where the signal originated?

In early 2024, after the US spot bitcoin ETF approvals, I led the integration of BlackRock's IBIT flow data into our fund's daily liquidity model here in Nairobi. The exercise was not about predicting price. It was about measuring transmission. What we found, running the flow series against on-chain exchange reserve data across a set of venues weighted toward emerging-market access, was a persistent lag of roughly fourteen days between a meaningful change in ETF creation and redemption activity and a corresponding, measurable shift in the liquidity conditions our counterparties actually faced. Fourteen days is not a rounding error. In a market where funding rates reset every eight hours, two weeks is an eternity, and it is the difference between entering at a level and entering after the level has already been repriced by someone with faster plumbing.

I raise this because the same lag logic applies to the tax reform, and nobody is modelling it. The mechanism is different but the structure is identical. A penalty reduction announced in the fourth quarter does not change behaviour in the fourth quarter. It changes behaviour when the first cohort of disclosures completes, when practitioners observe the actual penalty outcomes, and when those outcomes become the reference points that advisors quote to clients. That feedback loop is measured in quarters, not weeks, and it does not begin until the program is not only live but demonstrably predictable.

Which means the realistic timeline is something like this: the program launches, the first tranche of cases runs through it, practitioners compare notes, and a de facto penalty schedule emerges from observed outcomes rather than from published rules. That emergent schedule — not the announcement — is what the market will price. If the observed outcomes come in materially harsher than the rhetoric implied, the program will be quietly ignored and the backlog will persist. If they come in close to the rhetoric, the intake will fill, and the second-order effect will be a modest but real wave of asset sales as taxpayers fund settlements.

I want to be careful not to overstate that last point. In 2020, modelling the effect of MakerDAO stability fee increases on local USD-DAI arbitrage activity, I found that a liquidity gap affecting roughly forty smallholder farmers using stablecoin rails for remittances could be closed with a change in slippage tolerance — a technical parameter, not a policy intervention. The lesson I took from that was not that policy is powerless. It was that policy transmits through specific, identifiable mechanisms, and if you cannot name the mechanism, you are guessing. For the tax reform, the mechanisms are: timing of resolution, magnitude of the actual penalty outcome, and availability of payment terms. Nothing else in the announcement has a transmission channel into price.

On magnitude, the payment terms deserve more attention than they are getting. The explicit reference to installment agreements and offers in compromise is the most operationally significant element of the package, because it addresses the constraint that actually prevents people from disclosing: the inability to pay a lump sum on demand. An offer in compromise — a mechanism by which the agency accepts less than the full amount owed on grounds of doubt as to collectibility or doubt as to liability — is the formal acknowledgment that some portion of the liability is simply not recoverable. When an enforcement body starts building collection-flexibility into a disclosure program, it is telling you something about the composition of its inventory. It is telling you that a meaningful share of the backlog consists of people who owe more than they can pay.

I read that as a residual effect of the 2022 drawdown. A holder who realized gains in 2021 and watched the value of the underlying collateral fall by seventy percent or more is not refusing to pay. They are unable. A program designed without that reality would fail on contact, and the fact that the design includes it suggests the agency has looked at its own case files.

The Contrarian Read: This Is an Intake Story, Not a Leniency Story

The dominant interpretation forming around this announcement is that American crypto tax enforcement is softening. I think that interpretation is wrong in a way that will cost people money, and I want to state the alternative as sharply as I can.

Enforcement intensity and enforcement efficiency are not the same variable, and they move in opposite directions here. Every element of the reform that looks like a concession is, on closer inspection, an investment in the agency's ability to process cases faster and at greater volume. A shorter resolution cycle means more cases closed per unit of staff time. A lower penalty means more people enter the program voluntarily, which means fewer contested files and fewer resources burned on litigation. Installment agreements mean the agency collects revenue it would otherwise write off. Read as a portfolio, these are not the choices of an institution softening its stance. They are the choices of an institution that has decided its constraint is throughput rather than authority, and is optimizing accordingly.

The historical pattern supports a darker reading than the headlines do. Programs that formalize voluntary disclosure have, in the American context, frequently preceded periods of intensified automated enforcement. The logic is not conspiratorial; it is operational. You build the voluntary channel first because it is the cheapest way to clear the compliant-but-embarrassed segment, and clearing that segment frees capacity to pursue the segment that will never volunteer. The channel is not the alternative to enforcement. It is the sorting mechanism that makes enforcement more precise.

There is also a reading of the timing that I hold with lower confidence but consider worth flagging. The choice to preview the reform at an academic tax controversy conference, rather than through a formal notice, is a classic trial balloon. It lets the agency observe how practitioners and the industry react before committing to specific numbers, and it preserves the option to revise the discount downward if the reaction suggests the market would accept less. Anyone treating the current description as the terminal design is misreading the genre. Trial balloons are, by construction, not the final shape of the thing.

Where I part company with the more alarmist readings is on the question of intent toward the asset class. This is not a crypto-specific action, and treating it as one distorts the analysis. The agency is rebuilding a general intake function because its general inventory is overloaded and its general data inflow is about to multiply. Crypto happens to be the segment where the mismatch between the rules and the reality of record-keeping is most acute, which is why the segment will feel the reform most. But the reform is not aimed at crypto. It is aimed at volume, and crypto is where the volume is.

The Two-Track Problem

The genuine contrarian risk in this package is not the penalty discount or the timeline. It is the coexistence of two incompatible policy routes, and the compliance arbitrage that coexistence creates.

The administrative route runs through the agency's discretion: a unified disclosure process, no digital asset lane, a penalty framework set internally, a target date before December. The legislative route runs through the Digital Asset Taxation Certainty Act and its section 502, which would create a distinct disclosure program for inadvertent crypto errors, having already cleared the House Ways and Means Committee.

These are not complements. They are competing allocations of the same discretion. If the administrative route lands first with a defined penalty schedule, the legislative route becomes partially redundant and loses momentum. If the legislative route advances while the administrative details remain unpublished, taxpayers face a genuine question about which framework to file under — and the rational answer, for anyone whose facts might qualify as inadvertent, is to wait. Waiting is free if no clock is running and the penalties are not yet fixed. But waiting is not free in reality, because interest and failure-to-pay additions continue to accrue on the outstanding balance every month the file stays open.

That is the trap I would most want a reader to avoid. The rational response to policy uncertainty is delay, and delay is the one strategy that guarantees a worse outcome regardless of which route ultimately prevails. Interest compounds. The disclosure window does not.

There is a further wrinkle that practitioners will recognize immediately. If the legislative safe harbour is limited to inadvertent errors, and the administrative process applies uniformly, then characterization becomes the decisive variable — and characterization is exactly the question that is hardest to resolve without an adjudicated record. A taxpayer who believes their error was inadvertent will want the legislative route. A taxpayer whose facts are ambiguous will want the administrative route, because a uniform process is more predictable than a standard that turns on subjective state. Choosing between them requires an honest assessment of one's own file, which is precisely the assessment most people are least equipped to make about themselves.

I have watched this dynamic before, in a different register. In 2022, in the weeks after the Terra collapse, our fund's exposure committee spent days arguing over whether algorithmic stablecoin holdings should be cut to zero immediately or phased out over a quarter to avoid realizing losses at the bottom. The argument for phasing was reasonable on its face and wrong in substance, because it priced the known loss against an unknown tail. We cut to zero over a weekend, rebalanced into bitcoin and ether, and finished September down four percent against an industry average nearer thirty. The lesson was not that speed always wins. The lesson was that when the decision framework itself is unstable, the cost of waiting to see how it resolves is almost always underestimated by the people doing the waiting.

Applied here: do not wait for the two tracks to resolve. Build the file now, in a form that would satisfy either route. That means a complete six-year reconstruction of acquisition and disposal records, a narrative that is honest about what was known and when, and a payment plan that is realistic rather than optimistic. The work is the same under both frameworks. Only the label on the envelope differs.

What I Am Watching Between Now and December

I do not think the interesting question is whether the reform happens. The direction is set, the incentive for the agency is structural, and the data inflow arriving in 2026 and 2027 makes the intake rebuild close to unavoidable. The interesting question is what the reform actually delivers, and that will be legible in a small number of specific, observable places.

First, the Federal Register and the Internal Revenue Manual. If the change appears only as an internal memorandum, the penalty framework remains discretionary and unappealable, and the reform is weaker than advertised. If it appears as a published procedure with a defined penalty schedule, it becomes something a taxpayer can actually plan against.

Second, the revised Form 14457 and its instructions. The preclearance and narrative requirements are where the 120-day clock lives or dies. If the documentation burden increases while the resolution target shrinks, the target is arithmetic fiction.

Third, the observed acceptance rate on offers in compromise within disclosure cases. This is the quietest and most informative number in the entire package. A high acceptance rate tells you the agency is prioritizing collection over punishment. A low one tells you the flexibility language was aspirational.

Fourth, whether the section 502 route advances past committee. If it stalls, the administrative process becomes the only game, and the characterization fight moves inside the agency rather than into the statute.

And fifth, something that will not appear in any official document: whether practitioners start advising clients to file. Advisory behaviour is the real leading indicator. When the tax bar begins recommending the voluntary route in ordinary client conversations, the reform has worked. Until then, it is a press release with a good source.

I will close with the thing that has been true through every cycle I have watched, and which the current sideways market makes easy to forget because nothing is forcing the issue. Safety is the only yield that compounds over time. Not the highest return in a given quarter — the return you are still holding after the enforcement cycle, the reporting cycle, and the liquidity cycle have all run their course. A clean tax file is not an exciting asset. It is a durable one, and durability is what survives a regime change in Washington, a directive from Brussels, and a data exchange that arrives in 2027 whether or not anyone prepared for it.

We build walls not to keep out, but to keep safe. The agency is building one now, and it is building it with an intake valve rather than a battering ram. Whether that valve turns out to be a door or a filter depends entirely on numbers that have not yet been published — and on how many people decide, in the meantime, that a queue is a thing you can stand in forever.

You cannot. The clock starts whether you enter the program or not. The only question is who is holding the stopwatch when it does.